Robot controllers that plan safely in real time with fewer crashes
Online, Reachability-Aware, Sampling-Based Motion Planning
Robotics
Summary
Robots often need to plan their movement quickly and safely, especially in tricky situations like racing. The authors developed a way for robots to predict where they can safely go without spending a lot of time on complicated calculations beforehand. Their approach checks reachability—whether a robot can get to a certain place safely—as they move, helping avoid dangerous paths. Tests showed this method reduced crashes significantly and worked well on real robot racecars.
Sampling-Based Motion PlanningModel-Predictive ControlReachabilitySafety GuaranteesInterval AnalysisOnline ComputationRobotic NavigationReachable SetMotion PlanningRacecar Simulation
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.