Event

Method proposed for side-by-side autonomous navigation of two personal mobility vehicles

1 reports · 1 sources

Luis Yoichi Morales and Florent Ferreri propose a method for two personal mobility vehicles (PMVs) to drive autonomously side by side: after a global path is given, each vehicle runs a sampling-based model predictive planner, exchanging only the odometry information each can observe of the other, and then makes a joint decision on the utility matrix of candidate trajectory pairs through common rules. Behavior parameters were fitted by inverse reinforcement learning from 409 minutes of indoor driving data from ten pairs of human-driven PMVs; usable lanes are derived offline from corridors rather than manually set; the safety layer executes an action only when the entire body of the vehicle can brake to a stop, and a stuck vehicle reverses along its own path before continuing. Experiments are based on differential-drive kinematic simulation on ROS 2, and no real-vehicle validation results have yet been seen.