Development of the integrated control system for an intelligent agricultural vehicle

Abstract. This study aims to develop the integrated control system for an intelligent agricultural vehicle, which consist of several functions including self-localization, 2D/3D mapping, situation awareness, path planning, remote monitoring, and feedback control. The system was designed in accordance with the standards of automation level 3 of the guidelines for ensuring the safety of automatic driving of agricultural machinery established by the Ministry of Agriculture, Forestry and Fisheries (MAFF) in Japan. To satisfy the standards of automation level 3, agricultural machinery needs to recognize the surrounding environment and secure safety using various sensors. In addition, the operator must be able to check and control the working state of the machine using a monitoring device. The proposed system integrated with cameras, 3D-LiDAR, GNSS positioning devices, and IMU was instrumented on a commercial wheel tractor that can be controlled by drive-by-wire technology and made it a level 3 autonomous vehicle. An experiment with multi-field soil tillage and road driving between fields following the prescribed target paths were conducted using the fully autonomous tractor to evaluate the performance of the control system. A mannequin was arbitrarily installed around the experimental area to test the emergency stop function of the system when detecting obstacles. The experimental result indicated that the control system could enable the self-localization of the tractor in the agricultural environment and accurately guide it along the target paths during the soil tilling and the road driving. The average root mean square error (RMSE) of the lateral distance from the target path to the navigation point of the tractor (the center of rear axle) was about 0.05m and 0.06m for the navigations within fields and between fields, respectively. The system also could locate the mannequin at the danger area around the tractor, then performed the emergency stop and sent a warning message to a handheld device.