| The thesis was supported by national basic research project "Research on Basis of Synergetic Task and Reconfiguration Techniques for Heterogeneous Multi-mobile Agents"(A1420060159). Related works was subject to the part "multi-robots synergetic task" of the project. The thesis focuses on the problems of multi-robots coverage based multi-robots architecture,robot communication link,robot map description,coverage cell decomposition,coverage cell allocation and coverage performance metrics. The objectives are exploring new effective methods of robot coverage based on multi-robots synergetic cooperation in designated environments, and improving the cooperation of multi-robots, advancing the efficiency of robot coverage, achieving optimal coverage at last.The architecture is absolutely necessary basic of multi-robots coverage. In the thesis, firstly a self-configuration strategy was loaded on the layered architecture of multi-robot. Then a self-configurable and layered architecture of multi-robot (SCLA) was presented. Command layer, coordinate layer and execute layer are the three layers of SCLA from up to down. Instruction communication and self-configuration strategy are adopted by SCLA to achieve synergic mission. The simulation results indicate that it is robust, fault-tolerant to environment and mission. Communication is the key technology in multi-robots coordination and cooperation, and also is the basis of SCLA. In order to deal with some circumstances such as multi-robots communicating failure, CAM,CSM and SSS are presented, which are three strategies of robots establish communication links on-demand. So, the SCLA can run in all cases.Robot coverage needs environmental knowledge. The organization and description of environmental knowledge are called robot map. It is more important to the robot coverage. So, the 4-D environmental map for coverage is researched in this thesis. 4-D environmental map are composed from 3-D space coordinates and 1-D characteristics of the environment. The 3-D space coordinates can be observed by the robot's distance sensor, while the 1-D characteristics of the environment can be observed by using multi-sensor information at the same time. After sampling 4-D coordinate data of points in environment, and storing as a fully binary tree structure, a 4-D environmental map is attained for a environment region. Then some map transform methods from 4-D environmental map are discussed separately. They are grid map generating by Krige interpolation, cell decomposition map generating by elevation mutation operator, and topological map generating by binary tree finding. These maps, especially the cell decomposition map, are always used in robot coverage by many researchers.Length shift tangent (LST) cell decomposition is researched based on 4-D environmental map. According to certain rules, points of tangency or intersection with obstacle area are selected as the key points to cell decomposition by sweeping the target area with a tangent that shifts length timely. In 4-D environmental map, these key points can be got by binary tree finding simply. Compared to Boustrophedon cell decomposition, which is regarded as an excellent coverage method, LST cell decomposition not only reduced the complexity of algorithm but also decreased the number of cells. All these make robots system allocating and planning more optimal with smaller cost. In the SCLA system, cell allocation strategies based on cell square estimating by approximately rectangular are studied separately for n≈m, n>>m and n<cs, time consumption ratioηT, energy consumption ratioηE and effective path ratioη?. They are used to estimate the performance quantitatively. Then the coverage coordination index CCI is proposed, specifically for multi-robots synergetic coverage. |
