Autonomous Driving Control of a Robotic Mower on Slopes Using a Low-Cost Two-Frequency GNSS Compass and an IMU

Abstract

Highlights In this study, we developed an autonomous driving system for a robotic mower operating on a slope. RTK-GNSS, a GNSS compass, and an IMU were used as sensors for the robotic mower. Autonomous driving tests on a slope showed that the driving accuracy of the system was sufficient to meet the requirements of mowing work. Abstract. An autonomous driving control system for a robotic mower was designed using a low-cost dual-frequency global navigation satellite system (GNSS) compass, a low-cost dual-frequency real-time kinematic GNSS (RTK-GNSS), and an inertial measurement unit (IMU). The intended application is to save labor by autonomously mowing sloped terrains. To investigate the effectiveness of the prototype GNSS compass on an inclined terrain, the roll, pitch, and yaw angles were varied using an inclined stage, and the heading accuracy was evaluated. The root mean square (RMS) of the heading error was within 1.5° when the roll and pitch angles were both less than 45°. Autonomous driving tests were conducted on the slope of an agricultural dam with an incline of approximately 25°. The prototype GNSS compass was capable of continuous and consistent heading measurements even on the slope. The cross-track error for the set path was 7.2 cm RMS when starting near the top and driving down the slope. When the robot started near the bottom of the slope and ran up, the cross-track error was 6.6 cm RMS. For the mower and experimental conditions used in this study, this was sufficient accuracy for autonomous driving when mowing on a slope. Keywords: Heading angle, Inclination, MAVLink, Moving baseline, Robot operation system (ROS).