Research On Target Recognition And Motion Control Of Soccer

Soccer robot not only has a flexible, efficient autonomous movement,but also the ability to complete the task is very strong, with the rapid development of information technology,Soccer robot has become a hot topic. Research on soccer robot includes many fields, including machine learning, robotics, artificial intelligence, sensor technology, computer vision technology, multi-agent systems, image recognition.Vision systems and robot motion planning are the two research field about soccer robot, this paper is based on the RS-O-type robot, research and improvement the robot object recognition ability and motion control aspects.(1) The RS-0 type football machine comes with a camera, use the camera as the machine vision system.The system is more suitable HSI color space vision. In the pre-process image, uses median filtering denoising methods to de-noising the image, compared with the traditional method, this method can enhance the signal to noise ratio of the image, preserving the original image on edge and texture information.(2) In the target feature extraction and matching process, this paper is based on Scale Invariant Feature Transform SIFT operator, the traditional SIFT operator in the process of generating descriptors were optimized to improve the speed.Traditional SIFT match has low efficiency, match slow and other issues, proposed and implemented based on regional growth SIFT point matching method. This algorithm can basically unchanged the premise of matching reliability, significantly improved the matching speed.(3) In terms of soccer robot motion planning, design a new omni-directional wheel, its smooth movement and friction characteristics superior to the existing omni-directional wheel; In control of the movement of the mobile platform, this paper derive the model of the mobile platform, including the kinematic model, dynamic model, and control model based kinematic and dynamics model. According to the actual need simplified the model, then aimed a single motor shaft, used speed and current double closed loop control simulation, finally test speed cross-coupling experiments. Research results show that:the mobile platform has flexible movement ability, can quickly determine their own location, adjust its posture to achieve free movement in the plane.Experiments show that, by the above method used in the RS-O-type robot, the robot performance has been significantly improved, and was further verified by experimental in robot soccer.