Laser Scanner-Based Obstacle Detection System for Autonomous Tractor

Movement and Shape Detection Targeting at agricultural vehicle

Abstract

Detecting obstacles invading the region in front of a tractor, and creating an obstacle avoidance behavior are needed as a safety device for an autonomous tractor. This paper presents an obstacle detection system based on a laser scanner. The algorithm utilized a template matching for reconstructing a shape of a target obstacle and for detecting its movement. An extended Kalman filter worked for enhancing an obstacle detection accuracy. Field tests were conducted to evaluate the developed system. The results indicated that the system was capable of estimate a position, a speed, and a heading angle of an obstacle with 2.7cm, 0.13m/s, 1.4° error, respectively.