Space to Time: Out-of-Distribution Generalization of Safety Filters via Temporal Disturbance Encoding

Sander Tonkens, Nikhil Uday Shinde, Azra Begzadić, Michael C. Yip, Jorge Cortes, Sylvia Lee Herbert

Keywords: Distributionally robust robot learning, Domain adaptation and generalization

Abstract:

Safe operation is essential for autonomous systems in safety-critical environments such as urban air mobility.
Value function-based safety filters provide formal guarantees on safety, wrapping an autonomy stack with a layer of protection on its outputted action.
Recent approaches leverage offline learned value functions to scale these safety filters to high-dimensional systems.
Yet these methods assume knowing the full operational domain a priori - information that is typically unavailable in real world settings.
For example, detailed prior knowledge of all possible sources of model mismatch, in the form of disturbances, in the environment is highly unrealistic.
Even in well-mapped environments like urban canyons or industrial sites, drones encounter complex, spatially-varying disturbances arising from payload-drone interaction, turbulent airflow, and other environmental factors.
We introduce space2time, which enables safe and adaptive deployment of offline-learned safety filters, by generalizing across unknown, spatially-varying disturbances.
The key idea is to reparameterize spatial disturbances as a time-varying formulation, allowing the use of temporally varying precomputed value functions during online operation.
We validate space2time through extensive simulations on diverse quadcopter models and real-world hardware experiments, demonstrating significantly improved safety performance over worst-case and naive baselines.

Submission Number: 36