A Path Planning Algorithm and Application Based on Improved Cellular Neural Network

For the two problems in mobile robot path planning algorithm based on neural network, namely it did not take the adjacent position with largest active value that would be nonunique into account and it was possible for mobile robot to collide with obstacles' edge when it went to the next position, we proposed an improved cellular neural network (CNN) algorithm. Meanwhile, to address the problem of complete coverage path planning (CCPP) for mowing robots that had the greatest coverage rates and the lowest repetitive rates, we proposed an algorithm that combined an improved CNN algorithm with back-propagation neural network algorithm based on the plowing path planning. We adopted grid method to model the environment and used Matlab2010b to simulate them. Simulation results show the improved algorithm can make mowing robots avoid obstacles, out of the dead zone in the shorter path and continue to clear cut the blind area. Meanwhile, they achieve the CCPP, improve the greatest coverage rates, and reduce the lowest repetitive rates.