Study On The Path Planning And Trajectory Tracking Of The Coal Mine Detecting And Rescuing Robot

China is the largest country of the coal production and consumption in the world. However, because of poor conditions of the coal mining system, large number of highly-gas mine and lack of effective techniques and management, especially continuous increasing demand for the coal, accidents such as gas explosion, water inflow and ignition have happened frequently in recent years. Heavy casualties and negative social effects have made healthy development of the coal industry impossible. Therefore it is premise and necessary to develop the coal mine detecting and rescuing robot. Under the circumstance the causes and field conditions of the coal mine accidents were unknown, it could substitute the rescue staff to enter the accident area for implementation of detecting and rescuing, and information can be transferred real-timely to the rescue command center simultaneously. It can provide scientific basis for the rescue decision, and the rescue plan can be made as rapidly and accurately as possible. Therefore it has critical significance for safety production in coal mine and can reduce loss of the public property and lives. This dissertation focuses on the coal mine detecting and rescuing robot as research object, key techniques in robot autonomous navigation—path planning and trajectory tracking based on the design and analysis of its movement device were thoroughly and deeply studied. At first, the key techniques in autonomous mobile robot were summarized comprehensively and systematically, especially the research contents, existing problems and development trend of path planning and trajectory tracking were reviewed, performance differences in various algorithms were summarized and further research was developed. The main research contents included are as follows:A kind of movement device aiming at special environment of the coal mine was proposed. The motion system, crossing-capacity and structure of the coal mine detecting and rescuing robot were redesigned and further improved. Using land surface mechanics theory and with reference to the principles of tracked vehicle, the kinematic and dynamic model of the robot were built according to the features of differential steering, which provided a basis for subsequent research on trajectory tracking of the coal mine detecting and rescuing robot.Aiming at the features of non-structure environment in the coal mine, the hierarchical control system consisting of upper and lower computers was designed and a system architecture of layered cooperation and planning was proposed. Considering the particular application environment and demand, the control system was mainly composed of decision system (including monitor module of upper and lower computers, control module of the lower computer and sense and detection module, etc.), data communication system and motion control system. Then system architecture divided into such layers as decision layer, deliberation layer and control layer was proposed, which can meet the requirement of the navigation's complexity and uncertainty when the robot is under an unknown environment.Combining with advantages of the representative optimization algorithm such as genetic algorithm and ant algorithm, improvements were made respectively. Firstly, an improved genetic algorithm was designed with the height variation of underground were fully considered. Three-dimensional workspace was modeled by grid method, and fitness function design used path length and energy consumption as the criterion. The initial population was generated by Variable length coding and random guided searching strategy to ensure the generation of the obstacle-free path in the initial period, meanwhile, crossover operator and mutation operator were optimized to solve the problems of the simple genetic algorithm such as premature phenomena and slow convergence, thus make the whole planning process more simple and effective. Secondly, an improved ant algorithm based on Max—Min Ant system(MMAS) was proposed. Its heuristic Function, selection rules of the path and the regeneration methods of pheromone were all improved, so that the convergence speed was increased and the performance of the path planning was improved. Finally the experimental results show that the performance of improved algorithms were effective and feasible.A design method of fuzzy self-adaptive controller used in robot trajectory tracking was proposed. Firstly, the controller of the robot trajectory tracking control system was designed by Lyapunov direct method, which can control the process from current position to target point by the expected speed. When the reference speed was unequal to zero, Barbalat lemma was used to prove the global consistent asymptotic stability of the system. Furthermore, a fuzzy self-adaptive controller was designed on this basis, which solved problems such as the rules increased with the numbers of system's variables, and efficiency of the controller was greatly improved, fuzzy control logic also became more clear. The experimental results show these two trajectory tracking controller designed in this dissertation has excellent tracking effect for different initial errors and reference speeds.