Local Path Planning of an Autonomous Mobile Robot with Nonholonomic Constraints based on Control Barrier Functions and Elliptical Bounding Box

Autonomous mobile robots (AMRs) deployed in uncertain environments need to dynamically adjust their paths to avoid obstacles unmarked on the maps, such as walking people, moving vehicles, temporarily stacked objects, etc. This paper proposes a local path planner for avoiding obstacles and then returning to the global path. In planning the local path, nonholonomic constraints associated with the driving and steering mechanisms, and limits on the velocity and acceleration of the AMR are satisfied. In addition, a reasonable safety region is guaranteed by applying control barrier functions (CBFs) that use elliptical bounding boxes of obstacles and consider the heading angle of the AMR. Moreover, an artificial potential field (APF) is established to guide the AMR to return to the global path after bypassing the obstacles. Then simulations are conducted to verify the effectiveness of the proposed method in a complex environment with many static and dynamic obstacles.