Simulation Study to Develop Implement Control and Headland Turning Algorithms for Autonomous Tillage Operations

Abstract

This article reports a simulation study conducted using a three-dimensional tractor-driving simulator to develop an implement control algorithm and evaluate the tillage coverage and field efficiencies of a virtual autonomous tractor following paths generated based on different headland turning methods. To minimize the no-tilled or unnecessarily tilled areas that occurred in our previous study, a tillage implement control algorithm was designed by enabling the raising or lowering of a three-point hitch with appropriate delay times. The effects of headland turning methods, i.e., X-shaped, R-shaped, and C-shaped turns, on path tracking and full-path simulation of autonomous tillage operations were studied. The results of the simulation studies were evaluated in terms of tracking error, skipped area, and field efficiency. The developed implement control algorithm effectively reduced both no-tilled and unnecessarily tilled areas in comparison with those obtained without the implement control algorithm. The magnitudes of changes in no-tilled and unnecessarily tilled areas were from 4.5 to 0.4 m2 and from 4.6 to 0.3 m2 , respectively. In a study in which an autonomous tractor followed a desired path in a virtual field of 100 m ? 40 m at a constant traveling speed of 4 km h??1 designed to investigate the influence of the three different headland turning methods, the simulator allowed a quantifiable comparison of tillage operation performance for the various headland turns by showing lateral deviations < 9 cm, heading angle errors < 16째, ratios of skipped area < 2%, and field efficiencies ranging from 81.3 to 86.6%. Full-path simulation tests of the autonomous tillage operation conducted in virtual rectangular fields with a length of 100 m and various widths showed that field efficiency was inversely proportional to the field width. The use of the 3D tractor-driving simulator was effective for designing tillage implement control algorithms and studying the effects of headland turning methods on path tracking. These results are applicable to the development of path generation and tracking algorithms suitable for autonomous tillage operations