Research On Regional Coverage Path Planning Method For Indoor Mobile Robot

Learning the models of human to deal with the uncertain information has become a key issue of intelligent mobile robots. In navigation, mobile robots can get effective environment information to locate themselves and establish reasonable models. Robots completing coverage path planning has been widely used. Currently, completing tasks with reasonable, efficient and adaptive methods in limit spaces is one of the main research interests in regional coverage.To realize mobile robots regional coverage of path planning. Firstly, the thesis processes the environmental information collected by the robots, which means that the cost map is obtained from rough and uncertain sensor data. Then, based on comprehensive and penetrating research and analysis of indoor mobile robots path planning and regional coverage technology, the coverage strategy for multi-region and complex environment is proposed.The main contents of this work can be summarized as follows:1. Establishing a description of an environment model for indoor mobile robots regional coverage which includes environmental feature detection and the cost map development. In the aspect of the environmental feature detection, the Adaptive Hough transform (AHT) method is proposed, which aims at reducing effects of the quantization unit of the parameter space on Hough transform(HT) in detecting line features. Firstly, the sample model is built up by using samples and computing parameters of the model. Then, according to changes of the model parameters and sample distribution, a method is established to get the appropriate quantization parameters. Finally, the optimized quantization unit is obtained and applied to feature extraction in a structured environment. In the aspect of cost map development, the cost of cells in priori static map is first determined, and then the static map is turned into a dynamic cost map, which can provide real-time environmental information for the robots.2. In multi-region and complex environment, based on the cost map, a multi-region coverage method is proposed based on the minimal tree. Firstly, the cost map is divided into several regions, which are decomposed into sub-regions using Boustrophedon cellular decomposition. Then, a graph is extracted according to the adjacency relationship between the sub-regions, in which each node represents a sub-region. The sub-region planning method based on the minimal tree is applied to the graph in order to generate a sub-region planning sequence. The method based on the Dijkstra algorithm is employed to plan a transfer path among different regions, and the back and forth coverage strategy is applied to every sub-region to complete the coverage. Finally, the proposed multi-region coverage method is applied to complete coverage in different environments on-line, and the experimental results show that our algorithm has advantages in improving the efficiency of coverage over traditional methods.3. Proposing an indoor coverage method for car-like mobile robots. As this kind of robots are limited to the turning radius and speed constraints, and the performance of GPS, which is mostly used in outdoor environment, is constrained in indoor, therefore a new method is demanded. The spiral shift method is combined with the back and forth strategy, which makes car-like mobile robots have the ability of turning and reversing The method can ensure the whole process of the vehicle to be reliable and consecutive.In view of regional coverage path planning method for indoor mobile robots, the proposed and compared method are applied in different environments, and results show that our algorithms have advantages in reducing repeat coverage and coverage time, which ensures the validity and universality of the proposed method.