English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

When we rely on deep-learned models for robotic perception, we must recognizethat these models may behave unreliably on inputs dissimilar from the trainingdata, compromising the closed-loop system's safety. This raises fundamentalquestions on how we can assess confidence in perception systems and to whatextent we can take safety-preserving actions when external environmentalchanges degrade our perception model's performance. Therefore, we present aframework to certify the safety of a perception-enabled system deployed innovel contexts. To do so, we leverage robust model predictive control (MPC) tocontrol the system using the perception estimates while maintaining thefeasibility of a safety-preserving fallback plan that does not rely on theperception system. In addition, we calibrate a runtime monitor using recentlyproposed conformal prediction techniques to certifiably detect when theperception system degrades beyond the tolerance of the MPC controller,resulting in an end-to-end safety assurance. We show that this controlframework and calibration technique allows us to certify the system's safetywith orders of magnitudes fewer samples than required to retrain the perceptionnetwork when we deploy in a novel context on a photo-realistic aircraft taxiingsimulator. Furthermore, we illustrate the safety-preserving behavior of the MPCon simulated examples of a quadrotor. We open-source our simulation platformand provide videos of our results at our project page:https://tinyurl.com/fallback-safe-mpc.

Closing the Loop on Runtime Monitors with Fallback-Safe MPC