Complete Coverage Path Planning Of Indoor Robots

Complete coverage path planning of indoor robots is a special type of path planning for mobile robots, which requires the robot path to cover reasonably and efficiently every part of the workspace except obstacles. It has a very wide range of applications in our daily life,such as window cleaning robot, cleaning robot, etc.. In this paper, the grid map was used to model the indoor environment. And the complete coverage path planning algorithm based on the biologically inspired neural network was studied. However, there was redundant computation in the complete coverage path planning algorithm based on the biologically inspired neural network. And the complete coverage path planning results were influenced by the robot movement speed. In order to solve these problems, biologically inspired neural network was improved by discretization. The performance of the complete coverage path planning algorithm based on the discrete biologically inspired neural network was simulated by MATLAB. Eventually the complete coverage path planning algorithm based on the discrete biologically inspired neural network was applied to the Voyager-II A robot. The feasibility of the algorithm was verified. In this paper, the main research contents and original results were described as follows.(1) Several common map models were compared and analyzed, and their advantages and disadvantages were pointed out. Combined with the characteristics of the biologically inspired neural network, the grid map model was used to model the indoor environment.Then, each grid corresponded to a neuron. In order to simulate the performance of complete coverage path planning algorithm, the grid maps of typical indoor environment and random obstacle environment were established. The influence of the number of obstacles on the path planning results was studied, and the number of obstacles in the worst results of the complete coverage path planning was given.(2) Based on the grid maps of typical indoor environment and random obstacle environment, the influence of the robot’s moving speed on the complete coverage path planning algorithm based on the biologically inspired neural network was studied. The simulation results showed that the algorithm achieved the best performance when the robot’s moving speed and the rising time of the activity value of the uncovered neuron were reciprocal.(3) There was redundant computation in the complete coverage path planningalgorithm based on the biologically inspired neural network, and the complete coverage path planning results were influenced by the robot movement speed. In order to solve these problems, biologically inspired neural network was improved by discretization. The stability of discrete biologically inspired neural networks was analyzed, and the maximum sampling period was given. The effects of sampling period on the complete coverage path planning algorithm based on discrete biologically inspired neural network was studied. The simulation results showed that the algorithm achieved the best performance when the sampling period was equal to the rising time of the activity value of the uncovered neuron.The optimal complete coverage path planning results of the biologically inspired neural network and the discrete biologically inspired neural network were compared, and the simulation results showed that the total length of the coverage path and the number of turns were close, and the simulation time of discrete biologically inspired neural network was obviously reduced. The discrete biologically inspired neural network could effectively reduce the amount of computation.(4) In this paper, the Voyager-II A robot was used as the establish platform of complete coverage path planning algorithm. Then infrared and ultrasonic sensors were used to sense the environment, and to build the grid map. The robot was located by the method of dead reckoning. Finally, the feasibility of the algorithm was verified on the Voyager-II A robot.