Research On Humanoid Robot Vision Systems And Motion Planning System Of Shooting

The field of humanoid robots as an important branch of research in recent years has become the new darling of the robot, humanoid robot and the current application is hot robot soccer, it is a set of artificial intelligence, machine learning , communications, automation and many other fields of research and technology integration products, so the robot soccer game shows a country's level of technology in the field of computer and control from another angle. As the world's humanoid robot boom, some countries have also developed some of the more successful humanoid robot, like Honda's biped humanoid robot P2, P3 and enhanced version robot ASIMO, Sony the company's new entertainment humanoid robot SDR-3X, BIP2000 humanoid robot, the University of Michigan Rabbit, Technical University of Munich, Germany humanoid robot Johnnie, Tokyo University of biped robot H7, the U.S. developed SpringFlamingo MIT Stirling University and the United Kingdom launched RunBot are also very successful in humanoid robots. Relatively good domestic humanoid robot such as humanoid robot biped Tsinghua University THBIP-I, Harbin Institute of Technology Department of Computer Science Center for Multi-Agent Robot humanoid robot developed HIT-I, HIT-II and HIT-III , National Defense Science humanoid KDW-I KDW-II and KDW-III, Beijing Polytechnic University, a humanoid robot BHR-01.This paper will based on the hardware MF-ROBOT (Metal-Fighter) produced by MiniRobot cooperation and the Harbin Institute of Technology ,develop humanoid robot vision and motion planning system for the feet of the penalty, making the MF-ROBOT in the absence of the intervention, follow the contest rules, complete the penalty contest. Visual system as MF-ROBOT's eyes in the whole football game plays a vital role, through the visual system MF-ROBOT can distinguish ball, goal, teammates, the enemy, the results of the visual system go into the motion planning system generate action commands to control the MF-ROBOT's football action to complete the contest. Visual system is mainly through the camera calibration, color space selection, color histogram construction, image filtering algorithm, image segmentation algorithm, the target object classification algorithm, object localization algorithm use these algorithms for target object recognition. Shot motion planning system is the MF-ROBOT another penalty shoot-core system, the system will be designed for the MF-ROBOT to looking for the ball, near the ball, adjust the shooting angle, play and finished the contest, each the internal of state are composed of a complex strategy of action. MF-ROBOT put part of the data of visual system and results with its own current state, into the shooting motion planning system for analysis to determine the next step action MF-ROBOT need to complete and state need to go into.