{"key":"mind","title":"Orchestration & autonomy","kanji":"心","one_line":"A local, verifiable mind that runs every module, asks humans only for judgement, and survives being cut off.","thesis":"A fully automated ORIJO space is not one AI but a stack: a semantic nervous system, a deterministic safety kernel that owns every actuator, optimisers (MPC/RL) that propose setpoints, robots coordinated as fleets, and a remote operations centre that answers questions rather than steering. Field evidence says advanced control saves 9-21%, not 40%, and autonomy at scale still budgets human minutes (Waymo runs about one assistant per ~43 vehicles). The non-obvious advantage of ORIJO is that every module is a catalogue item: its semantic graph, emulator and fault signatures can ship in the fold passport, so commissioning labour, the main barrier to Brick, MPC and FDD adoption, disappears, and identical siblings across many sites create the labelled fault data that single buildings never have. The cautionary lessons from Freight Farms, Antarctic observatories and Gateway are that the mind must run locally, degrade gracefully for weeks, and outlive any one company or programme.","subsystems":[{"id":"MIND-S1","name":"Nervous system: field buses, protocols and the semantic model","function":"Get every sensor, actuator, battery, pump, hinge lock and robot onto one addressable, self-describing data plane so that controllers and agents can be written once and reused on every ORIJO module.","how_it_works":"Devices speak a field protocol (BACnet/IP or MS/TP for HVAC, Modbus RTU/TCP for inverters and pumps, Matter over Thread/Wi-Fi for consumer devices, CAN for battery packs). An edge gateway normalises them into a publish/subscribe bus (MQTT with Sparkplug B state-aware payloads, or OPC UA for industrial equipment) so that each point has a birth/death certificate and last-known-good value. A semantic graph (Brick ontology, ASHRAE 223P topology, Project Haystack tags) states what each point is and how equipment is connected (e.g. pump P1 feeds loop L2 that serves zone Z3), so that a fault rule or optimiser can query 'all supply-air temperature sensors downstream of this heat pump' instead of hard-coding point names. Tag systems without formal semantics drift into inconsistency; Fierro et al. showed Haystack's informal tags admit ambiguous interpretations, which motivated the formal Brick model. For ORIJO, the graph is generated from the configurator, not by a commissioning engineer, because every module and JO-PORT is a known catalogue item.","options":[{"name":"BACnet + Modbus field layer","principle":"Object/property model over IP or RS-485; polled or COV subscriptions","maturity":"COMMERCIAL","trl":9,"trl_why":"Decades of building deployment; but security optional and often absent","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7"],"limits":"Weak native security (BACnet/SC fixes this, but legacy devices dominate); no semantics beyond object names [src-kaur2015, src-granzer2010]"},{"name":"Matter 1.5 for residential devices","principle":"IP-based application layer over Thread/Wi-Fi/Ethernet with device-type clusters; 1.5 adds tariffs, grid carbon intensity, metering, solar data, bidirectional EV charging, closures and soil sensors","maturity":"COMMERCIAL","trl":7,"trl_why":"Specification released Nov 2025; energy clusters only now entering certified products","key_numbers":[{"label":"Matter 1.5 release date","value":"2025-11-20","unit":"date","basis":"reported","sources":["src-matter15"]}],"best_for":["E1","E2","E7"],"limits":"Consumer-grade scope; no ontology of plant topology; heat-pump and battery clusters not described in the 1.5 announcement [src-matter15]"},{"name":"MQTT + Sparkplug B bus","principle":"Broker-based pub/sub with stateful birth/death messages and report-by-exception","maturity":"COMMERCIAL","trl":9,"trl_why":"Standard in industrial IIoT; small footprint suits 10-15 W edge PC","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Broker is a single point of failure unless duplicated; semantics still external"},{"name":"Brick / ASHRAE 223P semantic graph","principle":"RDF/OWL graph of equipment, points and connections, queried with SPARQL","maturity":"PILOT","trl":7,"trl_why":"Field-demonstrated for scaling ASHRAE Guideline 36 sequences; 223P newer and less deployed","key_numbers":[{"label":"Brick founding paper citations","value":375,"unit":"citations (Semantic Scholar API, 2026-10-29)","basis":"measured","sources":["src-balaji2016"]}],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Tooling still research-grade; models are usually hand-authored after construction [src-roa2023, src-elias2026]"}]},{"id":"MIND-S2","name":"Supervisory optimiser: rules, MPC, reinforcement learning, LLM planner","function":"Choose setpoints and schedules for heat, power, water, crop climate and robot tasks that minimise cost, energy and wear while honouring comfort, crop and safety constraints.","how_it_works":"A grey-box (resistance-capacitance) or learned model predicts zone temperature, tank level, battery state of charge and crop climate over a 12-48 h horizon using weather and price forecasts; model predictive control solves a constrained optimisation every 5-15 min and applies only the first move. Reinforcement learning instead learns a policy from logged data; safe variants start from the existing rule-based policy and penalise deviation (KL regularisation, batch RL) so the first day of deployment is no worse than the baseline. DeepMind's data-centre system predicted future PUE with an ensemble of neural networks plus separate ensembles predicting temperature and pressure an hour ahead, and every five minutes sent selected actions back to the facility. Field evidence is consistently in the 9-21% savings band, not the 40% of the best-known headline, and 71% of published field demos used protocols that may give unreliable estimates. LLM agents are useful as translators of human intent into constraints and as code/model generators (e.g. auto-building MPC), but they fabricate schedules and fail commitments, so they sit above the optimiser, never in the actuation path.","options":[{"name":"Rule-based + ASHRAE Guideline 36 sequences","principle":"Deterministic if-then sequences with trim-and-respond resets","maturity":"COMMERCIAL","trl":9,"trl_why":"Industry default; the baseline every advanced controller is measured against","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Leaves 9-21% savings on the table in field comparisons [src-luo2022, src-mulayim2025]"},{"name":"Model predictive control (grey-box)","principle":"Receding-horizon constrained optimisation over a physics model","maturity":"PILOT","trl":7,"trl_why":"104 published field demonstrations, little commercial adoption","key_numbers":[{"label":"Reliable-protocol MPC/RL cost savings, residential","value":16,"unit":"% (duration-weighted)","basis":"measured","sources":["src-khabbazi2025"]},{"label":"Reliable-protocol MPC/RL cost savings, commercial","value":13,"unit":"% (duration-weighted)","basis":"measured","sources":["src-khabbazi2025"]},{"label":"MPC heat-pump energy saving vs constant setpoint (one house)","value":"18.1 (95% CI 4.4-30.9)","unit":"%","basis":"measured","sources":["src-mulayim2025"]}],"best_for":["E1","E2","E4","E5","E7","E8"],"limits":"Engineering effort per site; only 13 of 104 papers report deployment/maintenance cost [src-khabbazi2025]"},{"name":"Supervised/safe reinforcement learning","principle":"Policy learned from data, gated by uncertainty and operator constraints","maturity":"PILOT","trl":7,"trl_why":"Autonomous control in Google data centres since 2018 (TRL 8 there); two Trane-site live experiments 2022 are research trials (TRL 6-7); TRL 6 in homes","key_numbers":[{"label":"Data-centre cooling energy reduction (recommendation mode)","value":"up to 40","unit":"%","basis":"reported","sources":["src-dm2016"]},{"label":"Autonomous data-centre cooling, average saving after 9 months","value":"about 30 (from about 12 at launch)","unit":"%","basis":"reported","sources":["src-dm2018"]},{"label":"Commercial cooling RL saving at two live sites","value":"9 and 13","unit":"%","basis":"measured","sources":["src-luo2022"]},{"label":"RL heat-pump saving vs constant setpoint (one house)","value":"20.9 (95% CI 2.6-38.3)","unit":"%","basis":"measured","sources":["src-mulayim2025"]}],"best_for":["E1","E3","E4","E7"],"limits":"Distribution drift sim-to-real; RL caused three occupant discomfort reports during adaptation [src-mulayim2025, src-liu2022]"},{"name":"LLM / agentic planner above the optimiser","principle":"Tool-calling language model turns natural-language goals into constraints, schedules, queries and controller code","maturity":"EXPERIMENTAL","trl":4,"trl_why":"Benchmarks and 7-day rolling simulations; no audited field deployments in actuation path found","key_numbers":[],"best_for":["E1","E2","E7"],"limits":"Fabricated schedules, failed commitments and reasoning-to-action failures observed; low cost does not imply feasible behaviour [src-jonah2026]"}]},{"id":"MIND-S3","name":"Safety envelope: runtime assurance and functional safety","function":"Guarantee that no learned or remote command can push the habitat, robots or people outside a verified safe operating envelope, and that every failure ends in a known safe state.","how_it_works":"A simplex-style architecture splits control into an untrusted high-performance channel (MPC, RL, LLM, remote operator) and a small trusted channel (PLC or safety microcontroller) that checks every proposed action against hard limits and falls back to a verified baseline controller. DeepMind's autonomous cooling removed low-confidence actions, checked actions against operator constraints in the cloud and again in the local control system, and failed over to a neutral state on any violation; human override always superseded AI. NASA's Gateway Vehicle System Manager uses assume-guarantee contracts specified in PlusCal/TLA+ and checked at runtime. Industrial patents now formalise 'setpoint source selection' between AI and local controllers. For robots near people, ISO 10218-1/2:2025 (which absorbed ISO/TS 15066 power-and-force-limiting content) and ISO 13482:2014 for personal care robots set the hazard analysis and force/speed limits; the ORIJO design rule is that the karakuri (unpowered) state of each fold, valve and hinge is itself the fallback.","options":[{"name":"Two-layer verification + neutral failover (DeepMind pattern)","principle":"Uncertainty gating, cloud constraint check, local re-verification, automatic failover to rules","maturity":"COMMERCIAL","trl":8,"trl_why":"Operating in Google data centres since 2018","key_numbers":[{"label":"Action cadence of autonomous cooling AI","value":5,"unit":"minutes per decision cycle","basis":"reported","sources":["src-dm2018"]}],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Envelope only as good as the constraint set operators write"},{"name":"Formally specified runtime contracts (Gateway VSM)","principle":"Assume-guarantee contracts in TLA+ checked during development and at runtime","maturity":"EXPERIMENTAL","trl":5,"trl_why":"Developed and verified in lab for Gateway; Gateway paused in March 2026, never flown","key_numbers":[],"best_for":["E4","E8"],"limits":"Specification effort; programme risk [src-dabney2023, src-gateway-ifl]"},{"name":"Deterministic setpoint arbiter (patented industrial pattern)","principle":"Local controller keeps authority; AI setpoint accepted only if it passes checks, else local setpoint used","maturity":"PILOT","trl":6,"trl_why":"Granted US patent 2025-06-10, assignee Phaidra Inc.; product deployment of this exact pattern not independently verified","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7"],"limits":"Patent held by Phaidra (a potential competitor/partner): design-around or licence needed [src-pat-us12326701]"},{"name":"Robot functional safety (ISO 10218:2025, ISO 13482:2014)","principle":"Risk assessment, speed and separation monitoring, power and force limiting, protective stops","maturity":"COMMERCIAL","trl":9,"trl_why":"Published standards; cobot certifications routine","key_numbers":[{"label":"ISO 10218 parts 1-2 revision published","value":2025,"unit":"year","basis":"reported","sources":["src-iso10218-2025"]}],"best_for":["E1","E2","E3","E4","E5","E6","E7"],"limits":"ISO 13482 dates from 2014 and does not cover learned behaviours or teleoperation explicitly (not verified against full text) [src-iso13482]"}]},{"id":"MIND-S4","name":"Self-diagnosis: fault detection, prognostics and the calibrated digital twin","function":"Notice drift and failure before people do, isolate the faulty component, and decide whether to compensate, swap a module or call a human.","how_it_works":"Rule-based AFDD encodes physics (e.g. mixed-air temperature must lie between return and outdoor air; a pump drawing power with zero flow is dead-headed). Data-driven FDD trains supervised classifiers on labelled faults, unsupervised anomaly detectors on normal operation, or hybrids; reviews classify these and consistently note the lack of labelled real-world fault data. A digital twin is a simulation model continuously calibrated against telemetry, so residuals between prediction and measurement localise faults and the same model is reused by the MPC. Katipamula and Brambley's landmark review estimated that degraded and improperly controlled equipment wastes 15-30% of energy in commercial buildings, which is the economic case for FDD. In ORIJO, every module ships with known physics and identical siblings, so fleet-wide comparison (the same KURA-3 at 200 sites) gives a far stronger anomaly baseline than any single building has.","options":[{"name":"Physics/rule AFDD","principle":"Expert rules on sensor relationships","maturity":"COMMERCIAL","trl":8,"trl_why":"Built into commercial BMS analytics products","key_numbers":[{"label":"Energy wasted by degraded/poorly controlled equipment in commercial buildings","value":"15-30","unit":"%","basis":"reported","sources":["src-katipamula2005"]}],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Rule maintenance; high false alarm rates when sensors drift"},{"name":"Data-driven FDD (supervised/unsupervised/hybrid)","principle":"Classifiers and anomaly detectors on BMS time-series","maturity":"PILOT","trl":6,"trl_why":"Hundreds of studies; few long-term field deployments","key_numbers":[{"label":"Citations of Mirnaghi & Haghighat data-driven FDD review","value":360,"unit":"citations (Semantic Scholar)","basis":"measured","sources":["src-mirnaghi2020"]}],"best_for":["E1","E2","E7"],"limits":"Labelled fault data scarce; transfer between buildings weak [src-mirnaghi2020, src-chen2023fdd]"},{"name":"Calibrated digital twin (BIM + IoT)","principle":"Simulation model updated from telemetry; residual-based diagnosis","maturity":"PILOT","trl":6,"trl_why":"21 use cases and 6 enterprise solutions reviewed; mostly pilots","key_numbers":[{"label":"Building DT use cases in systematic review","value":21,"unit":"case studies","basis":"measured","sources":["src-cespedes2024"]}],"best_for":["E1","E2","E3","E4","E5","E7","E8"],"limits":"Older-building data gaps; model maintenance cost [src-cespedes2024]"},{"name":"Generative anomaly response for life support","principle":"ML on cyclic ECLSS processes to detect and respond to anomalies on a deep-space habitat","maturity":"EXPERIMENTAL","trl":3,"trl_why":"Simulation study","key_numbers":[],"best_for":["E4","E8"],"limits":"No flight or analog validation found [src-ibrahim2025]"}]},{"id":"MIND-S5","name":"Robot fleet orchestration","function":"Let several robots from different vendors (TSURU crease-rail robots, farm gantries, cleaning and delivery robots) share doors, lifts, corridors, chargers and tasks without collisions or deadlocks.","how_it_works":"Each vendor fleet keeps its own navigation; a fleet adapter translates it into a common API. Open-RMF holds a shared traffic schedule of predicted robot trajectories, negotiates conflicts, and treats doors, lifts and chargers as shared resources; lanes and vertices can be placed in mutex groups so only one robot holds them. Tasks (deliver, clean, patrol, charge) are bid for and allocated. VDA 5050 standardises the master-control-to-AGV message set in warehouses, and lightweight systems such as Robofleet handle secure, bandwidth-tolerant robot-to-operator links. For ORIJO, the crease rails are a pre-mapped graph, so the 'navigation' problem reduces to discrete-lane scheduling, exactly the case Open-RMF was built for.","options":[{"name":"Open-RMF on ROS 2","principle":"Shared traffic schedule, fleet adapters, door/lift/charger integration, task allocation","maturity":"PILOT","trl":7,"trl_why":"Open source since Oct 2021; used in airports, hospitals, malls per vendor-neutral press; maintained by OSRF/Intrinsic","key_numbers":[{"label":"Open-RMF public launch","value":"2021-10","unit":"date","basis":"reported","sources":["src-openrmf-news"]}],"best_for":["E1","E2","E6","E7"],"limits":"Integration effort per fleet adapter; single-site scope [src-openrmf-news, src-amrf2026]"},{"name":"VDA 5050 master control","principle":"Standard JSON/MQTT order and state messages between fleet manager and AGVs","maturity":"COMMERCIAL","trl":8,"trl_why":"Widely adopted in European intralogistics (not independently verified here)","key_numbers":[],"best_for":["E7"],"limits":"Warehouse-centric; no doors/lifts negotiation [src-amrf2026]"},{"name":"Robofleet-style secure remote link","principle":"Central server deduplicates and secures ROS messages over intermittent low-bandwidth links","maturity":"EXPERIMENTAL","trl":5,"trl_why":"Research system with field use at university scale","key_numbers":[],"best_for":["E2","E3","E4","E5"],"limits":"Not a scheduler; research software [src-robofleet]"}]},{"id":"MIND-S6","name":"Human-in-the-loop: remote operations centre and teleoperation","function":"Provide the scarce human judgement that autonomy still needs, at a ratio of one person to many sites, without humans ever being in the real-time safety loop.","how_it_works":"The autonomous system raises a structured request (what it sees, what options it considered, its confidence) to a queue; a remote agent answers with advice the system can accept or reject, rather than driving directly. Waymo runs about 70 remote-assistance agents at any time for about 3,000 vehicles across four redundant centres, and states it does not rely on remote drivers. Household humanoids invert this: 1X sells NEO at $20,000 (or $499/month) with owner-scheduled teleoperation that also collects training data, with blurring of people and owner-set no-go zones. Maritime regulators have codified the pattern: the IMO MASS Code (adopted May 2026, voluntary from 1 July 2026) centres on Remote Operations Centres while the master keeps overall responsibility. Situation-awareness research from MUNIN shows the hard part is handover: the operator must be brought into context before acting.","options":[{"name":"Remote assistance queue (advice, not driving)","principle":"System-initiated requests, human answers, system decides","maturity":"COMMERCIAL","trl":9,"trl_why":"At scale in robotaxi operations","key_numbers":[{"label":"Remote-assistance agents on duty","value":70,"unit":"agents on duty at any time (worldwide)","basis":"reported","sources":["src-waymo-futurism"]},{"label":"Fleet served","value":"about 3,000","unit":"vehicles","basis":"reported","sources":["src-waymo-futurism"]}],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Labour location and transparency became a Senate issue in Feb 2026 [src-waymo-futurism]"},{"name":"Scheduled teleoperation for manipulation","principle":"Remote human pilots robot body through cameras; data trains autonomy","maturity":"PILOT","trl":6,"trl_why":"Consumer deliveries targeted 2026; no independent data on autonomy share","key_numbers":[{"label":"NEO purchase price","value":20000,"unit":"USD","basis":"reported","sources":["src-1x-engadget"]},{"label":"NEO subscription","value":499,"unit":"USD per month","basis":"reported","sources":["src-1x-engadget"]}],"best_for":["E1","E2","E7"],"limits":"Privacy (camera feed into the home), latency, labour cost; chore autonomy 'not anytime soon' per trade press [src-1x-robotreport]"},{"name":"Remote Operations Centre under a regulatory code","principle":"ROC supervises several vessels; defined responsibility and fallback","maturity":"PILOT","trl":7,"trl_why":"Voluntary IMO MASS Code takes effect 1 July 2026; mandatory code targeted for adoption by 2030 and entry into force 1 Jan 2032","key_numbers":[{"label":"Mandatory MASS Code expected entry into force","value":"2032-01-01","unit":"date","basis":"target","sources":["src-imo-mass"]}],"best_for":["E6"],"limits":"Experience-building phase still under way [src-imo-mass]"},{"name":"Remote greenhouse growing via setpoint API","principle":"Teams set climate/irrigation setpoints remotely; process computer actuates","maturity":"PILOT","trl":7,"trl_why":"Four Autonomous Greenhouse Challenge editions 2018-2024","key_numbers":[{"label":"AI-controlled compartments beating human reference (2nd challenge)","value":"5 of 5","unit":"teams","basis":"measured","sources":["src-hemming2020"]}],"best_for":["E1","E7"],"limits":"Labour for crop work (pruning, harvest) still manual in the challenges [src-hemming2019]"}]},{"id":"MIND-S7","name":"Island mode: graceful degradation and Earth-independence","function":"Keep the space alive, dry and safe for days to months with no network, no visits and partial hardware failure, then recover without human intervention.","how_it_works":"Loads are ranked into survival tiers (life safety, freeze/overheat protection, data, comfort, production); a supervisor sheds tiers as stored energy and forecast fall, and drives each subsystem into a dormant state that survives indefinitely. The PENGUIn Automatic Geophysical Observatories ran unattended at about 100 W year-round on the Antarctic plateau, with about four days of battery coast and a rule that any two of three sources (solar, wind, battery) may fail; components shut down for more than six years still worked when powered up. Halley VI now runs autonomously through winter while staff monitor it from Cambridge, UK. Deep-space habitats add communication delay and blackout: Gateway's Vehicle System Manager was designed for fully autonomous uncrewed operation with minimal ground monitoring, and NASA material cites up to 44 minutes between asking Mars a question and getting the answer. The design consequence is that every decision the habitat must make faster than the worst-case link round trip has to live on board.","options":[{"name":"Tiered load shedding + dormant safe state","principle":"Priority-ranked loads, forecast-driven shedding, freeze-safe karakuri positions","maturity":"COMMERCIAL","trl":8,"trl_why":"Standard in remote telecom/polar instruments","key_numbers":[{"label":"AGO continuous power level","value":"about 100","unit":"W","basis":"reported","sources":["src-melville2014"]},{"label":"AGO battery coast time (~50 W instrument load)","value":"about 4","unit":"days","basis":"reported","sources":["src-melville2014"]}],"best_for":["E3","E4","E5","E6","E8"],"limits":"Requires generous oversizing (factor 5 or more recommended) [src-melville2014]"},{"name":"Uncrewed-season station autonomy (Halley VI pattern)","principle":"Station runs autonomously in winter, monitored and adjusted remotely from home institution","maturity":"COMMERCIAL","trl":8,"trl_why":"Operational practice since the station became summer-only (BAS: runs autonomously in winter, operated remotely from Cambridge)","key_numbers":[{"label":"Halley VI relocation inland (2017)","value":23,"unit":"km","basis":"reported","sources":["src-bas-halley"]}],"best_for":["E4"],"limits":"Not a habitat for people in winter; relies on summer maintenance crews [src-bas-halley]"},{"name":"Hierarchical spacecraft autonomy (Gateway VSM, MAST)","principle":"Distributed subsystem autonomy coordinated by vehicle-level planner, executive and fault manager","maturity":"EXPERIMENTAL","trl":5,"trl_why":"Ground-tested software; Gateway paused March 2026","key_numbers":[{"label":"Mars question-answer delay (upper bound)","value":"up to 44","unit":"minutes","basis":"calculated","sources":["src-nasa-mars-delay"],"note":"Not found in cited NTRS PDF; equals ~2x max one-way light time (~22 min)"}],"best_for":["E8"],"limits":"No flight heritage for full vehicle-level autonomy [src-anderson2024, src-gateway-ifl]"},{"name":"Self-contained robotic observatory (PLATO)","principle":"6-tonne module providing own heat, electricity and satellite comms for a year","maturity":"PILOT","trl":8,"trl_why":"Deployed Dome A Jan 2008, two full years of performance reported","key_numbers":[],"best_for":["E4"],"limits":"Instrument platform, not a human habitat [src-plato2009, src-plato2010]"}]},{"id":"MIND-S8","name":"Cyber boundary","function":"Make sure the space cannot be held hostage, spied on or turned against its occupants, even though it is remotely operated by design.","how_it_works":"ISA/IEC 62443 organises security around roles (asset owner, product supplier, integrator, service provider), a secure product development lifecycle (4-1), component technical requirements (4-2), system requirements and security levels (3-3), and patch management (2-3). Practically this means segmenting the habitat into zones (safety controller, plant control, robots, guest/occupant network, cloud) joined by conduits with explicit allowed flows; the safety controller accepts only signed setpoint envelopes and never inbound sessions. Legacy BACnet and KNX installations have weak or no authentication, and attackers now use building systems as leverage: a 2025 report describes ransomware groups threatening hospital HVAC, and Johnson Controls itself suffered a ransomware incident in 2023. Teleoperation of robots inside homes adds a privacy channel that must be consented per session.","options":[{"name":"IEC 62443 zones and conduits with signed setpoint envelopes","principle":"Segmentation, least privilege, secure-by-design components, patch process","maturity":"COMMERCIAL","trl":9,"trl_why":"Published, certifiable standard series","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Certification cost for a small company; ongoing patch obligations [src-isa62443]"},{"name":"Outbound-only edge with local autonomy","principle":"Edge initiates encrypted sessions; no inbound ports; site keeps operating if cloud is cut","maturity":"COMMERCIAL","trl":9,"trl_why":"Common IIoT practice","key_numbers":[],"best_for":["E1","E2","E3","E4","E5","E6","E7","E8"],"limits":"Remote assistance latency depends on the edge polling the queue"}]}],"projects":[{"id":"P-dm-dc","name":"Google data-centre cooling AI","org":"Google DeepMind / Google","country":"UK/US","year_or_period":"2016-2018 onward","what":"Neural-network ensembles predicting PUE, temperature and pressure; first recommendations, then direct autonomous control every five minutes","status":"shipped","outcome_or_lesson":"Up to 40% cooling-energy reduction (15% PUE-overhead reduction) when ML recommendations were applied in 2016 live tests; about 30% average in autonomous mode after nine months (from 12% at launch). Lesson: savings grow with operating time; safety layers made autonomy acceptable to operators.","sources":["src-dm2016","src-dm2018"]},{"id":"P-dm-trane","name":"RL control of commercial cooling","org":"DeepMind, Google, Trane Technologies","country":"US","year_or_period":"2022","what":"Live RL control of chillers at two commercial facilities","status":"research","outcome_or_lesson":"9% and 13% savings: one third of the data-centre headline once moved to messier commercial buildings with constraint and sensing challenges.","sources":["src-luo2022"]},{"id":"P-brainbox","name":"BrainBox AI autonomous HVAC","org":"BrainBox AI (acquired by Trane Technologies)","country":"Canada","year_or_period":"2017-2026","what":"Cloud AI that writes setpoints to existing BMS","status":"active","outcome_or_lesson":"Claims up to 25% energy and up to 40% GHG reduction across 14,000+ buildings; acquisition agreed Dec 2024 (terms undisclosed), completed start of 2025 per BetaKit; May 2026 claims of 68,000+ buildings. Lesson: supervisory AI is consolidating into HVAC OEMs.","sources":["src-trane-bbai","src-trane-lab"]},{"id":"P-phaidra","name":"Phaidra industrial/data-centre AI control","org":"Phaidra (founded by former DeepMind data-centre team members, per press)","country":"US","year_or_period":"2019-2026","what":"Autonomous control agents for data centre cooling","status":"active","outcome_or_lesson":"Cited PRWeb funding release now returns HTTP 404 (unverified). Independent evidence of activity: US12326701B2 'Deterministic industrial process control' assigned to Phaidra Inc., granted 2025-06-10. Performance figures not verified.","sources":["src-phaidra-news","src-pat-us12326701"]},{"id":"P-cmu-house","name":"Residential RL vs MPC field comparison","org":"Carnegie Mellon University / Cornell","country":"US","year_or_period":"2024-2025","what":"Same occupied house controlled alternately by MPC and RL","status":"research","outcome_or_lesson":"MPC 18.1%, RL 20.9% heat-pump energy savings with overlapping CIs; RL needed about one-third less engineering but caused three discomfort reports.","sources":["src-mulayim2025"]},{"id":"P-boptest","name":"BOPTEST benchmark","org":"IBPSA Project 1/2 (LBNL, KU Leuven et al.)","country":"International","year_or_period":"2019-2026","what":"Standard virtual building test cases and API for comparing controllers","status":"active","outcome_or_lesson":"Lets MPC/RL/LLM controllers be compared on identical emulators before touching a real building; ORIJO should publish its module emulators the same way.","sources":["src-blum2025","src-arroyo2022","src-wang2023rlmpc"]},{"id":"P-brick","name":"Brick schema consortium","org":"UC Berkeley, UCSD, CMU, Johnson Controls and others","country":"US","year_or_period":"2016-2026","what":"Open ontology for building equipment and points; converging with ASHRAE 223P and Haystack","status":"active","outcome_or_lesson":"Field demonstration showed Brick can scale ASHRAE G36 sequences across buildings; adoption limited by modelling labour.","sources":["src-balaji2016","src-fierro2019","src-roa2023"]},{"id":"P-matter","name":"Matter 1.5","org":"Connectivity Standards Alliance","country":"International","year_or_period":"2022-2025","what":"Consumer smart-home standard; 1.5 adds tariffs, carbon intensity, metering, solar data, bidirectional EV charging, soil sensors","status":"shipped","outcome_or_lesson":"Energy-aware residential devices become vendor-neutral; useful for ORIJO consumer appliances, not for plant topology.","sources":["src-matter15"]},{"id":"P-openrmf","name":"Open-RMF","org":"Open Source Robotics Foundation / Intrinsic (origin: OSRC Singapore, acquired by Intrinsic Dec 2022; Singapore-government programme origin not verified in cited source)","country":"Singapore/US","year_or_period":"2021-2026","what":"Open-source multi-fleet traffic, door, lift and task coordination on ROS 2","status":"active","outcome_or_lesson":"Proves vendor-neutral robot coordination is feasible; YouTube demos show simulated doors/lifts, real hospital footage less visible.","sources":["src-openrmf-news","src-openrmf-yt","src-frota360"]},{"id":"P-waymo-ra","name":"Waymo remote assistance","org":"Waymo","country":"US/Philippines","year_or_period":"2018-2026","what":"Fleet response agents answer vehicle-initiated questions; no remote driving","status":"active","outcome_or_lesson":"70 agents on duty at any time for about 3,000 vehicles (~1:43 calculated; Futurism says ~1:41), about half in the Philippines. Lesson: L4 autonomy still budgets human minutes, and the public cares where those humans sit.","sources":["src-waymo-futurism"]},{"id":"P-1x-neo","name":"1X NEO home humanoid","org":"1X Technologies","country":"Norway/US","year_or_period":"2025-2026","what":"Home humanoid sold with scheduled remote teleoperation to complete chores and gather training data","status":"pilot","outcome_or_lesson":"$20,000 or $499/month; owner approval per teleop session, people blurring, no-go zones. Lesson: teleop-first is the honest go-to-market for manipulation, but privacy is the product.","sources":["src-1x-engadget","src-1x-robotreport","src-1x-yt"]},{"id":"P-imo-mass","name":"IMO MASS Code","org":"International Maritime Organization MSC 111","country":"International","year_or_period":"2017-2032","what":"First global code for autonomous ships; ROC-centred, master retains responsibility","status":"active","outcome_or_lesson":"Voluntary from 1 July 2026, mandatory code expected to be adopted by 1 Jul 2030 and enter into force 1 Jan 2032 (target). A ready-made regulatory template for 'remotely supervised autonomous habitats'.","sources":["src-imo-mass"]},{"id":"P-munin","name":"MUNIN unmanned bulk carrier","org":"EU FP7 consortium (Fraunhofer CML, Chalmers et al.)","country":"EU","year_or_period":"2012-2015","what":"Concept and simulation of a 200 m bulk carrier unmanned during ocean passage with a shore control centre","status":"closed","outcome_or_lesson":"Shore-centre situation awareness and handover identified as the critical human-factors problem.","sources":["src-munin2013","src-porathe2014"]},{"id":"P-gateway","name":"Lunar Gateway Vehicle System Manager","org":"NASA","country":"US","year_or_period":"2018-2026","what":"Hierarchical autonomy for a station uncrewed most of the year","status":"paused","outcome_or_lesson":"Gateway paused/deferred 'in its current form' in March 2026 (not formally cancelled; hardware redirected to surface goals) in favour of a ~$20B surface base; autonomy software and TLA+ contracts remain reusable knowledge.","sources":["src-anderson2024","src-dabney2023","src-gateway-ifl","src-gateway-req"]},{"id":"P-iss-timeliner","name":"ISS Timeliner / HAL automated commanding","org":"NASA / Draper","country":"US","year_or_period":"2005 onward","what":"Onboard automated command sequencing for ISS payloads and systems","status":"shipped","outcome_or_lesson":"Moving routine ground-operator functions onboard was proven on ISS from April 2005.","sources":["src-stetson2011"]},{"id":"P-ago","name":"PENGUIn Automatic Geophysical Observatories","org":"US Antarctic Program (Virginia Tech, UC Berkeley et al.)","country":"Antarctica","year_or_period":"1996-2014+ (five stations operational in 2014)","what":"Unmanned ~100 W year-round observatories on the plateau","status":"shipped","outcome_or_lesson":"Propane thermoelectric generators' exhaust ducts froze and propane will not vaporise below about -40 C (abandoned 2001); shunt regulators failed shorted; authors recommend oversizing by 5x or more and surviving any two of three source failures.","sources":["src-melville2014"]},{"id":"P-plato","name":"PLATO robotic observatory","org":"UNSW / Polar Research Institute of China","country":"Antarctica (Dome A)","year_or_period":"2008-2010+","what":"6-tonne self-contained module with own heat, power and satellite comms","status":"shipped","outcome_or_lesson":"Two full years of autonomous performance reported; the closest existing analogue to an ORIJO KITA module.","sources":["src-plato2009","src-plato2010"]},{"id":"P-aastino","name":"AASTINO","org":"UNSW","country":"Antarctica (Dome C)","year_or_period":"2003-","what":"Automated site-testing observatory designed to run through winter without intervention","status":"closed","outcome_or_lesson":"Hybrid power plus automation can keep instruments alive through polar night.","sources":["src-aastino"]},{"id":"P-halley","name":"Halley VI winter autonomy","org":"British Antarctic Survey","country":"UK / Antarctica","year_or_period":"2017-2026","what":"Relocated 23 km inland in 2017; now summer-only, runs autonomously in winter monitored from Cambridge","status":"active","outcome_or_lesson":"Geohazards (ice cracks) turned a crewed station into an autonomous one: autonomy is also a hedge against site risk.","sources":["src-bas-halley"]},{"id":"P-agc","name":"Autonomous Greenhouse Challenge","org":"Wageningen University & Research (with Tencent sponsorship in early editions)","country":"Netherlands","year_or_period":"2018-2024","what":"International teams grow crops remotely with AI setpoints in 96 m2 compartments","status":"active","outcome_or_lesson":"Edition 1: one AI team beat the human reference; edition 2: all five did; edition 4: six algorithms all profitable on dwarf tomato.","sources":["src-hemming2019","src-hemming2020","src-maree2025","src-agc-yt"],"status_note":"Four editions 2018-2024 verified via papers; whether a 5th edition is running in 2026 is not verified"},{"id":"P-freightfarms","name":"Freight Farms container farms","org":"Freight Farms","country":"US","year_or_period":"2010-2025","what":"Container farms controlled through the Farmhand cloud platform","status":"failed","outcome_or_lesson":"Ceased operations and filed Chapter 7 on 30 Apr 2025 with ~$7.0M liabilities vs ~$0.59M assets; Farmhand cloud phased out, remote access and data history lost; growers reverse-engineered controls.","sources":["src-freightfarms","src-freightfarms-ch7"]},{"id":"P-plenty","name":"Plenty vertical farms","org":"Plenty Unlimited","country":"US","year_or_period":"2014-2025","what":"Highly automated vertical farms","status":"restructured","outcome_or_lesson":"Chapter 11 on 23 Mar 2025 after ~$961M raised; Compton farm halted Dec 2024; emerged from Chapter 11 around end of May 2025 as a smaller strawberry-only operator (Richmond, VA). Automation did not fix leafy-green unit economics; post-2025 commercial performance unverified.","sources":["src-plenty","src-plenty-emerge"]},{"id":"P-aerofarms","name":"AeroFarms","org":"AeroFarms","country":"US","year_or_period":"2004-2025","what":"Aeroponic vertical farming","status":"restructured","outcome_or_lesson":"Went through Chapter 11 (2023) and re-emerged smaller; Fast Company May 2025 headline describes a turnaround (article body returned 403, so details unverified).","sources":["src-aerofarms-news"]},{"id":"P-jci","name":"Johnson Controls ransomware incident","org":"Johnson Controls International","country":"US/Ireland","year_or_period":"2023-2024","what":"Ransomware encrypted internal servers of a major BMS vendor","status":"closed","outcome_or_lesson":"Customer environments reportedly not affected (expert comment in Cybersecurity Dive, Sep 2023). The often-cited >$27M cleanup cost could not be verified (Dark Reading page 403; no other primary source fetched) - unverified.","sources":["src-jci2023","src-jci-27m"]},{"id":"P-fanuc","name":"FANUC highly robotised factories","org":"FANUC","country":"Japan","year_or_period":"2000s-2026","what":"Highly automated plants with 'long hours of continuous unmanned machining'","status":"active","outcome_or_lesson":"Widely repeated claims of robots running unattended for weeks are not confirmed on FANUC's own page; only 'long hours' is stated.","sources":["src-fanuc"]},{"id":"P-hiseas","name":"HI-SEAS / BASALT analog communications","org":"University of Hawaii / NASA","country":"US","year_or_period":"2013-2019","what":"Mars analog missions with delayed communication between crew and mission support","status":"research","outcome_or_lesson":"Analogs validated delay-tolerant file exchange and crew autonomy practices for deep-space comms.","sources":["src-seibert2019"]},{"id":"P-mast","name":"NASA MAST framework","org":"NASA JSC","country":"US","year_or_period":"2019","what":"Component-based distributed hierarchical autonomy framework for spacecraft subsystems","status":"research","outcome_or_lesson":"Each subsystem keeps local autonomy; vehicle-level coordinator respects their independence: the right pattern for module-based habitats.","sources":["src-badger2019"]},{"id":"P-growy","name":"Growy","org":"Growy","country":"Singapore","year_or_period":"2024-2025","what":"Automated vertical farm","status":"failed","outcome_or_lesson":"Winding up within a year of official launch (headline-level evidence).","sources":["src-growy"]},{"id":"P-vf","name":"Vertical Future","org":"Vertical Future","country":"UK","year_or_period":"2016-2025","what":"Automated vertical farming technology","status":"failed","outcome_or_lesson":"Put up for sale on an insolvency market after pre-tax losses above GBP 10M (~USD 13.2M) (AgFunderNews, 31 Jul 2025).","sources":["src-vertical-future"]}],"papers":[{"id":"PA-luo2022","title":"Controlling Commercial Cooling Systems Using Reinforcement Learning","authors":"Luo, Paduraru, Voicu et al. (36 authors)","year":2022,"venue":"arXiv 2211.07357","doi_or_url":"https://arxiv.org/abs/2211.07357","finding":"RL controlling real commercial chiller plants saved about 9% and 13% energy at two live sites.","type":"field-trial"},{"id":"PA-mulayim2025","title":"Comparative Field Deployment of Reinforcement Learning and Model Predictive Control for Residential HVAC","authors":"Mulayim, Pergantis, Reyes Premer, Chen, Qu, Kircher, Berges","year":2025,"venue":"arXiv 2510.01475","doi_or_url":"https://arxiv.org/abs/2510.01475","finding":"MPC saved 18.1% and RL 20.9% weather-normalised heat-pump energy vs constant setpoint; RL needs about one-third less engineering effort.","type":"field-trial"},{"id":"PA-khabbazi2025","title":"Lessons learned from field demonstrations of MPC and RL for residential and commercial HVAC: A review","authors":"Khabbazi, Pergantis, Reyes Premer et al.","year":2025,"venue":"Applied Energy 399:126459","doi_or_url":"https://doi.org/10.1016/j.apenergy.2025.126459","finding":"Of 104 field demonstrations, 71% used protocols that may give unreliable estimates; reliable ones average 16% (residential) and 13% (commercial) cost savings; only 13 report deployment costs.","type":"review"},{"id":"PA-liu2022","title":"Safe HVAC Control via Batch Reinforcement Learning","authors":"Liu, Balaji, Gao et al.","year":2022,"venue":"ICCPS 2022","doi_or_url":"https://doi.org/10.1109/iccps54341.2022.00023","finding":"KL-regularised batch RL starting from the rule-based policy beats existing control from day one of deployment in a real multi-zone building (savings number not captured in abstract).","type":"field-trial"},{"id":"PA-bengea2014","title":"Implementation of model predictive control for an HVAC system in a mid-size commercial building","authors":"Bengea, Kelman, Borrelli et al.","year":2014,"venue":"HVAC&R Research","doi_or_url":"https://doi.org/10.1080/10789669.2013.834781","finding":"Landmark full-scale MPC proof-of-concept on a multi-zone dual-duct unit (142 citations); savings figure not verified here.","type":"field-trial"},{"id":"PA-zhang2022","title":"Model Predictive Control for Demand Flexibility: Real-world Operation of a Commercial Building with Photovoltaic and Battery Systems","authors":"Zhang, Prakash, Paul et al.","year":2022,"venue":"Advances in Applied Energy","doi_or_url":"https://doi.org/10.1016/j.adapen.2022.100099","finding":"Real-world MPC coordinated HVAC, PV and battery for demand flexibility (107 citations); quantitative results not verified here.","type":"field-trial"},{"id":"PA-wang2023","title":"Comparison of reinforcement learning and model predictive control for building energy system optimization","authors":"Wang, Zheng, Wang et al.","year":2023,"venue":"Applied Thermal Engineering","doi_or_url":"https://www.sciencedirect.com/science/article/pii/S1359431123004593","finding":"Head-to-head RL vs MPC on the BOPTEST hydronic heat pump case (168 citations); numbers not verified here.","type":"model"},{"id":"PA-katipamula2005","title":"Methods for Fault Detection, Diagnostics, and Prognostics for Building Systems - A Review, Part I","authors":"Katipamula, Brambley","year":2005,"venue":"HVAC&R Research 11(1)","doi_or_url":"https://doi.org/10.1080/10789669.2005.10391123","finding":"Degraded and improperly controlled equipment wastes an estimated 15-30% of energy in commercial buildings (1,024 citations).","type":"review"},{"id":"PA-mirnaghi2020","title":"Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: A comprehensive review","authors":"Mirnaghi, Haghighat","year":2020,"venue":"Energy and Buildings 229:110492","doi_or_url":"https://doi.org/10.1016/J.ENBUILD.2020.110492","finding":"Classifies data-driven FDD into supervised, unsupervised and hybrid, recommending hybrids; 360 citations.","type":"review"},{"id":"PA-chen2023","title":"A review of data-driven fault detection and diagnostics for building HVAC systems","authors":"Chen, O'Neill, Wen et al.","year":2023,"venue":"Applied Energy 339:121030","doi_or_url":"https://doi.org/10.1016/j.apenergy.2023.121030","finding":"Recent comprehensive data-driven FDD review (299 citations); identifies data and transferability gaps.","type":"review"},{"id":"PA-cespedes2024","title":"A review of building digital twins to improve energy efficiency in the building operational stage","authors":"Cespedes-Cubides, Jradi","year":2024,"venue":"Energy Informatics","doi_or_url":"https://doi.org/10.1186/s42162-024-00313-7","finding":"Systematic selection of 21 DT use cases plus 6 enterprise solutions for O&M energy efficiency.","type":"review"},{"id":"PA-balaji2016","title":"Brick: Towards a Unified Metadata Schema For Buildings","authors":"Balaji, Bhattacharya, Fierro et al.","year":2016,"venue":"ACM BuildSys","doi_or_url":"https://doi.org/10.1145/2993422.2993577","finding":"Introduced the Brick ontology for portable building applications (375 citations).","type":"standard"},{"id":"PA-fierro2019","title":"Beyond a House of Sticks: Formalizing Metadata Tags with Brick","authors":"Fierro, Koh, Agarwal et al.","year":2019,"venue":"ACM BuildSys","doi_or_url":"https://doi.org/10.1145/3360322.3360862","finding":"Shows Project Haystack's informal tags have inherent interpretability and consistency issues; proposes formally defined Brick+.","type":"model"},{"id":"PA-hemming2019","title":"Remote Control of Greenhouse Vegetable Production with Artificial Intelligence","authors":"Hemming, de Zwart, Elings et al.","year":2019,"venue":"Sensors 19(8):1807","doi_or_url":"https://doi.org/10.3390/s19081807","finding":"Five teams each remotely ran a 96 m2 cucumber compartment for 4 months; one AI team outperformed the manual reference.","type":"field-trial"},{"id":"PA-hemming2020","title":"Cherry Tomato Production in Intelligent Greenhouses - Sensors and AI for Control","authors":"Hemming, de Zwart, Elings et al.","year":2020,"venue":"Sensors 20(22):6430","doi_or_url":"https://doi.org/10.3390/s20226430","finding":"Over 6 months of cherry tomato in six compartments, all five AI-supported teams outperformed the human-operated reference (net-profit objective).","type":"field-trial"},{"id":"PA-maree2025","title":"Autonomous Greenhouse Cultivation of Dwarf Tomato: Performance Evaluation of Intelligent Algorithms for Multiple-Sensor Feedback","authors":"Maree, Zhang et al.","year":2025,"venue":"Sensors 25(14):4321","doi_or_url":"https://doi.org/10.3390/s25144321","finding":"In the 4th challenge, all six intelligent algorithms that fully controlled dwarf tomato achieved a profitable strategy; best reached ~45 kg/m2/year (reported), above commercial high-wire cherry tomato.","type":"field-trial"},{"id":"PA-melville2014","title":"Sustainable energy at the 100 W level for scientific sites on the Antarctic Plateau: lessons learned from the PENGUIn-AGO project","authors":"Melville, Stillinger, Gerrard, Weatherwax","year":2014,"venue":"Review of Scientific Instruments 85:045117","doi_or_url":"https://doi.org/10.1063/1.4871555","finding":"About 100 W year-round unattended (solar, wind, batteries); ~4 days battery coast for the ~50 W instrument load; recommends oversizing by 5x or more and tolerating loss of any two of three sources.","type":"field-trial"},{"id":"PA-plato2010","title":"Performance of the autonomous PLATO Antarctic Observatory over two full years","authors":"Luong-Van, Ashley et al.","year":2010,"venue":"Proc. SPIE","doi_or_url":"https://doi.org/10.1117/12.857910","finding":"A self-contained robotic observatory at Dome A operated autonomously across two full years (detailed uptime not verified here).","type":"field-trial"},{"id":"PA-anderson2024","title":"Gateway Autonomy for Enabling Deep Space Exploration","authors":"Anderson, Badger","year":2024,"venue":"IEEE Aerospace Conference","doi_or_url":"https://doi.org/10.1109/AERO58975.2024.10521067","finding":"Autonomy needed to handle long communication outages and reduce sustaining operations costs via the Vehicle Systems Manager.","type":"model"},{"id":"PA-dabney2023","title":"Trustworthy Autonomy for Gateway Vehicle System Manager","authors":"Dabney, Badger, Rajagopal","year":2023,"venue":"IEEE SCC","doi_or_url":"https://doi.org/10.1109/SCC57168.2023.00018","finding":"VSM spans 4 function categories and levels from fully autonomous uncrewed operation to advisory automation, verified with TLA+ assume-guarantee contracts.","type":"model"},{"id":"PA-badger2019","title":"A Distributed Hierarchical Framework for Autonomous Spacecraft Control","authors":"Badger, Strawser","year":2019,"venue":"IEEE Aerospace Conference","doi_or_url":"https://doi.org/10.1109/AERO.2019.8742199","finding":"MAST framework: component-based, distributed hierarchical autonomy that respects subsystem independence.","type":"model"},{"id":"PA-baker2025","title":"Spacecraft Autonomy Levels","authors":"Baker, Phillips","year":2025,"venue":"arXiv 2503.01049","doi_or_url":"https://arxiv.org/abs/2503.01049","finding":"Proposes 6 levels of spacecraft autonomy deliberately parallel to automotive and aircraft levels.","type":"standard"},{"id":"PA-alfus2005","title":"Autonomy Levels for Unmanned Systems (ALFUS) framework: an update","authors":"Huang, Pavek et al.","year":2005,"venue":"Proc. SPIE","doi_or_url":"https://doi.org/10.1117/12.603725","finding":"Autonomy measured on three axes (mission complexity, environmental difficulty, human independence) rather than one ladder (134 citations).","type":"standard"},{"id":"PA-porathe2014","title":"Situation awareness in remote control centres for unmanned ships","authors":"Porathe, Prison, Man","year":2014,"venue":"RINA Human Factors in Ship Design","doi_or_url":"https://scholar.google.com/scholar?q=%22Situation+awareness+in+remote+control+centres+for+unmanned+ships%22","finding":"Shore-based operators of unmanned ships lose situation awareness; handover design is critical (138 citations).","type":"experiment"},{"id":"PA-cheng2023","title":"A novel system-theoretic approach for human-system collaboration safety: two degrees of autonomy for autonomous ships","authors":"Cheng, Utne, Wu et al.","year":2023,"venue":"Reliability Engineering & System Safety","doi_or_url":"https://doi.org/10.1016/j.ress.2023.109388","finding":"STPA-style analysis of human-system collaboration across two IMO degrees of autonomy (63 citations).","type":"model"},{"id":"PA-sikand2021","title":"Robofleet: Open Source Communication and Management for Fleets of Autonomous Robots","authors":"Sikand, Zartman, Rabiee et al.","year":2021,"venue":"IROS","doi_or_url":"https://doi.org/10.1109/IROS51168.2021.9635830","finding":"Lightweight secure multi-robot/multi-user communication robust to intermittent low-bandwidth links.","type":"experiment"},{"id":"PA-jonah2026","title":"LLMs for Agentic Home Energy Management","authors":"Jonah, Moses et al.","year":2026,"venue":"arXiv 2607.04569","doi_or_url":"https://doi.org/10.48550/arXiv.2607.04569","finding":"Five LLM backends scheduling appliances vs a MILP benchmark over a 7-day rolling test: near-optimal on ordinary days but fabricated schedules and failed commitments under constraint conflicts.","type":"experiment"},{"id":"PA-granzer2010","title":"Security in Building Automation Systems","authors":"Granzer, Praus, Kastner","year":2010,"venue":"IEEE Trans. Industrial Electronics","doi_or_url":"https://doi.org/10.1109/TIE.2009.2036033","finding":"Building automation protocols originated when security was a side issue; analyses BACnet/KNX/LonWorks protection gaps (167 citations).","type":"review"},{"id":"PA-seibert2019","title":"Developing Future Deep-Space Telecommunication Architectures: A Historical Look at the Benefits of Analog Research on the Development of Solar System Internetworking for Future Human Spaceflight","authors":"Seibert, Lim, Miller et al.","year":2019,"venue":"Astrobiology","doi_or_url":"https://doi.org/10.1089/ast.2018.1915","finding":"BASALT and HI-SEAS analogs validated delay-tolerant communication architectures for Mars operations.","type":"field-trial"},{"id":"PA-stetson2011","title":"The HAL 9000 Space Operating System","authors":"Stetson, Knickerbocker et al.","year":2011,"venue":"IEEE Aerospace Conference","doi_or_url":"https://doi.org/10.1109/AERO.2011.5747602","finding":"Timeliner-based automated commanding moved ground-operator functions onboard ISS from April 2005.","type":"field-trial"},{"id":"PA-arroyo2022","title":"Comparison of optimal control techniques for building energy management","authors":"Arroyo, Spiessens, Helsen","year":2022,"venue":"Frontiers in Built Environment","doi_or_url":"https://doi.org/10.3389/fbuil.2022.849754","finding":"Benchmarks MPC, RL and hybrids on BOPTEST in a unified framework (52 citations).","type":"model"}],"numbers":[{"id":"N01","label":"Data-centre cooling energy reduction from ML recommendations","value":"up to 40","unit":"%","basis":"reported","context":"Google 2016; equals 15% reduction of PUE overhead","sources":["src-dm2016"]},{"id":"N02","label":"PUE overhead reduction","value":15,"unit":"%","basis":"reported","context":"Same 2016 deployment","sources":["src-dm2016"]},{"id":"N03","label":"Autonomous data-centre cooling savings after nine months","value":"about 30 (about 12 at launch)","unit":"%","basis":"reported","context":"Direct AI control with two-layer verification, 2018","sources":["src-dm2018"]},{"id":"N04","label":"AI decision cycle in autonomous cooling","value":5,"unit":"min","basis":"reported","context":"'Every five minutes' cloud AI takes snapshot and sends actions; operator constraints checked in cloud and again on site","sources":["src-dm2018"]},{"id":"N05","label":"RL savings at two commercial cooling sites","value":"9 and 13","unit":"%","basis":"measured","context":"DeepMind/Google/Trane live experiments","sources":["src-luo2022"]},{"id":"N06","label":"MPC / RL heat-pump energy savings in one occupied house","value":"18.1 / 20.9","unit":"%","basis":"measured","context":"95% CIs 4.4-30.9 and 2.6-38.3; overlap","sources":["src-mulayim2025"]},{"id":"N07","label":"RL engineering effort relative to MPC for a fresh house","value":"about 2/3","unit":"ratio","basis":"reported","context":"Authors' estimate","sources":["src-mulayim2025"]},{"id":"N08","label":"Field demos with possibly unreliable protocols","value":71,"unit":"% of 104","basis":"measured","context":"24 residential + 80 commercial papers","sources":["src-khabbazi2025"]},{"id":"N09","label":"Duration-weighted cost savings in reliable field demos","value":"16 residential / 13 commercial","unit":"%","basis":"measured","context":"MPC and RL pooled","sources":["src-khabbazi2025"]},{"id":"N10","label":"Field-demo papers reporting deployment/O&M cost","value":"13 of 104","unit":"papers","basis":"measured","context":"The business-case gap","sources":["src-khabbazi2025"]},{"id":"N11","label":"Energy wasted by degraded/poorly controlled commercial equipment","value":"15-30","unit":"%","basis":"reported","context":"Landmark FDD review estimate","sources":["src-katipamula2005"]},{"id":"N12","label":"BrainBox AI claimed energy reduction / deployed buildings","value":"up to 25 / 14,000+","unit":"% / buildings","basis":"claimed","context":"Figures stated in Facilities Dive at agreement announcement, 20 Dec 2024 (deal closed early 2025 per BetaKit). By May 2026 BetaKit reports BrainBox claiming a network of 68,000+ buildings; all company claims, not independently verified","sources":["src-trane-bbai","src-trane-lab"]},{"id":"N13","label":"Remote-assistance agents per vehicle (Waymo)","value":"70 agents on duty at any time / about 3,000 vehicles (calculated ~1:43; Futurism states ~1:41)","unit":"agents : vehicles","basis":"reported (ratio calculated)","context":"Feb 2026; four operations centres (Arizona, Michigan, two in the Philippines); agents advise, Waymo states it does not rely on remote drivers. 'Redundant' not stated in source","sources":["src-waymo-futurism"]},{"id":"N14","label":"1X NEO price / subscription","value":"20,000 / 499","unit":"USD / USD per month","basis":"reported","context":"Teleoperation scheduled by owner","sources":["src-1x-engadget"]},{"id":"N15","label":"IMO MASS Code voluntary in force / mandatory expected","value":"2026-07-01 / 2032-01-01","unit":"date","basis":"reported (mandatory date is a target)","context":"Adopted at MSC 111 (13-22 May 2026). Roadmap: mandatory MASS Code adoption expected by 1 Jul 2030, entry into force 1 Jan 2032","sources":["src-imo-mass"]},{"id":"N16","label":"Unattended polar observatory power","value":"about 100","unit":"W","basis":"reported","context":"PENGUIn AGO, operating continuously since 1996 (five stations operational in 2014); solar + wind + batteries","sources":["src-melville2014"]},{"id":"N17","label":"AGO battery coast time","value":"about 4","unit":"days","basis":"reported","context":"About 4 days for the ~50 W instrument load (not the full ~100 W) on new batteries; lead-acid carry-over about 2 days in wind outage","sources":["src-melville2014"]},{"id":"N18","label":"Recommended oversizing for unattended polar systems","value":">=5","unit":"x","basis":"reported","context":"Balance logistics against failure costs","sources":["src-melville2014"]},{"id":"N19","label":"Dormant survival of shut-down equipment","value":">6","unit":"years","basis":"reported","context":"AGO4 science equipment unused >6 years, worked on power-up in 2010-11 season","sources":["src-melville2014"]},{"id":"N20","label":"Mars question-to-answer delay (worst case)","value":"up to 44","unit":"min","basis":"calculated","context":"NASA NTRS 20180007474 PDF could not be text-extracted; the 44-min figure was not found in the source. Physically consistent: max Earth-Mars distance ~2.67 AU gives ~22 min one-way light time, so ~44 min round trip (calculated, not source-verified)","sources":["src-nasa-mars-delay"],"verification":"unverified in cited source; retained as calculated light-time"},{"id":"N21","label":"Autonomous Greenhouse Challenge compartment / duration (ed. 1)","value":"96 / 4","unit":"m2 / months","basis":"measured","context":"Five teams, cucumber","sources":["src-hemming2019"]},{"id":"N22","label":"AI teams beating human reference (ed. 2, cherry tomato, 6 months)","value":"5 of 5","unit":"teams","basis":"measured","context":"Second Autonomous Greenhouse Challenge, six compartments, six months, net-profit objective; abstract: all AI-supported teams outperformed the human-operated reference","sources":["src-hemming2020"]},{"id":"N23","label":"Freight Farms liabilities / assets at Chapter 7","value":"7.01 / 0.594","unit":"million USD","basis":"reported","context":"Chapter 7 filed 30 Apr 2025 (D. Mass. 1:25-bk-10908); $7,010,826 liabilities vs $593,916 assets per Official Form 206Sum; Farmhand cloud phased out","sources":["src-freightfarms","src-freightfarms-ch7"]},{"id":"N24","label":"Plenty capital raised before Chapter 11","value":"about 961","unit":"million USD","basis":"reported","context":"Chapter 11 filed 23 Mar 2025 (S.D. Texas); about $961M over eight rounds per Bondoro. Emerged from Chapter 11 about 29 May 2025 (Bloomberg Law headline via search; plenty.ag shows 'completes restructuring, emerges from chapter 11') and now focuses on Richmond VA strawberries","sources":["src-plenty","src-plenty-emerge"]},{"id":"N25","label":"NASA surface base budget after pausing Gateway","value":"about 20","unit":"billion USD","basis":"reported","context":"IFLScience 24 Mar 2026 (amended 25 Mar): Gateway paused 'in its current form' / deferred, not cancelled; phases 2028-2030, 2031-2033, 2033-2036","sources":["src-gateway-ifl"]},{"id":"N26","label":"Halley VI former winter crew","value":16,"unit":"people","basis":"reported","context":"BAS: 16 people formerly wintered; now summer-only (~40 summer staff), runs autonomously in winter, operated remotely from Cambridge","sources":["src-bas-halley"]},{"id":"NG-14","label":"EU Machinery Regulation 2023/1230 mandatory application","value":"2027-01-20","unit":"date","basis":"reported","context":"Adds provisions for AI-powered safety functions and cyber-safety of safety control software; replaces Directive 2006/42/EC","sources":["src-ec-machinery"]}],"failures":[{"what":"Freight Farms (2025): container farms controlled through the Farmhand cloud","why":"Company went into Chapter 7; cloud platform phased out, remote access and production history lost, parts and support discontinued.","lesson_for_anyground":"Every ORIJO module must run its full control stack locally, export its data in open formats (Brick + CSV/Parquet), and carry an escrowed firmware/source licence that activates on insolvency. Cloud is advisory only."},{"what":"Plenty (2025), AeroFarms (2023), Growy (2025), Vertical Future (2025)","why":"High capex/opex vertical farms; automation did not fix unit economics or energy cost. Plenty and AeroFarms survived only by restructuring through Chapter 11 into much smaller operations; Growy wound up; Vertical Future put up for sale.","lesson_for_anyground":"Autonomy is sold on labour minutes saved per kg and per guest-night, not as a technology story; the mind layer must expose that metric from day one."},{"what":"Headline-to-field shrinkage in AI HVAC control","why":"40% data-centre cooling headline became 9-13% in commercial buildings; 71% of published field demos use protocols that may be unreliable.","lesson_for_anyground":"Promise 10-20% and measure it with switch-back (A/B by day) protocols against a rule-based baseline; never quote data-centre numbers for homes."},{"what":"RL comfort complaints during adaptation (residential field trial)","why":"RL explored cooler setpoints early; three occupant discomfort reports.","lesson_for_anyground":"Warm-start from rules (batch RL with KL penalty), hard comfort bands in the safety kernel, and shadow mode before control."},{"what":"PENGUIn AGO propane generators and shunt regulators","why":"Exhaust ducts froze, propane would not vaporise below about -40 C; regulators failed shorted; screw terminals loosened by thermal cycling.","lesson_for_anyground":"Polar ORIJO (E4) uses no combustion in the critical path, oversizes 5x, tolerates loss of any two energy sources, and uses spring-clamp terminals."},{"what":"Johnson Controls ransomware (2023)","why":"A major BMS vendor's own IT (incl. ESXi servers) was encrypted; customer environments reportedly unaffected. A cleanup cost above $27M is widely reported but unverified here.","lesson_for_anyground":"Vendor cloud is part of the attack surface; ORIJO sites must keep operating with all vendor clouds cut, and the safety kernel accepts only signed envelopes."},{"what":"NASA Lunar Gateway paused (March 2026)","why":"Programme-level budget and architecture change despite mature autonomy work.","lesson_for_anyground":"Do not tie the autonomy stack to one flagship customer or programme; build it as reusable modules (MAST-style) that move between typologies."},{"what":"Waymo remote-assistance controversy (Feb 2026)","why":"Senator Markey's inquiry and public concern (Feb 2026) over remote-assistance staffing, about half located in the Philippines.","lesson_for_anyground":"Publish the human-minutes ledger and ROC locations per site; guests and residents must know when a human can see or act."},{"what":"LLM home energy agents fabricate schedules","why":"Reasoning-to-action failures and failed commitments under constraint conflicts.","lesson_for_anyground":"LLMs never write actuators directly; their output is a proposed constraint set that a deterministic solver and the safety kernel must accept."}],"interfaces":[{"with_layer":"energy","exchange":"data (PV forecast, battery SoC, tariff, carbon intensity), kWh dispatch commands, load-shed tiers","design_rule":"Energy controller owns the battery; mind sends a 24-48 h dispatch plan and survival tier; any plan violating minimum reserve for 4 island days is rejected locally."},{"with_layer":"water","exchange":"L (tank levels, flows), quality sensor data, valve commands, heat from recycler","design_rule":"Potable-water release requires an independent quality interlock in the water module; mind can schedule but never override a quality trip."},{"with_layer":"grow","exchange":"crop climate setpoints, irrigation/nutrient recipes, images, labour tasks for robots","design_rule":"Use the Autonomous Greenhouse Challenge pattern: setpoint API between remote/AI grower and local process computer, with the process computer keeping crop-safe limits."},{"with_layer":"build","exchange":"fold state, hinge lock status, deployment sequence commands, structural health data","design_rule":"Fold/unfold motions are permitted only in a commissioning mode with a human present or on camera and with wind below the module limit; never by an autonomous optimiser."},{"with_layer":"structure","exchange":"strain, tilt, settlement (ASHI legs), snow/wind loads","design_rule":"Structural alarms map to evacuation-grade notifications and freeze all non-essential robot motion."},{"with_layer":"care","exchange":"occupant/guest context, privacy consent, service requests, teleoperation sessions","design_rule":"Any remote human view inside a private space requires per-session consent, blurring and no-go zones (1X pattern); logs are visible to occupants."},{"with_layer":"loop","exchange":"nutrient and waste stream states (urine, larvae digester, compost), labour","design_rule":"Loop processes run on long time constants; mind schedules them for energy surplus windows and flags drift, but biological safety limits are enforced in the loop module."},{"with_layer":"mind","exchange":"fleet telemetry, model updates, remote-assistance requests between sites and ROC","design_rule":"Site-to-cloud is outbound-only and advisory; model updates are staged (shadow, then 10% of fleet, then all) and reversible."}],"autonomy_ladder":[{"level":"L0","description":"Manual. People operate every system; no remote visibility. Island operation 0 days without a person on site. MTBI not defined.","human_minutes_per_week":">600 per site","enabling_tech":"Thermostats, manual valves, paper logs"},{"level":"L1","description":"Assisted. Local rules and alarms; remote monitoring dashboard. Needs a person within 1 day for most faults. MTBI <1 day; island <1 day.","human_minutes_per_week":"300-600","enabling_tech":"BACnet/Modbus/Matter devices, rule-based sequences, SMS alarms"},{"level":"L2","description":"Supervised automation. Optimiser (MPC/RL in shadow or advisory) sets schedules; remote human approves changes daily; robots do scripted tasks. MTBI >=1 day; island >=3 days with comfort maintained.","human_minutes_per_week":"60-300","enabling_tech":"Semantic model (Brick), MPC, rule FDD, single-vendor robots"},{"level":"L3","description":"Conditional autonomy. System controls within a verified envelope and asks the ROC for advice when confidence is low (Waymo pattern); FDD isolates faults; multi-fleet robots coordinate. Human on call, response <=4 h. MTBI >=7 days; island >=14 days.","human_minutes_per_week":"15-60","enabling_tech":"Safety kernel + supervised RL/MPC, data-driven FDD, Open-RMF, remote-assistance queue"},{"level":"L4","description":"High autonomy. Self-heals by redundancy and module swap by robot; humans visit only for scheduled renewal (shikinen) or swap of a failed sealed module. MTBI >=30 days; island >=90 days (Halley-winter class).","human_minutes_per_week":"1-15","enabling_tech":"Calibrated twin, fleet-wide anomaly baselines, crease-rail service robots, tiered load shedding"},{"level":"L5","description":"Earth-independent. No routine dependence on outside humans or networks; all decisions faster than the worst link round trip are onboard. MTBI >=180 days; island >=365 days (PLATO/AGO class, Mars conjunction).","human_minutes_per_week":"<1 (remote), crew self-sufficient if present","enabling_tech":"Hierarchical autonomy (MAST/VSM), formal runtime contracts, 5x oversizing, onboard planning"}],"orijo_design":{"recommended_stack":[{"subsystem":"MIND-S1","choice":"Modbus/CAN/BACnet devices -> MQTT Sparkplug B on the KURA-3 edge; Matter 1.5 only for consumer appliances; Brick + 223P graph generated from the configurator and stored in each module's fold passport","why":"Every ORIJO module is a catalogue item, so the semantic model can be auto-generated, removing the main adoption barrier of Brick."},{"subsystem":"MIND-S2","choice":"Rules (G36-style) as baseline; grey-box MPC for heat/battery/water tank; batch RL warm-started from rule logs for crop climate and HVAC fine-tuning; LLM only as intent-to-constraint translator and ops copilot","why":"Field evidence supports 13-21% savings for MPC/RL; LLM actuation is not yet safe."},{"subsystem":"MIND-S3","choice":"BANSHO safety kernel (see below) on a separate microcontroller per KURA-3, with karakuri fallback positions","why":"Same pattern that made DeepMind autonomy acceptable, applied at home scale."},{"subsystem":"MIND-S4","choice":"Rule AFDD + fleet-sibling anomaly detection across identical modules; BOPTEST-style emulator per module type","why":"Identical modules at many sites give the labelled-fault data that single-building FDD lacks."},{"subsystem":"MIND-S5","choice":"ROS 2 + Open-RMF with crease rails modelled as lanes and JO-PORT chargers as shared resources","why":"Discrete lane graph is exactly what Open-RMF schedules well; open source avoids lock-in."},{"subsystem":"MIND-S6","choice":"YOBIDASHI remote-assistance queue with human-minutes ledger; teleoperation only with per-session consent","why":"Proven at robotaxi scale and codified by IMO MASS for ships."},{"subsystem":"MIND-S7","choice":"FUYU-GOMORI island mode with five survival tiers and 4-14 day coast target","why":"Antarctic AGO and Halley VI lessons."},{"subsystem":"MIND-S8","choice":"IEC 62443 zones (safety, plant, robots, occupant, cloud), outbound-only edge, signed envelopes","why":"Building systems are now ransomware leverage."}],"new_mechanisms_or_modules":[{"name":"BANSHO (guardhouse) safety kernel","what":"A small deterministic controller in every KURA-3 that owns the actuators and accepts only signed, bounded setpoint envelopes from any AI, remote human or LLM.","how_it_works":"The kernel holds per-port hard limits (temperature bands, tank minimums, battery reserve for N island days, robot speed/force zones) compiled from the Brick graph. Each incoming command is checked against limits and rate-of-change; if the untrusted channel is silent, uncertain or violating for more than one cycle, the kernel reverts to the baseline rule controller, and if that fails to the karakuri state (valves spring-closed, hinges locked, heat-pump in frost protect). All decisions are logged to an append-only ledger visible to occupants.","prototype_12mo":{"build":"STM32/RP2040-class safety MCU + industrial PC + 8 relays + Modbus I/O on a bench KURA-3 mock-up with heat pump, tank and battery emulator; BOPTEST case for the AI channel","bom_usd":"about 1,500 (estimate, not quoted)","test":"10,000 randomised and adversarial command sequences from RL, LLM and fuzzers; 200 h hardware-in-the-loop run with injected sensor faults","pass":"Zero hard-limit violations; fallback within one cycle; comfort and tank constraints held","kill":"Any limit violation that cannot be fixed by specification, or fallback causing a worse state than the karakuri default"}},{"name":"KIROKU fold passport graph","what":"Each module ships with a machine-readable Brick/223P graph and calibrated emulator; docking at a JO-PORT merges graphs automatically.","how_it_works":"The configurator emits RDF for every module; JO-PORT contacts carry a module ID (e.g. NFC + wired 1-Wire) that the edge reads on docking, pulls the graph from the passport, and connects it to the site graph (feeds/isFedBy relations). Controllers query the merged graph, so a new YOSHI-6G greenhouse becomes controllable without commissioning labour.","prototype_12mo":{"build":"Graph generator in the configurator; 3 mock modules with ID tags on a bench JO-PORT; SPARQL-driven control app","bom_usd":"about 800 (estimate)","test":"Dock/undock 100 times in random order; run the same control and FDD apps on 5 configurations","pass":"Apps run unchanged on all configurations; commissioning under 10 min per module","kill":"If manual point mapping is still needed for more than 10% of points"}},{"name":"YOBIDASHI remote-assistance queue","what":"Habitat-initiated requests to a remote operations centre for advice, with a public human-minutes ledger per site.","how_it_works":"When FDD confidence or a robot planner's confidence falls below threshold, the site packages context (graph snippet, last 24 h trends, camera stills with blurring, candidate actions) into a ticket; an operator chooses or edits an option; the site validates it through BANSHO before acting. Minutes per ticket are logged to compute human_minutes_per_week, the KPI of the autonomy ladder.","prototype_12mo":{"build":"Ticket service + operator console on top of MQTT; 3 pilot sites (bench, one rooftop, one farm)","bom_usd":"about 3,000 software/hosting for 12 months (estimate)","test":"6-month operation logging all interventions","pass":"<=60 human min/week per site (L3) by month 6 with no safety event","kill":">180 min/week per site at month 6, or any action bypassing BANSHO"}},{"name":"FUYU-GOMORI (winter retreat) island mode","what":"Tiered graceful degradation that lets any ORIJO site run without network or visits, then recover by itself.","how_it_works":"Five tiers: T0 life safety and freeze protection, T1 water integrity, T2 data and comms beacon, T3 comfort, T4 production. A daily energy forecast vs stored energy decides the highest affordable tier; robots park on rails at chargers; a low-rate satellite beacon reports status. Recovery replays the reverse order with self-tests.","prototype_12mo":{"build":"Firmware on BANSHO + energy controller; test in a cold chamber or winter site with 1 module","bom_usd":"about 2,000 incl. chamber time (estimate)","test":"14-day simulated island with 50% PV deficit and one of three energy sources removed","pass":"T0-T2 held for 14 days; automatic recovery to T4 within 24 h of power return","kill":"Freeze damage or loss of data beacon in any run"}}],"sizing_reference":{"formulae":"points = sum(module_points); telemetry_bytes_day = points x 1440 samples x 8 bytes; edge_energy_kWh_day = P_edge_W x 24 / 1000; ROC_sites_per_operator = 2,400 productive min/week / human_min_per_site_week. Point counts and edge power are design assumptions, not sourced.","home_4p":{"points":"about 300 (KURA-3 about 120 + 7 other modules x about 25) - assumed","telemetry":"300 x 1440 x 8 = 3.46 MB/day raw, average 0.32 kbit/s - calculated","edge_compute":"industrial PC 10-15 W + safety MCU about 1 W -> 0.26-0.38 kWh/day - calculated from assumed power","target_level":"L3 at launch (15-60 min/week), L4 by year 3 - target","roc_ratio":"2,400 / 60 = 40 homes per operator at L3; 160 at 15 min - calculated; compare Waymo about 43 vehicles per agent (reported)"},"hotel_6key":{"points":"about 440 (core 200 + 6 rooms x 40) - assumed","telemetry":"440 x 1440 x 8 = 5.1 MB/day - calculated","robots":"2-3 TSURU + 1 cleaning robot coordinated by Open-RMF - target","target_level":"L3 for building systems (<=120 min/week excluding housekeeping) - target","roc_ratio":"2,400 / 120 = 20 hotels per operator - calculated"},"farm_module":{"points":"about 300 for 4 HATA-4 cassettes + climate - assumed","vision_data":"8 cameras x 24 images/day x 0.5 MB = 96 MB/day, processed at edge, only features uplinked - calculated from assumptions","control":"Setpoint API identical to Autonomous Greenhouse Challenge pattern - design","target_level":"L3 climate (AI setpoints) with labour tasks robotised progressively; 30-60 min/week remote grower time - target","roc_ratio":"One remote grower per 40-80 modules - calculated"}}},"open_problems":["No audited field evidence yet for LLM agents in the actuation path of buildings; current results show fabricated schedules under constraint conflicts.","Business case for advanced control is under-reported: only 13 of 104 field-demo papers report deployment or O&M costs.","No accepted autonomy-level standard for buildings/habitats; ALFUS, SAE J3016, IMO MASS degrees and spacecraft levels are not harmonised. ORIJO's L0-L5 with measurable minutes/MTBI/island-days is a proposal.","Brick/223P models are still hand-authored for most buildings; auto-generation from configurators is unproven at scale.","Teleoperation inside homes and guest rooms lacks a privacy and labour standard; 1X's consent model is company policy, not regulation.","Functional safety standards (ISO 13482:2014) predate learning-based and teleoperated home robots.","Long-duration (>90 day) uncrewed habitat autonomy with life support has no flight or analog heritage; Gateway VSM never flew.","Fleet-sibling FDD across identical modules is a hypothesis; no published evidence for prefab modular homes.","Insolvency resilience of cloud-controlled hardware (Freight Farms) has no legal template beyond source escrow.","Widely quoted 'FANUC robots run unattended for 30 days' could not be confirmed on FANUC's own pages."],"sources":[{"id":"src-dm2016","title":"Google DeepMind blog (20 Jul 2016) DeepMind AI reduces Google data centre cooling bill by 40%","url":"https://deepmind.google/discover/blog/deepmind-ai-reduces-google-data-centre-cooling-bill-by-40/","type":"company","accessed":"2026-10-29","read_primary":true},{"id":"src-dm2018","title":"Google DeepMind blog (17 Aug 2018) Safety-first AI for autonomous data centre cooling and industrial control","url":"https://deepmind.google/discover/blog/safety-first-ai-for-autonomous-data-centre-cooling-and-industrial-control/","type":"company","accessed":"2026-10-29","read_primary":true},{"id":"src-luo2022","title":"Luo, Paduraru, Voicu et al. (2022) Controlling Commercial Cooling Systems Using Reinforcement Learning, arXiv:2211.07357 (DeepMind, Google, Trane)","url":"https://arxiv.org/abs/2211.07357","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-mulayim2025","title":"Mulayim, Pergantis, Reyes Premer, Chen, Qu, Kircher, Berges (2025) Comparative Field Deployment of RL and MPC for Residential HVAC, arXiv:2510.01475","url":"https://arxiv.org/abs/2510.01475","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-khabbazi2025","title":"Khabbazi, Pergantis, Reyes Premer et al. (2025) Lessons learned from field demonstrations of MPC and RL for residential and commercial HVAC: A review, Applied Energy 399:126459 (arXiv:2503.05022)","url":"https://arxiv.org/abs/2503.05022","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-liu2022","title":"Liu, Balaji, Gao et al. (2022) Safe HVAC Control via Batch Reinforcement Learning, ICCPS 2022","url":"https://doi.org/10.1109/iccps54341.2022.00023","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-bengea2014","title":"Bengea, Kelman, Borrelli et al. (2014) Implementation of model predictive control for an HVAC system in a mid-size commercial building, HVAC&R Research 20(1)","url":"https://doi.org/10.1080/10789669.2013.834781","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-zhang2022","title":"Zhang, Prakash, Paul et al. (2022) Model Predictive Control for Demand Flexibility: Real-world Operation of a Commercial Building with PV and Battery Systems, Advances in Applied Energy","url":"https://doi.org/10.1016/j.adapen.2022.100099","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-wang2023rlmpc","title":"Wang, Zheng, Wang et al. (2023) Comparison of reinforcement learning and model predictive control for building energy system optimization, Applied Thermal Engineering (BOPTEST)","url":"https://www.sciencedirect.com/science/article/pii/S1359431123004593","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-arroyo2022","title":"Arroyo, Spiessens, Helsen (2022) Comparison of optimal control techniques for building energy management, Frontiers in Built Environment","url":"https://doi.org/10.3389/fbuil.2022.849754","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-blum2025","title":"Blum, Lu, Urrutia, Zheng et al. (2025) IBPSA Project 2 BOPTEST: overview, new software features, Building Simulation 2025","url":"https://publications.ibpsa.org/conference/paper/?id=bs2025_1666","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-trane-bbai","title":"Facilities Dive (20 Dec 2024) Trane Technologies agrees to acquire BrainBox AI","url":"https://www.facilitiesdive.com/news/trane-technologies-agrees-to-acquire-brainbox-ai/736155/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-phaidra-news","title":"PR Newswire / PRWeb items on Phaidra (Aligned Data Centers partnership; $12M raise) - search results only","url":"https://www.prweb.com/releases/phaidra-raises-12-million-to-help-data-centers-increase-compute-infrastructure-and-solve-the-most-urgent-problem-facing-ai-302214004.html","type":"news","accessed":"2026-10-29","read_primary":false,"status":"URL returned HTTP 404 on 2026-10-29 (prweb and prnewswire); claim unverified"},{"id":"src-katipamula2005","title":"Katipamula & Brambley (2005) Methods for Fault Detection, Diagnostics, and Prognostics for Building Systems - A Review, Part I, HVAC&R Research 11(1)","url":"https://doi.org/10.1080/10789669.2005.10391123","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-mirnaghi2020","title":"Mirnaghi & Haghighat (2020) Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: A comprehensive review, Energy and Buildings 229:110492","url":"https://doi.org/10.1016/J.ENBUILD.2020.110492","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-chen2023fdd","title":"Chen, O'Neill, Wen et al. (2023) A review of data-driven fault detection and diagnostics for building HVAC systems, Applied Energy 339:121030","url":"https://doi.org/10.1016/j.apenergy.2023.121030","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-cespedes2024","title":"Cespedes-Cubides & Jradi (2024) A review of building digital twins to improve energy efficiency in the building operational stage, Energy Informatics 7","url":"https://doi.org/10.1186/s42162-024-00313-7","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-balaji2016","title":"Balaji, Bhattacharya, Fierro et al. (2016) Brick: Towards a Unified Metadata Schema for Buildings, ACM BuildSys","url":"https://doi.org/10.1145/2993422.2993577","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-fierro2019","title":"Fierro, Koh, Agarwal et al. (2019) Beyond a House of Sticks: Formalizing Metadata Tags with Brick, ACM BuildSys","url":"https://doi.org/10.1145/3360322.3360862","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-roa2023","title":"Duarte Roa, Prakash, Peffer (2023) Field demonstration of the Brick ontology to scale up the deployment of ASHRAE Guideline 36 control sequences (OSTI 1996791)","url":"https://www.osti.gov/biblio/1996791","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-elias2026","title":"Elias, Hutchins, Issa (2026) Operational digital twins in the built environment: semantic infrastructure (Brick, Haystack, ASHRAE 223P), ITcon","url":"https://itcon.org/papers/2026_43-ITcon-Elias.pdf","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-matter15","title":"CE Pro (20 Nov 2025) Connectivity Standards Alliance releases Matter 1.5, adding cameras, closures and new energy features","url":"https://www.cepro.com/news/connectivity-standards-alliance-releases-matter-1-5-adding-cameras-closures-and-new-energy-features/623552/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-openrmf-news","title":"The Robot Report (8 Jan 2024) New version of Open-RMF hits ROS 2","url":"https://www.therobotreport.com/new-version-of-open-rmf-hits-ros2/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-openrmf-yt","title":"YouTube: Fleet Management with Open-RMF - Episode 1: Doors; OpenRMF an introduction and overview (shows simulated door/lift integration, not a hospital deployment)","url":"https://www.youtube.com/watch?v=r_VIxVeLcw8","type":"video","accessed":"2026-10-29","read_primary":false},{"id":"src-frota360","title":"Freitas, Mesquita, Pimenta, Raffo et al. (2026) Frota 360: fleet management for heterogeneous robots on offshore platforms, IEEE Access (notes Open-RMF use cases in hospitals)","url":"https://ieeexplore.ieee.org/abstract/document/11678973/","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-amrf2026","title":"Nangsue, Shubanchong (2026) AMRF: software architecture for AMR fleet management in warehouses (compares Open-RMF, VDA 5050)","url":"https://doi.org/10.1145/3816713.3819506","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-robofleet","title":"Sikand, Zartman, Rabiee et al. (2021) Robofleet: Open Source Communication and Management for Fleets of Autonomous Robots, IROS","url":"https://doi.org/10.1109/IROS51168.2021.9635830","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-waymo-futurism","title":"Futurism (23 Feb 2026) Here's how many remote operators Waymo has per self-driving taxi (citing Waymo letter to Sen. Markey and Waymo ops blog)","url":"https://futurism.com/advanced-transport/waymo-remote-operators","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-1x-engadget","title":"Engadget (29 Oct 2025) 1X Neo is a $20,000 home robot that will learn chores via teleoperation","url":"https://www.engadget.com/ai/1x-neo-is-a-20000-home-robot-that-will-learn-chores-via-teleoperation-040252200.html","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-1x-robotreport","title":"The Robot Report (21 Nov 2025) Teleop, not autonomy, is the path for 1X's NEO humanoid","url":"https://www.therobotreport.com/teleop-not-autonomy-the-path-for-1x-neo-humanoid/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-1x-yt","title":"YouTube: WSJ 'I Tried the First Humanoid Home Robot. It Got Weird.' (shows NEO tasks performed under remote teleoperation)","url":"https://www.youtube.com/watch?v=f3c4mQty_so","type":"video","accessed":"2026-10-29","read_primary":false},{"id":"src-imo-mass","title":"IMO press briefing (22 May 2026) IMO adopts first global Code for autonomous ships (MSC 111)","url":"https://www.imo.org/en/mediacentre/pressbriefings/pages/imo-adopts-mass-code.aspx","type":"agency","accessed":"2026-10-29","read_primary":true},{"id":"src-porathe2014","title":"Porathe, Prison, Man (2014) Situation awareness in remote control centres for unmanned ships (EU MUNIN)","url":"https://scholar.google.com/scholar?q=%22Situation+awareness+in+remote+control+centres+for+unmanned+ships%22","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-munin2013","title":"Burmeister, Porathe et al. (2013) Maritime Unmanned Navigation through Intelligence in Networks: The MUNIN project","url":"https://scholar.google.com/scholar?q=%22Maritime+Unmanned+Navigation+through+Intelligence+in+Networks%3A+The+MUNIN+project%22","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-cheng2023","title":"Cheng, Utne, Wu et al. (2023) A novel system-theoretic approach for human-system collaboration safety: two degrees of autonomy for autonomous ships, Reliability Eng. & System Safety 109388","url":"https://doi.org/10.1016/j.ress.2023.109388","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-melville2014","title":"Melville, Stillinger, Gerrard, Weatherwax (2014) Sustainable energy at the 100 W level for scientific sites on the Antarctic Plateau: lessons learned from the PENGUIn-AGO project, Rev. Sci. Instrum. 85:045117, doi:10.1063/1.4871555","url":"https://pubs.aip.org/rsi/article/85/4/045117/357636/Sustainable-energy-at-the-100-W-level-for","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-plato2010","title":"Luong-Van, Ashley et al. (2010) Performance of the autonomous PLATO Antarctic Observatory over two full years, Proc. SPIE","url":"https://doi.org/10.1117/12.857910","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-plato2009","title":"Ashley, Allen et al. (2009) The PLATO observatory: robotic astronomy from the Antarctic plateau, Proc. IAU (6-tonne self-contained, deployed Dome A Jan 2008)","url":"https://doi.org/10.1017/S1743921310010811","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-bas-halley","title":"British Antarctic Survey: Halley VI Research Station facility page","url":"https://www.bas.ac.uk/polar-operations/sites-and-facilities/facility/halley/","type":"agency","accessed":"2026-10-29","read_primary":true},{"id":"src-anderson2024","title":"Anderson & Badger (2024) Gateway Autonomy for Enabling Deep Space Exploration, IEEE Aerospace Conference","url":"https://doi.org/10.1109/AERO58975.2024.10521067","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-dabney2023","title":"Dabney, Badger, Rajagopal (2023) Trustworthy Autonomy for Gateway Vehicle System Manager, IEEE SCC","url":"https://doi.org/10.1109/SCC57168.2023.00018","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-badger2019","title":"Badger & Strawser (2019) A Distributed Hierarchical Framework for Autonomous Spacecraft Control (MAST), IEEE Aerospace Conference","url":"https://doi.org/10.1109/AERO.2019.8742199","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-gateway-ifl","title":"IFLScience (24 Mar 2026) NASA pauses Lunar Gateway, announces ~$20B moon base plan","url":"https://www.iflscience.com/nasa-kills-lunar-gateway-announces-ambitious-plans-for-moon-base-ready-in-early-2030s-82959","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-stetson2011","title":"Stetson, Knickerbocker et al. (2011) The HAL 9000 Space Operating System (Timeliner automated command software on ISS), IEEE Aerospace Conference","url":"https://doi.org/10.1109/AERO.2011.5747602","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-nasa-mars-delay","title":"NASA NTRS 20180007474, Emerging Technologies for Human Mars Exploration (search snippet: up to 44 min question-answer delay; ~12 h Earth contact per sol); PDF could not be text-extracted here","url":"https://ntrs.nasa.gov/api/citations/20180007474/downloads/20180007474.pdf","type":"agency","accessed":"2026-10-29","read_primary":false,"status":"PDF fetched but text not extractable; 44-min statement not located"},{"id":"src-baker2025","title":"Baker & Phillips (2025) Spacecraft Autonomy Levels, arXiv:2503.01049","url":"https://arxiv.org/abs/2503.01049","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-alfus2005","title":"Huang, Pavek et al. (2005) Autonomy Levels for Unmanned Systems (ALFUS) framework: an update, Proc. SPIE 5804","url":"https://doi.org/10.1117/12.603725","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-seibert2019","title":"Seibert, Lim, Miller et al. (2019) Developing Future Deep-Space Telecommunication Architectures: A Historical Look at the Benefits of Analog Research on the Development of Solar System Internetworking for Future Human Spaceflight, Astrobiology 19(3)","url":"https://doi.org/10.1089/ast.2018.1915","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-hemming2019","title":"Hemming, de Zwart, Elings et al. (2019) Remote Control of Greenhouse Vegetable Production with Artificial Intelligence, Sensors 19(8):1807","url":"https://doi.org/10.3390/s19081807","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-hemming2020","title":"Hemming, de Zwart, Elings et al. (2020) Cherry Tomato Production in Intelligent Greenhouses - Sensors and AI for Control, Sensors 20(22):6430","url":"https://doi.org/10.3390/s20226430","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-maree2025","title":"Maree, Zhang et al. (2025) Autonomous Greenhouse Cultivation of Dwarf Tomato: 4th Autonomous Greenhouse Challenge, Sensors 25(14):4321","url":"https://doi.org/10.3390/s25144321","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-agc-yt","title":"YouTube: WUR Autonomous Greenhouse Challenge videos (team intros, 4th edition winner IDEAS; show remote dashboards, not robots)","url":"https://www.youtube.com/watch?v=Y0D8LfEhlPw","type":"video","accessed":"2026-10-29","read_primary":false},{"id":"src-freightfarms","title":"HortiDaily (5 May 2025) Freight Farms' bankruptcy filing reveals $7M in liabilities","url":"https://www.hortidaily.com/article/9728610/freight-farms-bankruptcy-filing-reveals-7m-in-liabilities-and-how-the-industry-is-responding/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-plenty","title":"Bondoro (Apr 2025) Case Summary: Plenty Chapter 11","url":"https://bondoro.com/plenty/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-aerofarms-news","title":"Fast Company (9 May 2025) AeroFarms turned itself around after bankruptcy (headline only)","url":"https://www.fastcompany.com/91330232/aeorfarms-turned-itself-around-after-bankruptcy-is-it-a-sign-that-vertical-farming-can-succeed","type":"news","accessed":"2026-10-29","read_primary":false},{"id":"src-growy","title":"Straits Times (5 Nov 2025) S'pore vertical farm Growy winding up after official launch in Nov 2024 (headline only)","url":"https://www.straitstimes.com/singapore/environment/spore-vertical-farm-growy-winding-up-after-official-launch-in-nov-2024","type":"news","accessed":"2026-10-29","read_primary":false},{"id":"src-jci2023","title":"Cybersecurity Dive (28 Sep 2023) Johnson Controls hit by 'severe' cyberattack","url":"https://www.cybersecuritydive.com/news/johnson-controls-ransomware-attack/695130/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-jci-27m","title":"Dark Reading (31 Jan 2024) Johnson Controls ransomware cleanup costs top $27M (headline; page returned 403)","url":"https://www.darkreading.com/ics-ot-security/johnson-controls-ransomware-cleanup-costs-27m","type":"news","accessed":"2026-10-29","read_primary":false,"status":"403 on re-check 2026-10-29; $27M figure unverified"},{"id":"src-hvac-ransom","title":"Security Magazine (24 Sep 2025) Disabling hospital HVAC is now a bargaining chip in ransomware (headline only)","url":"https://www.securitymagazine.com/blogs/14-security-blog/post/101927-disabling-hospital-hvac-is-now-a-bargaining-chip-in-ransomware","type":"news","accessed":"2026-10-29","read_primary":false},{"id":"src-granzer2010","title":"Granzer, Praus, Kastner (2010) Security in Building Automation Systems, IEEE Trans. Industrial Electronics 57(11)","url":"https://doi.org/10.1109/TIE.2009.2036033","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-kaur2015","title":"Kaur, Tonejc et al. (2015) Securing BACnet's Pitfalls, IFIP SEC","url":"https://doi.org/10.1007/978-3-319-18467-8_41","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-isa62443","title":"ISA: ISA/IEC 62443 Series of Standards (overview page)","url":"https://www.isa.org/standards-and-publications/isa-standards/isa-iec-62443-series-of-standards","type":"standard","accessed":"2026-10-29","read_primary":true},{"id":"src-iso10218-2025","title":"Control Engineering (19 Feb 2025) Revised industrial robot safety standards now available (ISO 10218-1/-2:2025; snippet; page 403) and A3/Robot Report coverage noting ISO/TS 15066 content absorbed into ISO 10218-2:2025","url":"https://www.controleng.com/revised-industrial-robot-safety-standards-now-available/","type":"standard","accessed":"2026-10-29","read_primary":false},{"id":"src-iso13482","title":"ISO 13482:2014 Robots and robotic devices - Safety requirements for personal care robots (catalogue page; 403 on fetch)","url":"https://www.iso.org/standard/53820.html","type":"standard","accessed":"2026-10-29","read_primary":false},{"id":"src-fanuc","title":"FANUC Corporation: Production facilities page ('long hours of continuous unmanned machining'; highly robotized factories)","url":"https://www.fanuc.co.jp/en/profile/production/index.html","type":"company","accessed":"2026-10-29","read_primary":true},{"id":"src-jonah2026","title":"Jonah, Moses et al. (2026) LLMs for Agentic Home Energy Management, arXiv:2607.04569","url":"https://doi.org/10.48550/arXiv.2607.04569","type":"paper","accessed":"2026-10-29","read_primary":true},{"id":"src-agenticmpc2026","title":"Ren, Hong et al. (2026) AgentiControl-MPC: automated MPC for buildings using an LLM-based multi-agent system, Advanced Engineering Informatics 105062","url":"https://doi.org/10.1016/j.aei.2026.105062","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-autocontrolmpc2026","title":"Hu, Li, Han, Li, Wang (2026) AutoControl-MPC: automated MPC development framework based on LLMs, Energy and Buildings (tested on BOPTEST)","url":"https://www.sciencedirect.com/science/article/pii/S0378778826010856","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-simuhome2026","title":"Seo, Yang et al. (2026) SimuHome: a temporal- and environment-aware benchmark for smart-home LLM agents, ICLR 2026","url":"https://proceedings.iclr.cc/paper_files/paper/2026/hash/43456e781fac3a034842f0807860474a-Abstract-Conference.html","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-us12326701","title":"US12326701B2 Deterministic industrial process control (AI setpoints vs local control setpoints, setpoint-source selection)","url":"https://patents.google.com/patent/US12326701B2/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-us12013673","title":"US12013673B2 Building control system using reinforcement learning","url":"https://patents.google.com/patent/US12013673B2/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-us20240393750","title":"US20240393750A1 Building management system with building domain system coordination (recommend action, predict affected entities)","url":"https://patents.google.com/patent/US20240393750A1/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-us12181844","title":"US12181844B2 Building management system with natural language model-based data structure generation","url":"https://patents.google.com/patent/US12181844B2/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-cn113848711","title":"CN113848711A Data center refrigeration control algorithm based on safe-model reinforcement learning (safe exploration)","url":"https://patents.google.com/patent/CN113848711A/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-jp7788110","title":"JP7788110B2 Method and system for remote assistance of autonomous agents","url":"https://patents.google.com/patent/JP7788110B2/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-pat-us20230202467","title":"US20230202467A1 Automatic imposition of vehicle speed restrictions depending on road situation analysis (applies during vehicle teleoperation sessions)","url":"https://patents.google.com/patent/US20230202467A1/en","type":"patent","accessed":"2026-10-29","read_primary":false},{"id":"src-gateway-req","title":"Adamek (2019) Gateway System Requirements, NASA NTRS 20190029153 (uncrewed operations via robust autonomous functions)","url":"https://ntrs.nasa.gov/citations/20190029153","type":"agency","accessed":"2026-10-29","read_primary":false,"status":"NTRS record confirms title/author/date (10 Jul 2019); 'uncrewed operations via autonomy' wording not visible in record description - unverified"},{"id":"src-wur-agc-cucumber","title":"Elings, Righini et al. (2020) Remote control of greenhouse cucumber production with AI - results from the first international autonomous challenge, Acta Hortic. 1294","url":"https://doi.org/10.17660/actahortic.2020.1294.9","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-freightfarms-ch7","title":"Vertical Farm Daily (1 May 2025) Freight Farms files for Chapter 7 bankruptcy, leaving growers disconnected (headline)","url":"https://www.verticalfarmdaily.com/article/9728357/freight-farms-files-for-chapter-7-bankruptcy-leaving-growers-disconnected/","type":"news","accessed":"2026-10-29","read_primary":false},{"id":"src-vertical-future","title":"AgFunderNews (31 Jul 2025) Automated vertical farming startup Vertical Future up for sale after posting $13m loss (headline)","url":"https://agfundernews.com/automated-vertical-farming-startup-vertical-future-up-for-sale-after-posting-13m-loss","type":"news","accessed":"2026-10-29","read_primary":false},{"id":"src-trane-lab","title":"BetaKit (20 May 2026) Trane and BrainBox AI open Montreal lab to cut energy waste in buildings (read: confirms acquisition at start of 2025; claims 68,000+ buildings, up to 25% utility cost and 40% carbon cut)","url":"https://betakit.com/trane-and-brainbox-ai-open-montreal-lab-to-cut-energy-waste-in-buildings/","type":"news","accessed":"2026-10-29","read_primary":true},{"id":"src-aastino","title":"Lawrence, Ashley et al. (2005) A remote, autonomous laboratory for Antarctica with hybrid power generation (AASTINO, Dome C, since Jan 2003)","url":"https://doi.org/10.1080/1448837X.2005.11464109","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-ibrahim2025","title":"Ibrahim, Eshima, Gebraeel, Nabity et al. (2025) A generative ML framework for anomaly response in cyclical ECLSS processes on a deep space habitat, Acta Astronautica","url":"https://www.sciencedirect.com/science/article/pii/S0094576525006654","type":"paper","accessed":"2026-10-29","read_primary":false},{"id":"src-plenty-emerge","title":"Plenty Unlimited homepage (plenty.ag, (c) 2026) news item 'Plenty completes restructuring, emerges from chapter 11'; Bloomberg Law headline 29 May 2025 'Plenty Concludes Restructuring, Emerges From Chapter 11' (search snippet only)","url":"https://www.plenty.ag/","type":"company","accessed":"2026-10-29","read_primary":true},{"id":"src-ec-machinery","title":"European Commission - Machinery (Regulation (EU) 2023/1230 applies on a mandatory basis from 20 January 2027; AI-powered safety functions; cyber-safety)","url":"https://single-market-economy.ec.europa.eu/sectors/mechanical-engineering/machinery_en","type":"regulator","accessed":"2026-10-29","read_primary":true,"added_by":"gaps-2026-10-29"}],"verification":{"date":"2026-10-29","method":"Adversarial check of the most load-bearing claims against primary sources via WebFetch (DeepMind blogs, arXiv abstracts, Europe PMC abstracts for MDPI papers, IMO press briefing, AIP abstract, BAS page, trade press), Crossref/Semantic Scholar DOI resolution for 21 paper DOIs, Google Patents page checks for 7 patents, one SERP call (layer SERP cap then exhausted at 48/48).","checked":26,"confirmed":16,"confirmed_items":["N01/N02 DeepMind 2016: up to 40% cooling energy, 15% PUE overhead (src-dm2016)","N03/N04 DeepMind 2018: ~30% after nine months, 12% at launch, 5-minute cycle, two-layer constraint check (src-dm2018)","N05 Luo et al. 2022: ~9% and ~13% at two live Trane sites (src-luo2022)","N06/N07 Mulayim et al. 2025: MPC 18.1% (CI 4.4-30.9), RL 20.9% (CI 2.6-38.3), RL ~1/3 less engineering, 3 discomfort reports (src-mulayim2025)","N08-N10 Khabbazi et al. 2025: 104 demos (24 res/80 com), 71% possibly unreliable, 16%/13% duration-weighted, 13 of 104 report costs; Applied Energy 399:126459 (src-khabbazi2025)","N14 1X NEO USD 20,000 or USD 499/month, owner approval, blurring, no-go zones (src-1x-engadget); 'not anytime soon' (src-1x-robotreport)","N15 IMO MASS Code adopted MSC 111, voluntary from 1 Jul 2026, mandatory EIF 1 Jan 2032 (src-imo-mass)","N18 AGO 'oversize by a factor of 5 or more' (src-melville2014)","N21 AGC ed.1: five teams, 96 m2, cucumber, 4 months, one team beat reference (src-hemming2019)","N22 AGC ed.2: all five AI teams beat human reference, 6 compartments, 6 months (src-hemming2020)","PA-maree2025 AGC ed.4: six algorithms all profitable, best ~45 kg/m2/yr (src-maree2025)","N23 Freight Farms Chapter 7: USD 7,010,826 liabilities vs USD 593,916 assets; Farmhand phased out (src-freightfarms, src-freightfarms-ch7)","N26 Halley VI: 16 former winter crew, now summer-only, autonomous in winter, 23 km relocation 2017 (src-bas-halley)","PA-jonah2026 arXiv:2607.04569 exists (submitted 6 Jul 2026), five LLM backends vs MILP, 7-day rolling test, fabricated schedules (src-jonah2026)","Matter 1.5 (20 Nov 2025): tariffs, carbon intensity, metering, solar, bidirectional EV charging, soil sensors, closures, cameras (src-matter15)","All 21 checked paper DOIs resolve to the stated titles/first authors (Crossref/Semantic Scholar); 7 patent numbers resolve to matching titles; sizing formulae recompute"],"corrected":[{"item":"N12 / P-brainbox acquisition","was":"acquired Dec 2024; 14,000+ buildings","now":"agreement announced 20 Dec 2024, completed start of 2025; 14,000+ at announcement, 68,000+ claimed May 2026 (all claims)","source":"src-trane-bbai, src-trane-lab"},{"item":"N13 Waymo remote assistance","was":"about 70 per about 3,000; four redundant centres","now":"70 agents on duty at any time; ratio ~1:43 calculated (Futurism states ~1:41); four centres (AZ, MI, two in Philippines) - 'redundant' not in source","source":"src-waymo-futurism"},{"item":"N17 AGO battery coast","was":"about 4 days (implied at ~100 W)","now":"about 4 days for the ~50 W instrument load; ~2 days lead-acid carry-over","source":"src-melville2014"},{"item":"N16/N19 AGO basis","was":"measured","now":"reported (values stated in abstract, not our measurement)","source":"src-melville2014"},{"item":"N20 Mars 44-min delay","was":"reported from NASA NTRS","now":"calculated (2 x ~22 min max one-way light time); not located in cited PDF","source":"src-nasa-mars-delay"},{"item":"P-plenty status","was":"failed","now":"restructured: Chapter 11 23 Mar 2025, emerged ~end May 2025, strawberry-only in Richmond VA","source":"src-plenty, src-plenty-emerge"},{"item":"P-aerofarms status","was":"pilot","now":"restructured (Chapter 11 2023, re-emerged smaller; headline-level)","source":"src-aerofarms-news"},{"item":"P-gateway status","was":"closed","now":"paused (deferred 'in its current form', not cancelled)","source":"src-gateway-ifl"},{"item":"Deterministic setpoint arbiter option","was":"TRL 7, 'in industrial process control products'","now":"TRL 6; patent US12326701B2 granted 2025-06-10 to Phaidra Inc.; product use not independently verified","source":"src-pat-us12326701"},{"item":"Supervised/safe RL TRL justification","was":"'Commercial ... two Trane-partner commercial sites'","now":"Trane sites were 2022 research experiments (TRL 6-7); data centres TRL 8","source":"src-luo2022, src-dm2018"},{"item":"P-freightfarms date","was":"Chapter 7 in 2025","now":"ceased operations and filed Chapter 7 on 30 Apr 2025 (D. Mass. 1:25-bk-10908)","source":"src-freightfarms-ch7"},{"item":"P-vf loss currency","was":"$13M loss","now":"pre-tax loss above GBP 10M (~USD 13.2M); for sale on insolvency market","source":"src-vertical-future"},{"item":"Home sizing telemetry rate","was":"3.5 MB/day, 0.33 kbit/s","now":"3.46 MB/day, 0.32 kbit/s (300 x 1440 x 8 B)","source":"calculated"},{"item":"src-seibert2019 / PA-seibert2019 title","was":"truncated title 'Benefits of Analog Research'","now":"full Crossref title","source":"doi:10.1089/ast.2018.1915"},{"item":"P-openrmf origin","was":"Singapore government programme","now":"OSRC Singapore (acquired by Intrinsic Dec 2022); government-programme origin not verified in cited source","source":"src-openrmf-news"}],"removed":[],"unverifiable":["Johnson Controls cleanup cost > USD 27M (Dark Reading 403; alternative URLs 404) - kept, labelled unverified","Phaidra USD 12M raise / Aligned partnership (PRWeb and PR Newswire URLs return 404) - kept as unverified; company activity indirectly supported by 2025 patent grant","N11 Katipamula & Brambley 15-30% energy waste (paper exists, 1,028 citations, but abstract elided by publisher; widely cited secondary) - kept as reported","N20 Mars 'up to 44 min' wording in NTRS 20180007474 (PDF not text-extractable) - basis changed to calculated","Growy winding-up (Straits Times not fetchable) - headline-level","src-gateway-req 'uncrewed operations via robust autonomous functions' wording not visible in NTRS record","Whether the Autonomous Greenhouse Challenge has a 5th edition active in 2026"],"internal_consistency":"All cited source ids exist (one source added: src-plenty-emerge). 16 sources are listed but not cited anywhere (mostly patents and secondary papers) - retained as bibliography. Sizing formulae recompute correctly except a rounding fix (0.33 -> 0.32 kbit/s).","verdict":"Substantially sound. Core quantitative evidence (DeepMind, Luo, Mulayim, Khabbazi, AGC papers, AGO, IMO, Freight Farms) matches primary sources with correct units. Main errors were status labels (Plenty and AeroFarms restructured rather than failed/pilot; Gateway paused not closed), a battery-coast figure applied to the wrong load, a Mars-delay figure not actually found in its source, and an over-stated TRL for the patented setpoint arbiter (which is Phaidra's patent). No fabricated papers or DOIs found. Three trade-press claims (JCI USD 27M, Phaidra USD 12M, Growy) remain headline-level."},"certification":[{"id":"C-M1","regime":"EU Machinery Regulation 2023/1230","jurisdiction":"EU","status":"mandatory from 2027-01-20","applies_to":"Machinery incl. robots and AI-powered safety functions","requirement":"Provisions for AI-powered safety functions and cyber-safety of safety-relevant software","implication_for_orijo":"Keep learned components out of safety functions: the local deterministic kernel stays the certified safety function; ML proposes only.","sources":["src-ec-machinery"]},{"id":"C-M2","regime":"ISA/IEC 62443","jurisdiction":"international","status":"in force","applies_to":"Industrial automation and control security","requirement":"Zones and conduits, secure development lifecycle","implication_for_orijo":"Already adopted in mind dossier; required for remote operations credibility.","sources":["src-isa62443"]}],"revision_log":[{"date":"2026-10-29","by":"gaps-2026-10-29","change":"Added certification paths, new sources/numbers/papers/projects from the completeness-critic pass; see research/v4/gaps-log.md"}]}