Abstract
Humanoid robots maintain dynamic balance continuously, and disturbances such as collisions, surface irregularities, or payload shifts can lead to falls. Such falls cause damage and can injure nearby people. Balance recovery is conventionally fully autonomous, and remote operators are excluded from the recovery loop, even though many falls develop over several seconds. This disclosure describes a staged approach in which a composite stability health score is computed from multiple balance metrics. When instability is detected, the predicted time to the point of no return is used to classify the event into one of four temporal classes. Each class is routed to an autonomous, shared, or operator-led handler. Escalation is permitted only when a computed time budget is positive and an operator readiness score is sufficient. Feasibility-verified recovery options are then presented for selection. Human judgment is thereby applied to slowly developing instability, while fast falls remain under autonomous protection.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Garikapati, Divya, "Predictive Instability-Driven Escalating Teleoperator Authority with Temporal Fall Triage and Menu-Based Recovery Selection for Humanoid Robots", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/12102