Real-time Path Planning Algorithm based on Ridge Detection for Solid-State LiDAR-based Autonomous Agricultural Robot

Abstract

Autonomous robots driving in agricultural environments need to create accurate routes to prevent damaging farmland. Herein, a real-time path planning algorithm based on ridge detection using the Livox Mid70 LiDAR is proposed. The field has a steep slope between the furrow and the ridge, and we find data corresponding to this slope from the lidar data. Using two Mid70 LiDARs, we collect the left and right slope data and linearly fit the slope data using the RANSAC algorithm. Finally, the driving route is obtained from the average of the two straight lines. To verify the algorithm, the steering angle needed to follow the driving path was calculated, and an autonomous driving test was conducted in the actual field environment: it was observed that the robot drove stably. Additional tests will be conducted in more rugged terrain.