AUTONOMOUS GREENHOUSE SPRAYER VEHICLE USING MACHINE VISION AND LADAR FOR STEERING CONTROL

摘要

A six-wheel differential steering vehicle was designed and built to provide a vehicle platform for a greenhouse sprayer. The vehicle was designed to operate as a self-contained system, which would carry the power supply, spray tank, pump, and boom. Independently controlled DC motors were used to provide differential steering for the vehicle. A fuzzy logic-based proportional-derivative controller was developed and tested to navigate the vehicle through the aisle way using machine vision and laser radar range information. The vehicle was tested under three different path configurations using a flexible curvature track. The track was configured for straight, moderate and heavy curved conditions. Path tracking comparisons were made between the vision-based control and a laser radar (ladar) based control. The error was monitored using the ladar as the most accurate measure of the position error between the path centerline and the vehicle centerline. Error comparisons were made using RMS error and average instantaneous error. It was found that the ladar-based control was the better overall guidance sensor under all track conditions. Although, average errors were fairly low at approximately 10mm, the maximum errors left room for improvement at approximately 40mm. It was proposed, that a control scheme which used both machine visions and ladar may provide a more robust guidance control then using independent sensor range data.