Research On Complete Coverage Navigation Technology For Cleaning Robot

Service robots are slowly accessing to our daily life, and playing increasingly more important roles in family service, restaurant service, exhibition hall host and so on. Cleaning robots are most widely used to keep people away from heavy and boring physical laboring. It can be used for the ground cleaning either in domestic place or public place. According to the movement characteristics of cleaning robot, this paper focus on the complete coverage navigation and try to improve the movement efficiency. Localization, obstacle avoidance and complete coverage path planning were discussed in this paper. The main contents focused in this dissertation are listed as follows.First of all, the model of the working environment for service robot was introduced. The environment information is represented as the formation of grid map. In order to limit the navigation precision and meet the needs of the processing speed of coverage algorithm concurrently, double precision maps are adopted in this paper. The high accurate grid map obtained through laser sensor could be turned into a low accurate grid map. In order to improve the navigation precision, the localization and obstacle avoidance are executed in the high accurate grid map. During the cleaning process, the coverage status is marked in the low accurate map to improve the processing speed of the complete coverage path planning algorithm.Secondly, considering the complexity of working environment and the efficiency of the task, a novel obstacle avoidance system of cleaning robot was proposed. Obstacle were established by building rolling widows and categorized as the size of obstacle. In order to improve the coverage rate around obstacles, cleaning robot moved around close to the edge of small obstacles. Fuzzy control decision was used to avoid large obstacle. Only large obstacles are regarded as the boundary of the cells during cellular decomposition. The problem of complex path and high repeat rate due to the large number cells could be avoided.Then, on the basis of analyzing and comparison of the existing algorithms, this paper proposed an improvement complete coverage path planning algorithm. The environment was divided into a series of cells without repetition using cellular decomposition. The methods of path planning in subareas and the path connection between subareas are proposed. A method of cells connection that could reduce the repetition rate was put forward in this dissertation. In every cell, cleaning robot moved with a boustrophedon motion which is a sequence of simple back-and-forth motions until meeting a chokepoint. Computing the path connection parameters (the distance from chokepoint to uncovered cells’vertexes), the cell whose path connection parameter has the smallest value would be the next cell to be covered. This method could decrease the length of repetitive path and reduce the repetition rate. In this way, the coverage efficiency of the cleaning robot is improved.Finally, the complete coverage path plan method proposed in this paper was compared with other algorithms through the simulation experiment. The efficient of this method was verified. The software system based on ROS (Robot Operation System) has been designed. The experiment in real environment was made on the modified Turtlebot robot platform. The practicability and effectiveness of the coverage navigation method proposed in this paper was proved in the experiment.