Vision-based Omni-directional Mobile Robot Path And Behavior Planning

Mobile robot is an intelligent result of fusing lots of subjects including automation, mechanism, communication, computer etc. It is the full combination of technology and human activities. The platform model of omni-directional mobile robot is established in this work. In sequence, the solutions for intelligent path planning (PP) are applied in the platform. In addition, according to the characteristics of robot vision, four basic behaviors including roaming, shooting and obstacle avoiding are achieved for mobile robots, which then lead to the obtaining of a set of reasonable behavior conversion schemes of mobile robots.The hardware platform of the robot system is analyzed in the first instance. The corresponding mathematic model is constituted based upon the omni-directional mobile frame. Moreover, the improved genetic algorithm (GA), namely adaptive F method is introduced to finish the design of path planning, improving the executing effect. The numerical simulations under the MATLAB and VC++ environment are carried out to testify the feasibility of the approach. The results demonstrate that F adaptive algorithm results in the better performances, realizing the obstacle avoidance in short term, compared with the conventional GA under the same conditions. The examinations are set up on the omni-directional mobile frame, and through the comparison to the artificial potential field method, the F adaptive algorithm presents the remarkable obstacle avoidance performances, overcoming the limitations of the grabbling under the objective surroundings with obstacles.According to the identified information measured from the vision system of robots, four basic behaviors are divided. They are tracking, roaming, shooting and obstacle avoiding. That is, the panoramic tracking is obtained by adopting polar coordinates technology, and the front tracking is achieved by T-style for fore-field of vision. For the roaming behavior, the motive fixed-curve roaming should be gained considering the locations of the obstacles. Besides, the shooting area is set according to the structures of the mobile robots, and two shooting modes, which consists of lob shooting and pushes shooting are realized by dividing the shooting grids of the fore-field camera. The behaviors mentioned above influence the determinant terms, which then lead to the behavior conversion scheme around the obstacle avoidance. After the practical examinations being implemented, the smooth behavior conversions are obtained as expected based upon the designs above.