A Fuzzy Path Preview Algorithm for the Rice Transplanting Robot Navigation System

摘要

Path preview is the key to realizing automatic navigation control that can guide the navigational direction of the vehicle trace. The preview point must be regulated dynamically according to the predetermined navigational features of the actual path preview control to obtain the ideal path trace result. Based on the “preview follower theory” combined with the driver’s experience and utilizing the Matlab (The MathWorks, Inc., Natick, Massachusetts, USA) fuzzy control toolbox, a fuzzy algorithm of the preview point has been proposed in the paper. The navigation control system embedded in the path planning and preview control algorithm has been applied in the experiment to trace a predetermined navigational straight way, 90°curve and arc way to operate a rice transplanting robot. The results showed that the lateral DRMS (distance root mean square error) of the straight trace is approximately 0.33 m, and the maximum error is approximately 0.71 m; the lateral DRMS of the 90°curve trace is approximately 0.73 m, and the maximum error is approximately 1.85 m; the lateral DRMS of the arc way trace is approximately 1.03 m, and the maximum error is approximately 2.56 m; and the total mean error is approximately 0.6 m. According to the outcome of the experiment, the relevant algorithm can basically meet the requirements of rice transplanting robot guidance.