Distributionally Robust Risk Map for Learning-Based Motion Planning and Control: A Semidefinite Programming Approach

• Published in: IEEE transactions on robotics Vol. 39; no. 1; pp. 1 - 20
• Main Authors: Hakobyan, Astghik, Yang, Insoon
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.02.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Collision avoidance; Computational modeling; Control methods; Errors; Gaussian process; Learning; Motion control; Motion planning; Neural networks; Optimization; Planning; Predictive control; probability distribution; Robot dynamics; Robot kinematics; Robot sensing systems; Robots; Robustness; Safety; Semidefinite programming; Upper bounds

In this article, we propose a novel safety specification tool, called the distributionally robust risk map (DR-risk map), for a mobile robot operating in a learning-enabled environment. Given the robot's position, the map aims to reliably assess the conditional value-at-risk (CVaR) of collision...