Research On Path Planning Of Agricultural Robots Based On Extended A* And Improved Apf Algorithms

Based on the rapid development of agricultural technology,relying on mobile robots to effectively improve the efficiency of agricultural operations is an inevitable development direction.The path planning problem of mobile robots is again an important part of its technological research.The meaning of path planning is the optimal path planning problem of the robot,which is to seek an optimal route from the starting point to the end point to ensure safe and effective avoidance of obstacles in a real operating environment according to the lowest cost of work,the least travel time,and the best path to travel.In this thesis,the experimental orchard-based bi-directional A* and APF algorithms are used and optimized to achieve a relatively smooth path planning exploration for agricultural robots.The key inquiry essence of this thesis includes the following aspects.A vector e-map of the orchard environment based on ARCGIS was developed.Based on the specific situation of the orchard,the article investigates the production method of vector e-map under the ARCGIS platform from three aspects: general and detailed design of the map,coordinate system conversion and vectorization of the electronic base map.In static environments,the bi-directional A* algorithm is based on the shortcomings of useless interval search,long pathfinding time and more nodes.In this paper,we address these shortcomings and introduce corresponding strategies for improvement.In addition,the shortest path is smoothed using the cubic spline interpolation method.Finally,a static simulation environment is designed for comparison,and the results show that the paths planned by cubic spline interpolation combined with EB-A* algorithm are shorter and less time-consuming than those planned by bidirectional A* algorithm,and the smoothed paths are more conducive to the safe driving of the robot.Then it is an improved APF algorithm is proposed in a dynamic environment.Firstly,the basic principle of APF is explained and the shortcomings of APF are dissected.This is followed by solving the deficiencies of the algorithm by correcting the potential field so that the robot cannot reach the target point accurately,and by adding the escape force to solve the local minimum deficiency encountered by the robot during the exercise process.Finally,simulation experiments are conducted in matlab,and the experimental results verify the effectiveness of the improved APF in planning paths in dynamic environments.The improved APF can both make full use of the existing advantages of its own technology and overcome its shortcomings to supply a practical and implementable path planning algorithm for robot operation in real environments.A solid foundation is laid for the next step.The final result is a hybrid application of algorithms in dynamic and static environments for path planning of an agricultural robot,in which the hybrid path planning: a global path is planned using cubic spline interpolation combined with EB-A* global path planning algorithm,and a modified APF to avoid unexpected situations in the orchard in real time:agricultural machinery,livestock,etc.An experimental platform for the agricultural robot was established and also calibrated for algorithm feasibility.