Research On Path Planning Of Mobile Robot Based On Second Division Of The Environment

Path planning of mobile robots is one of the most basic issues in robotics research, including environment modeling and path planning. The common grid method because of its simple calculation, easy to realize and maintenance, has been widely applied to model environment of the path planning, but the granularity of grid cells is difficult to control has failed to solve properly. Therefore a improved grid method is proposed. In the method, the size of grid cells in robot's field is divided again according to the information of obstacles in the environment, so as to solve this problem properly.Genetic algorithm has good robustness for the species problem, therefore has been successfully applied to solve the problem of global known path planning. However, the traditional genetic algorithm cause the initial population's ergodicity and feasibility poor in the search space, at the same time, the inherent defect of genetic algorithm is low convergence speed. Therefore this paper uses environmental second division method to provide excellent environment modeling for path search algorithm; a new method based on insurance matrix to generate initial population is proposed, it can improve initial population's ergodicity and feasibility in the search space; at the same time, insurance matrix as environmental knowledge to generate fitness function and related genetic operator, which overcomes the slow convergence speed of GA and improves the efficiency of path planning. The simulation show that robot can not only plan a optimal path but also avoid collision safety using the algorithm in the complicated environment, the convergence speed is improved more significantly than existing congener representative algorithms, the effect is very satisfactory.In the perception environment, the algorithm based on rolling window is a classic algorithm, but the planning strategy is the majority nodes deemed invalid nodes in local path planning, it can not only waste storage space but also influence planning real-time. The limitation of the robot vision domain may bring about planning stalled. Therefore, the algorithm based on high quality nodes and steer window is proposed. First this algorithm use environmental second division method to provide excellent environment modeling for path search algorithm, then forage quality value evaluates distance length, the safety degree, the possibility of trap obstacles. Furthmore, the new planning strategy is to choose high quality value's nodes realizing the reasonable combination the current optimal choice and the anticipation of trap obstacles, which ensures successful planning in the local path and solves the problem of poor effectiveness. Finally, this algorithm gives steer window's strategy to solve the problem of the planning stagnation. The simulation show that this algorithm is a effective method, the results are satisfactory.