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/

Zendy Plus

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

Zendy Tools

Zendy Open

In order to safely operate around humans, robots can employ predictive modelsof human motion. Unfortunately, these models cannot capture the full complexityof human behavior and necessarily introduce simplifying assumptions. As aresult, predictions may degrade whenever the observed human behavior departsfrom the assumed structure, which can have negative implications for safety. Inthis paper, we observe that how "rational" human actions appear under aparticular model can be viewed as an indicator of that model's ability todescribe the human's current motion. By reasoning about this model confidencein a real-time Bayesian framework, we show that the robot can very quicklymodulate its predictions to become more uncertain when the model performspoorly. Building on recent work in provably-safe trajectory planning, weleverage these confidence-aware human motion predictions to generate assuredautonomous robot motion. Our new analysis combines worst-case tracking errorguarantees for the physical robot with probabilistic time-varying humanpredictions, yielding a quantitative, probabilistic safety certificate. Wedemonstrate our approach with a quadcopter navigating around a human.

Probabilistically Safe Robot Planning with Confidence-Based Human Predictions