Research On Mobile Robot Obstacle Avoidance Based On Optical Flow Method

Mobile robot autonomous obstacle avoidance function is an essential function of the robot in the unknown environment.Vision is the main method to get the scene information in scientific research.Visual sensor can acquire more obstacle avoidance information than other methods in robot obstacle avoiding.Robot obstacle avoidance system based on optical flow method is a new hot spot in visual method.To realize the self-avoiding function and improve the accuracy of obstacle avoidance,the robot obstacle avoidance system will be designed based on the HS(Horn-Schunck)method with wheeled model robot as the experimental platform.The validity of this system is verified by the obstacle avoidance test in real environment.The main research work includes:firstly,image sequences are processed by image processing using image gray scale,image filtering and image scaling then using HS,LK and improved HS algorithm calculate and analyze optical flow in image sequences.FOE and TTC are calculated in optical flow field.Robot's moving strategy is designed according to the change of TTC.Then obstacle avoidance system of robot is designed and the obstacle avoidance experiment platform is established.The instruction transmission of the Sabertooth motor drive board from the notebook computer to the robot is realized through the serial communication.The obstacle avoidance experiment of the robot is designed in real environment,obstacle avoidance data of the robot is located and recorded by the Opti Track system and the effectiveness of the system is verified.Meanwhile,compared with obstacle avoidance effect of the improved HS algorithm then verify validity of improved method.It is proposed to establish the relationship between the rotation angle of the robot and the optical flow vector model so that the robot can rotate the corresponding angle in different range to optimize strategy.Then establishing relationship between weight coefficient and the obstacle avoidance effect of the robot to calculate the range of the weight coefficient corresponding to the optimal obstacle avoidance effect.