Research On Key Technologies Of Autonomous Tractor Driving And Field Tractor Operation

Agricultural mechanization and informatization can significantly improve labor productivity and have become one of the important pillars of the agricultural economy;automatic navigation of agricultural machinery can effectively improve the operation efficiency and operation quality of agricultural machinery,avoid missing and overlapping operations,and decrease the labor intensity of drivers.In the past decade,China’s agricultural machinery automatic navigation and assisted driving technology has matured,coupled with the continuous advancement of tractor electronic control technology,the degree of autonomy and intelligence has been continuously improved,and the automatic driving operation technology of agricultural machinery has gradually become one of the hotspots in the intelligent research of agricultural machinery equipment.The technological advancement and innovation from the automatic navigation of agricultural machinery to the automatic driving of agricultural machinery not only promote the progress of intelligent agricultural machinery,but also the accuracy of information acquisition,the stability and adaptability of control algorithms,the comprehensive performance of the system and the level of intelligence,put forward higher requirements.I take the wheeled tractor as the platform of agricultural machinery and I will elaborate the key technologies of its autonomous driving operations in this paper,mainly including the methods for planning path of automatic driving operations,the methods for obtaining the tractor’s position and attitude,the algorithm of path tracking control,the design and development of the tractor’s automatic driving operation system,etc,and the corresponding experiment and test.The specific research content is as follows:(1)According to the characteristics of the tractor’s automatic driving operation,I put forward two types of path planning namely a field route and a migration route.And I also put forward a general path planning method for two scenarios: a structure is used to represent the straight line operation section of the field route and the migration route.The third method contains the starting and ending plane position coordinates,the driving and operating status that the agricultural machine should adopt in this straight line segment,the way that the agricultural machine switches from the current straight line segment to the next one.And I finally adopt a linked list-based data structure to store,manage and use the decision results of the path planning algorithm,so as to realize the path planning of the tractor’s automatic driving operation.(2)According to the characteristics of multi-sensor fusion in the acquisition of tractor pose information,I converted the positioning points and navigation control points based on the conversion of coordinate system;I designed a Kalman filter based on RTK-GNSS and AHRS to effectively suppress the cumulative heading of the inertial system and correct the divergence of the GNSS positioning information.It is proved in the test that when the navigation of the tractor is controlled by the processing results of the information fusion processing algorithm,the maximum position deviation is less than 3cm,and the maximum heading deviation is less than 0.5 degrees.(3)I utilized a simplified tractor two-wheeler model to design a navigation controller based on the feedforward of the heading and the feedback of the position deviation.The feedforward control part is used to eliminate the heading point heading angle deviation in advance and reduce the cumulative deviation,while the feedback control part is used to make up for the inaccuracy of feedforward part by monitoring radial deviation and other unmeasured interference factors.Finally,I used MATLAB and its component Simulink as the platform to carry out the simulation and analysis of the steering control method,which proved the feasibility and superiority of this control algorithm.(5)Based on ROS(Robot Operating System),I built the overall architecture of the main controller software for automatic driving operations of tractors.I used NVIDIA Jetson TX2 as the hardware platform to design and develop nodes such as information acquisition,path tracking control,system communication and data interaction.The tractor’s expected front wheel steering angle is also adopted.Finally,applying the above main controller,GNSS positioning board,attitude measurement system,SPC-STW-2612 CM motion controller,electric steering wheel and other hardware,I built a tractor automatic driving operating system to achieve steering control.(6)The field test was carried out with Lovol Tx1104-D tractor as the test platform,mainly including the field operation route tracking control performance test and the hangar to field transportation route tracking control performance test.As is showed in the test,the method and system proposed and developed in this paper can realize the tracking control of the tractor’s movement route from the hangar to the field and the path tracking control under the conditions of field rotary tillage.During the tracking process of the moving route,the tractor has stable speed and steering process,coupled with small deviation of position and course.During the tracking of the straight line of field operations,the maximum error is0.062 m,the average error is 0.0233 m,and the standard deviation is 0.0213 m,which meets the needs of operation.