Research On The Problem Of Path Planning For Mobile Robot Based On Improved Ant Colony Algorithms

The path planning for mobile robot plays an important role in the field of robotic. It has the features of complexity, binding and nonlinear. Over the years research on the path planning is in full swing, the existing optimization algorithms due to the complex background environment with the defects of poor robustness, non-optimal and low efficiency. In view of the advantages of the ACO in solving combinatorial optimization problems and its own characteristics, comply with the intellective and bionics tendency of the field's development, this article designs two new ACO and applicats them in solving the problem of path planning for mobile robot successfully with the expected results. The main research work is as follows:Firstly,2D and 3D grids environment were modeled with the grids method. For the defaults of traditional evaluation function on feasible path, with sufficient analysis about the different environment characteristics, respectively establishing the evaluation functions which are applicable to static and dynamic environment.Secondly, with the large experimental analysis about the parameters of the algorithm, parameters on algorithm performance and the range of them were got, as a guide for later algorithm design. The convergence analysis for the basic ACO based on the shortest path problem was given at the same time.Thirdly, for the disadvantages of basic ACO as dis-convergence and finding the local solution, from the angles about the algorithm itself and hybrid algorithm, fuzzy adaptive window ACO (FWACO) and differential evolution chaos ACO (DECACO) were designed. The standard TSP proved that the ability of global optimization and convergence rate of the designed methods get significant improvement.Finally, FWACO and DECACO were applicated in the path planning for mobile robot. A large of emulation studies in 2D and 3D barricade environment were realized by MATLAB with good results. For the complex environment with many obstacles, the improved methods have strong robustness, ability of global optimization, fast convergence, moderate complexity algorithm and high efficiency, and they have respective strengths for the different barricade environments at the same time. The improved methods can complete the task about path planning in complex background environment with many obstacles better and faster.