Risk Aware Warning and Control for Interactive Traffic Safety
-
2024-07-31
-
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report
-
Corporate Publisher:
-
Abstract:In this project, we proposed a CBF-inspired risk assessment toolbox, measuring the aggregated risk faced by individual agents due to multi-agent interactions, to help empower the ego robot with a comprehensive understanding of the dynamic environment it is in. We also demonstrate two complementary methodologies of embedding the resulting risk assessment into robot controller design, especially for multi-robot scenarios. We show how to utilize such risk information to perform partition in robots’ joint state and action space respectively to construct robot controllers, resulting in the proposed Risk-aware WBVC and Risk-aware decentralized CBF. We further demonstrated the general applicability of our risk assessment in deployment to a physical F1-Tenth testing platform with existing MPPI-based vehicle controllers. A safe and efficient overtaking scenario is demonstrated. In the future, we are interested in extending these methods for safe navigation for self-driving cars with learning-enabled components.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: