An integrated path planning system for a robot designed for oil palm plantations

Abstract

This paper presents a path planning system that would enable a robot to navigate through an oil palm plantation. The system is comprised of a distance estimation algorithm, which can accurately identify the location of a detected oil palm tree, an obstacle detection method that accurately detects obstacles and a path planning algorithm which can plan the shortest collision-free path for the robot to follow. The Dlite algorithm was selected as the path planning algorithm due to its exceptional performance. The tree detection algorithm uses the KINECT camera and a Viola and Jones detector based method to detect trees, which is then used to estimate the location of the nearest detected tree via further image processing. The distance estimation technique for tracking the distance to the nearest detected tree resulted in an estimation accuracy of 80% within a 4-6 meter range from the tree. All subsystems were integrated as a whole so that they function smoothly as an integrated system. Real time testing of the integrated system showed that the developed auto navigation system was capable of planning the shortest collision-free path to the goal and obstacles ahead were accurately detected.