WANTED-10K combines established time-to-event statistics, human-centered task research, long-term autonomy, digital-twin testing, and robot-safety practice. This record makes the intellectual lineage—and its boundaries—auditable.
Each source informs a defined part of the protocol. None is presented as validation of the complete WANTED construct.
Kaplan–Meier estimates retention when observations are censored. Restricted mean survival time is the area under that curve to a fixed horizon. WANTED adapts both methods from time-to-event analysis to time until voluntary robot rejection.
Stanford’s BEHAVIOR-1K grounds 1,000 household activities in human surveys and tests long-horizon mobile manipulation in realistic scenes. It is a strong prequalification layer for capability coverage.
Oxford’s GOALS research targets robots that operate for days, weeks, and months in dynamic environments. STRANDS reported 104 combined deployment days and 116 km across four sites, demonstrating the value of longitudinal autonomy evidence.
NVIDIA Isaac Sim and Isaac Lab support common robot descriptions, physics-based simulation, sensors, synthetic data, domain randomization, and parallel robot-learning workflows. WANTED uses these capabilities for prequalification and failure injection.
Harvard’s field research on robot overtrust measured both stated trust and behavioral compliance, showing why a robot being trusted is not, by itself, proof that the interaction is safe or appropriate.
RFC 8785 defines a JSON Canonicalization Scheme so equivalent event objects have one deterministic representation for hashing and signing. WANTED 0.2 uses JCS before SHA-256 hash-chain and signature operations.
IEC 60300 treats dependability as the ability to perform as required when required, spanning reliability, availability, maintainability, and supportability. ISO 14224 provides a mature pattern for exchanging equipment, failure, maintenance, resource, and downtime data across operators and manufacturers.
The relevant legal and standards framework depends on the robot, environment, jurisdiction, and intended use. A qualified assessor determines applicability.
Personal care robot product safety and physical human–robot contact within its stated scope.
VERIFY ↗Proposed broader service-robot safety requirements. Track status before making conformity claims.
VERIFY ↗Safety requirements for industrial robots, where that classification applies.
VERIFY ↗Integration, commissioning, operation, and maintenance of industrial robot applications and cells.
VERIFY ↗Electrical, battery, functional, mechanical, mobility, fire, and related hazards within its product scope.
VERIFY ↗It cannot reproduce the full social environment or generate human-retention outcomes.
It cannot show that the people sharing the environment chose continued coexistence.
It cannot excuse safety failures or isolate why a participant kept or rejected the robot.
It is cohort- and context-specific, and must travel with uncertainty, burden, incident, and support disclosures.