| With the increasing popularity of tractor autopilot technology,it is gradually faced with major problems such as variable work terrain,low navigation work accuracy,construction of the turning paths,and optimization of the turning paths.This paper conducts in-depth research on these issues,builds a hardware platform,and proposes new lateral error assessment method,optimized global coverage path planning algorithm,and finally through field trial analysis,the main research content is as follows:(1)Set up the hardware platform,establish the coordinate system of tractor positioning,calculate the coordinates of the tractor's ground projection point based on RTK-DGPS and IMU acquisition data,and implement three lateral error assessment methods for manual measurement,2D-GPS measurement and 3D-GPS measurement;(2)According to the historical track of the tractor and the actual boundary data of the acquired farmland,optimizing the parameters of the boundary of the ground,calcullating the end point of the straight path,determining the optimal working direction,and selecting the linear path operation sequence;(3)Constructing a continuous curvature curve model for the non-working area of the boundary's head to form a global path planning algorithm for the working tractors;(4)In summary,the assessment method of lateral navigation error based on RTK-DGPS and IMU can truly reflect the lateral error of farmland on gentle slope,and the global path planning based on actual farmland boundary information,optimization of optimal operation direction and continuous curvature turning model.The design can fully automate the farmland operation process.The test results show that the root mean square error of the straight path lateral error is 7.05cm,and the root mean square error of the automatic turning path is 10.02cm,which can fully meet the tractor's fully automatic driving operation in the field.Thus,laid the foundation for real-time online path planning and design of farmland automatic driving operations. |
PDF Full Text Request