Online planning method reduces safety risks in robot navigation
Online, Reachability-Aware, Sampling-Based Motion Planning
Robotics
Summary
Planning safe paths for robots as they move is tricky because the robots need to avoid crashes while staying efficient. The authors came up with a fast online method that predicts where a robot can safely go by computing 'reachable sets,' which are like safety bubbles around the robot's future positions. This method works as well as slower, pre-calculated techniques but can handle more complex robots and situations. Tests showed it cut safety problems by over 99% in race simulations and kept a model racecar from crashing in real runs.
What this means in practice
- •For robotics engineers: Use fast online reachability calculations to ensure safe navigation of robots without costly pre-computation.
- •For autonomous vehicle developers: Control real autonomous racecars or similar vehicles with reduced crashes through guaranteed safety checks during motion planning.
Authors
Brendan Gould, Zhiyuan Zhang, Panagiotis Tsiotras, Samuel Coogan
Abstract
Sampling-Based Model-Predictive Control (MPC) algorithms are a flexible class of controllers used for navigation on a wide range of robotic systems. Historically, such approaches have lacked hard safety guarantees, a shortcoming which we remedy in this work by computing guaranteed reachable-set overapproximations online with a fast, interval-based pipeline. We show that our method achieves similar performance to a state-of-the-art reachability-based planner without the need for the expensive pre-computation step, and can be scaled to systems that are infeasible using existing approaches. Finally, we demonstrate that our technique reduces safety violations by over 99% in a racing simulation and successfully controls a model racecar on real hardware experiments without crashes.