Research On Safe Route Planning Of Factory AGV Based On Improved Genetic Algorithm

Automated Guided Vehicle(AGV)has become one of the important tools in the industrial production,Automatic storage and port transportation system.The Path Planning is one of the key technologies to study the AGV system.That is,in the map of known environmental information,the shortest route around the obstacles to reach the target location is planned.In this paper,the minimum collision probability and the shortest distance are studied,and a path planning algorithm based on improved Dijkstra algorithm and improved genetic algorithm is proposed.The main research contents of this paper are as follows:(1)Aiming at the deficiency of traditional grid map construction which only contains obstacle information,an improved grid map construction method based on fusion of collision probability of AGV is proposed.By analyzing the size and movement state of two-wheel differential-driven AGV,a calculation method of Global Grid Collision Probability is designed and integrated into the traditional grid map.(2)Aiming at the inefficiency of traditional Dijkstra algorithm,a cross-layer search Dijkstra algorithm is proposed.A method of searching k-order adjacent grids is designed,which adopts the cross-layer searching method in a large area of free grids and layer-by-layer searching method in an area near the edge of obstacles at the same time,the pre-processing method is used to further improve the computational efficiency.Finally,the simulation results show that the improved Dijkstra algorithm can improve the efficiency of 41% at most.(3)Aiming at the problem of path safety and shortest path,a safe path planning algorithm based on improved genetic algorithm is proposed.Based on the grid map of collision probability,using the path planned by cross-layer search Dijkstra algorithm as the baseline path,the improved genetic algorithm is used to optimize the baseline path.Finally,an optimal path with short distance is obtained under the premise that the collision probability is the lowest.The fitness function combining the collision probability of the path nodes and the node offset is constructed,and the coding method based on the node offset is designed.At last,we use different path planning algorithms in the map of complex environment.The results show that the collision probability can be reduced by 47.3%.(4)According to the need,the hardware system of modular AGV platform is designed and built based on ROS System,and the software of AGV platform including grid map construction,AGV location,AGV path planning and drive control is developed.The validity and practicability of the improved algorithm are verified in both simulated and real environment.