Research On Electric Tractor Navigation System In Facility Greenhouses

China’s facilities greenhouse space closed,high temperature and high humidity,the exhaust gas generated during the operation of the human health and crop growth has a certain degree of harm.At present,China’s facilities greenhouse machinery intelligence level is low,the lack of special miniaturized operating machinery,with the scale of China’s facility agriculture is increasing,the intelligent demand for facility greenhouse machinery is increasing.The application of navigation technology to facility greenhouse operation machinery can improve the level of automation and intelligence of facility production,reduce the demand for labor in facility agriculture production,which is of great significance to ensure the sustainable development of facility agriculture in China.This study addressed the problems such as the low degree of intelligence of the facility greenhouse machinery,and on the basis of summarizing and analyzing the current situation of domestic and foreign research,the research of navigation system algorithm was carried out with the electric tractor as the platform and laser navigation.The main work and conclusions were as follows:(1)Aiming at the problem of less research on the existing small-scale operation machinery in the facility greenhouse,the electric tractor was used as the research platform,and LIDAR was used to sense the environmental information of the facility greenhouse.The navigation system can realize the monitoring of tractor working status,environment building map and the release of navigation target point instruction through PC remote and tractor onboard display.The navigation hardware system mainly consists of environment sensing module,control module,communication module,motion module and power module.The software system consists of application layer,data processing layer and execution layer.The control system was an industrial control computer,which realized communication with PC through wifi,communication with on-board display through VGA,and sended speed control information to motor controller through CAN communication.(2)For the problem of map construction and positioning of electric tractor in the facility greenhouse,raster map was selected to build the facility greenhouse environment map;by comparing different SLAM(Simultaneous Localization And Mapping)schemes,Gmapping algorithm was used to build the facility greenhouse environment map;The AMCL(Adaptive Monte Carlo Localization)algorithm was used for tractor localization.The experiment showed that the Gmapping algorithm established a raster map with high consistency and accuracy with the environment of the facility greenhouse,the particle convergence was high,the AMCL algorithm had good localization effect,the maximum lateral deviation of the built map was 6cm,and the maximum vertical deviation was 16 cm,which could meet the map accuracy requirements required for tractor navigation in the facility greenhouse.(3)To address the problems of unsmooth path and long planning time in the process of autonomous tractor navigation,a path planning method for electric tractors in facility greenhouses integrating the improved A~* algorithm and the improved DWA algorithm was proposed.The heuristic function was improved by setting the weight coefficients,the key point selection strategy was used to reduce the path turning points,the Bessel curve was introduced for path smoothing,and the path key points of the improved A~* algorithm were used as the path guidance points of the DWA algorithm.Simulation and comparison experiments were conducted for Dijkstra algorithm,RRT(Rapid-expansion Random Trees)algorithm,A~*algorithm and improved A~* algorithm in five different raster maps,and the experiments showed that the improved A~* algorithm had the least operation time and the smallest path length,which was beneficial to the tractor to execute the path trajectory to the target point.Finally,simulation experiments of the navigation system were carried out,which showed that the fusion algorithm could plan an optimal path and the tractor could successfully complete the obstacle avoidance task.(4)Based on the design and construction of the navigation system,the study of map building and positioning methods,and the study of the path planning algorithm of the navigation system,navigation field experiments were conducted in the laboratory-built scenario and the facility greenhouse,respectively.Firstly,straight-line driving and ground turn tests were conducted.The tests showed that the route planned by the improved A~* algorithm was smoother,and the tractor could achieve straight-line driving and ground turn smoothly.The maximum lateral deviation in the two environments was 14.30 cm and 11.20 cm,and the maximum heading deviation was 14.80 ° and 13 °,respectively,at a tractor speed of 1 m/s.Finally,the obstacle avoidance test was carried out,and the test showed that the fusion algorithm can successfully complete the obstacle avoidance task on the basis of ensuring the global optimal path,which could meet the accuracy requirements of the autonomous navigation operation of the electric tractor in the facility greenhouse.