Research On Spatiotemporal Association Map Building For Mobile Robot In Dynamic Environment

Map building for mobile robots is a key to implementing the navigation of intelligent mobile robots, and an essential foundation when mobile robots engage path planning, environment exploration and autonomous localization. Conventional map building methods are based on the assumption of static environments. However, most working environments of mobile robots in practical applications are constantly changing. So research on map building in dynamic environments is of important pragmatic significance for the navigation of intelligent mobile robots.This thesis is supported by the key project of National Natural Science Foundation of China under grant no.60234030. Research on Theories and Methods of Navigation Control for Mobile Robots in Unknown Environments. The states of the working environments of mobile robots are constantly changing over time due to presence of moving obstacles. The association of spatial and temporal properties of obstacles can better represent the characteristics of the dynamically changing environments, so the precision of maps built for mobile robots can be improved. MORCS-1 developed by Center of Intelligent System and Software of Central South University as a prototype system, spatiotemporal map building for a mobile robot in two-dimensional dynamic unknown environments is studied with a focus on the following three aspects:1. A real-time method of identifying dynamic and static obstacles on the basis of spatiotemporal association property in dynamic environment is proposed. A novel obstacle representation which associates the spatial and temporal properties of obstacles is introduced to detect dynamic and stationary obstacles fast and effectively. Sensor readings in different scanning time are transformed into the same world reference frame. A mobile robot will be able to identify dynamic and static obstacles by analyzing the spatial and temporal properties of obstacles. A great deal of storage and computational time will be saved because obstacles don't have to be mapped in grid map. 2. A real time approach to building spatiotemporal association map for a mobile robot in dynamic environments is presented. The method of maintaining dynamic and static grid maps respectively to build maps in dynamic environments leads to imprecise maps because of not taking localization error into consideration, and it is time-consuming when updating the global grid map. Three steps are taken to improve the method. First, a spatiotemporal map representation is introduced. Second, the eight-neighborhood scrolling window method is adopted to handle the drift error. Third, the time complexity of updating global grid map is greatly reduced using selective update mechanism. Spatiotemporal association map building can improve the precision of environmental map when a mobile robot works in dynamic environments, and the capability to survive in dynamic unknown environments for a mobile robot is also improved.3. A software system which integrates dynamic and static obstacle detection and map building in dynamic environment is designed and implemented based on the experimental platform MORCS-1.Ample experiments have demonstrated the effectiveness and reliability of the proposed methods.