A Research On Mathod Of Path Planning For Intelligent Mobile Robot

The path planning of intelligent mobile robot is the basis to complete its task. In the thesis, we researched on Grid method and PRM which are used most commonly in practical applications, designed and implemented these methods and compared them on the basis of experiments, then optimized PRM.Firstly, analyzed the basic principles of grid method for path planning, classify and summarize their characteristics according to the description of its environmental information is accurate or approximate description. Also detailed the basic principles of several common path search algorithm (greedy algorithm, Djikstra algorithm, depth-first algorithms, breadth-first algorithm and the heuristic search algorithm), then introduced design and implementation of the grid method of planning in detail presentation, including the design of modeling environment, grid size selecting, the choice of the grid, grid information coding, and the design of path searching algorithm and conduct a software simulation to verify the correctness and effectiveness of design of algorithm in this paper .Analyzed the basic principles of PRM path planning algorithm, introduced design and implementation of PRM path planning algorithm in detail presentation, mainly including the probability of the map building process , the sampling strategy and methods for testing the rationality, finally a software simulation experiment designed to validate the correctness and effectiveness of the PRM algorithm.Based on using the grid method and PRM path planning method in the design and implementation respectively, through several sets of simulation experiments, two planning methods were analyzed. Firstly, sum up that the two kinds of methods are path planning methods based on graph theory, change these environmental information in to geometric form, converted into a graph problem, used graphs to model the environment. Then in the same environmental conditions of planning, the two kinds of planning methods were analyzed by the cost in time and space consumed (search time, storage space, path quality, etc.).Two methods'environmental adaptability were compared and analyzed under the premise of occupying the same storage memory and the path searching time. Finally through the summary of the assessing standard of the path planning algorithm, PRM method and the grid method were compared find the respective advantages and disadvantages of the two algorithms, and pointed out the necessity of the optimizing of PRM algorithm.Finally, optimized the PRM algorithm, obtained that the distribution of sampling points is the main effect of the success and ideals of PRM planning algorithm through experiments, so a good sampling strategy is essential. Describes the dispersion of sampling sites (low dispersion and low differences sampling rate of sampling), pointed out the validity of pseudo-random sampling, as well as it is very important to alternate use these two methods in motion planning. Used the pseudo-random sampling strategy which meet the low-dispersion rate and a low difference between certainty and completely random to ensure the uniformity of the distribution of sampling points. Against the connected region which is not connected in the free space of the figure, and enhanced the nodes of roadmap, toned up the connectivity of the map, advance the optimization of the path by using strategy of nodes enhancing based on elliptic. Finally, the effectiveness is verified by the simulation and experiment of the optimization strategy.