Research On Inspection And Obstacle Avoidance Methods Of Agro-machinery Obstructions

With the promotion and application of precision agriculture technology in China,agricultural automatic navigation technology is becoming more and more necessary in agricultural production.In the field of automatic navigation technology for agricultural machinery,automatic obstacle avoidance is the basic key technology to ensure the reliability and safety of automatic navigation applications for agricultural machinery.In this paper,sensors such as lasers and inertial measurement units were mounted on automatic navigation vehicle for agricultural machinery based on beidou navigation positioning system.With the help of multi-sensor fusion technology,continuous monitoring was carried out in an unstructured farmland environment to realize the study of obstacle identification,positioning and obstacle avoidance of obstacles with different dynamic characteristics and relative positions.The main contents and conclusions of this research were as follows:(1)The vehicle information acquisition and control hardware platform was constructed.The electric vehicle platform was selected as the test carrier,and an environmental information perception system including a laser scanner,a Beidou satellite locator and an inertial measurement unit was constructed.The automatic steering system that controls the direction of the test carrier was built.The front wheel declination detection system was constructed to provide a reliable basis for the feedback adjustment in the automatic steering control.The construction of the hardware platform established the basic skeleton for obstacle detection and obstacle avoidance systems proposed in this study.(2)A method for detection and identification of agricultural machinery obstacles in an unstructured environment was developed.For the farmland environment,Beidou,inertial measurement unit and two-dimensional laser scanner were selected as external sensors,and a method for real-time detection of dynamic obstacles was proposed.The Kalman filter algorithm was used to integrate Beidou and inertial sensor data to locate the position of the agricultural machine in real time.The results show that the proposed Kalman filter algorithm can reduce the average deviation of the distance between two points from 0.0895 m before filtering to 0.0355 m after filtering.The two-dimensional laser scanner was used to detect the distance between obstacles and agricultural machinery.On this basis,the absolute position and motion state of the obstacle were obtained by relative coordinate transformation.(3)An obstacle detection system based on ROS software platform was developed.The obstacle detection and recognition system was built on the ROS robot operating system platform,and functions such as sensor information acquisition,coordinate conversion,information fusion,and serial communication were realized.Static performance,real-time performance and dynamic performance of obstacle detection systems were tested.In the static test,the average deviation of the distance and size detected by the system when the obstacle was longitudinally offset were 1.02 cm and 1.08 cm.The lateral deviations and the average size deviations detected by the system when the obstacle was laterally offset were 1.13 cm and 1.34 cm.Real-time test measured system update frequency was 5.04 Hz.The experimental results proved the effectiveness and reliability of the proposed method in the automatic navigation of agricultural machinery.(4)An obstacle avoidance strategy for agricultural machinery navigation system based on Beidou positioning was developed.According to the obstacle type,status information and security level,a strategy for obstacle avoidance of agricultural navigation system based on Beidou positioning was proposed.The obstacle hierarchy was developed,and different obstacle avoidance measures were implemented for different state obstacles in different regions.A modified artificial potential field method was proposed to implement obstacle avoidance operations for static obstacles in agricultural machinery.(5)An obstacle detection and obstacle avoidance system for agricultural machinery operations was built and simulation and field test were conducted.On the Gazebo physics simulation software,by building an electric vehicle motion simulation model,a real-world simulation model,and a sensor simulation model,adding the algorithm file proposed by the research to simulate the obstacle avoidance test.In the simulation environment,the simulation model vehicle could complete the path planning task,and the generated path was smooth,basically meeting the test requirements.On the other hand,a large number of field trials were conducted on specific environments.The results of the tests showed that the agricultural machinery smoothly bypassed the static obstacles and verified the effectiveness of the inspection and obstacle avoidance methods of agro-machonery obstructions proposed in this paper.