A Loosely Coupled Extended Kalman Filter Algorithm for Agricultural Scene-Based Multi-Sensor Fusion

With the arrival of aging society and the development of modern agriculture, the use of agricultural robots for large-scale agricultural production activities will become a major trend in the future. Therefore, it is necessary to develop suitable robots and autonomous navigation technology for agricultural production. However, there is still a problem of external noise and other factors causing the failure of the navigation system. To solve this problem, we propose an agricultural scene-based multi-sensor fusion method via a loosely coupled extended Kalman filter algorithm to reduce interference from external environment. Specifically, the proposed method fuses inertial measurement unit (IMU), robot odometer (ODOM), global navigation and positioning system (GPS), and visual inertial odometry (VIO), and uses visualization tools to simulate and analyze the robot trajectory and error. In experiments, we verify the high accuracy and the robustness of the proposed algorithm when sensors fail. The experimental results show that the proposed algorithm has better accuracy and robustness on the agricultural dataset than other algorithms.

Automated summary

With the aging society and modern agriculture development, the use of agricultural robots for large-scale agricultural production will be a major trend in the future. To solve the problem of navigation system failure caused by external noise and other factors, a multi-sensor fusion method based on agricultural scenes is proposed, utilizing a loosely coupled extended Kalman filter algorithm to reduce interference from the external environment. Experimental results demonstrate the high accuracy and robustness of the proposed algorithm in case of sensor failures, showing better accuracy and robustness on agricultural datasets compared to other algorithms.