基于神经网络的机器人路径规划仿真系统的设计与实现-软件工程专业论文.docxVIP

哈尔滨工业大学工程硕士学位论文 哈尔滨工业大学工程硕士学位论文 - - II - Abstract The application of robotics and the development of robots in many areas have been applied, entities simulation in robot research is also more expensive. when the robotics structure is completed, it will be difficult to change, And some special tasks and environment can not been experimented. All of these requires the adoption of the computer for simulation. In the simulation process, robotics architecture can arbitrarily change to achieve various tasks. Through a virtual simulation platform, not only to improve efficiency, reduce the research and development cycles, but also enable the pilot study from the environment and climate impact, and ensure safety. Meanwhile, robot path planning is an important robotics technology applications, using good robot path planning techniques can save a lot of time which spend on operating robots, reduce wear and tear, but also can save human resources, reducing capital investment. for robot applications in a variety of industries to lay a good foundation. In this paper, neural network control and planning trails combined to form the new intelligent path planning methods, strive to improve robot path planning avoiding obstacles precision, and speed up the pace of planning to meet the needs of practical applications, in order to do something meaningful to try and explore the robot simulation and mobile robot path planning. This paper based on the actual situation, the road map plan was turned into two situations: a detailed road map of trails planning and no detailed maps . For a detailed map of trails, in connection with the use of grid reverse dissemination algorithm trails planning. For a No detailed maps of trails, use a method of approaching energy separated by the negative gradient direction, the path is set to determine the trend of the campaign. The application of neural network methods to tap the mobile robot dynamic path planning solutions to the potential. These two p