The Research Of Picking Robot Multi-objective Path Planning Based On Improved Ant Colony Algorithm

The path planning algorithm is a core technology of the picking ball robot to search the effective path. At present, it has become a hot research topic in robotics. There are a lot of path planning algorithms, but for the solution of multi-objective path planning problem, each algorithm has different defects. Although the ant colony algorithm has its own deficiencies, it has parallelism and positive feedback, and easy to combines with other intelligent algorithms. Relatively, it is more suitable for solving multi-objective path planning problem. This paper will study the problem of ant colony algorithm. The main work is as follows:(1) Designing an intelligent picking ball robot and modeling the picking environment, which contributes to program and implement remote monitoring terminal for the work of picking ball robot, and builds the hardware platform for the subsequent algorithm experiment. If picking robot working in the tennis court, it has access to obtain environmental information through the top of visual sensor and processes the information data directly. After completing all the environment data, the method uses grid method for dividing the environment, which the grid size can automatically adjust according to the actual need and the ability to deal. This paper will mark on each grid, and only considering the information of the robot in a 2D space on XY.(2) Researching a kind of improved ant colony algorithm for picking robot multi-target path planning, using MFC programming to achieve the simulation software of path planning algorithm, and completing the algorithm experiment in the simulation software. In order to solve the slow search speed of ant colony algorithm and easy to fall into local optimum problem, through in-depth analysis of the ant colony algorithm searching path mechanism in the process of multi-objective path planning, thus, it can be seen that it has different needs of the algorithm parameters in different search stages. By considering the algorithm parameter contributing to search performance, the paper puts forward an improved ant colony algorithm according to the adaptive iteration algorithm parameters. Therefore, the algorithm can plan ideal path rapidly and efficiently.(3) Aiming at the problem of slow search speed, and based on the improved ant colony algorithm, the paper puts forward a kind of fusion algorithm which is suitable for picking robot multi-target path planning, and completing the algorithm experiment in the simulation software. Ant colony algorithm depends on pheromone accumulation and renewal to search path, all path pheromone are equal can't reflect the diversity of solution in the initial stage, which leads to slow the convergence speed, thus, the planned path is often not optimal or near optimal. The genetic algorithm can rapidly establish the initial path information in the initial searching time, and increase the diversity of solutions, but it ignores the feedback information, and leads to massive redundancy iteration and low efficiency. Based on this, Firstly, using the fast random search of genetic algorithm to get initial path information, then making full use of the advantages of the improved ant colony algorithm, by fusing two algorithms, it gets a fusion algorithm which further improves the path search efficiency.