Research Of Path Planning Method Oriented To Real-Used For Multi-Task Robot In The Supermarket

With the continuous improvement of people's living standard, the size and number of the supermarket are increasing, the service and management of the supermarket also need to significantly improve, the existing services quality and work efficiency of supermarket system need to further improve, therefore, it will become an important research topic to the service robot go into the supermarket, the research and development on the supermarket robot system makes robot be able to finish shopping guide, floor-cleaning and other functions in the supermarket. This topic is proposed under the background, the aim is to solve the supermarket path planning of shopping guide and floor-cleaning to improve the work efficiency and service quality of the supermarket.Firstly, this paper conducts a depth research on robot floor-cleaning and shopping guide at the supermarket environment, it based on multi-level hierarchical, region segmentation, general topology, characteristics, knowledge utilization, making use deliberating and behavior evolution mixed system structure to complete the area, the shelf and commodity modeling in a semantic level and geometric level.Secondly, this thesis researched deeply vertex-viewed method, and applied to the robotic global path planning, the use of the vertex-viewed method can be combined with A* algorithm to finish robotic global path planning of shopping guide, through simulation comparison shows that this method more real-time and adaptability than the global free space method.Thirdly, it analyzed the problems that the floor-cleaning behavior planning faces in supermarket and optimized the region segmentation, sub-region internal walking style, and operation order between the sub-regions to complete to optimized floor-cleaning task in the whole region with obstacle. The minimum time and the shortest distance are used as planning constraints, and it uses the sub-region internal walking style optimization method based on region segmentation including obstacles and operation order between the sub-regions to obtain the optimum result.Finally, the design and functional modules of experimental system have been elaborated in detail, and the experimental results of each module have been given. At last, path planning has been implemented on simulation platform.