Research On Vision System And Low-Level Motion Control Algorithm For Soccer Robot

Soccer robot has become one of the hot subjects in the field of robot researching, it has referred to many kinds of advanced research from different fields, and it is a research platform which is full of high technologies as challenge .As a branch of the machine vision, soccer robot vision system is one of the crucial parts in soccer-robot system, because it provides the exact position information for strategy subsystem and it is the foundation of the whole system control. A robot vision system contains the following processing: the obtaining of digital image, image pre-processing, image segmentation and object identification.Taking the Mirosot robot 3vs3 as study object, the paper made a deep study on vision system and low-level motion control algorithm. The study mainly lies in the following aspects:1. The paper analyze the several color space models normally used in vision system, and compare with the strong point and flaw of different color space models. RGB color model used to process image can reduce the calculation in the system, and improve real time of vision system. But the illumination influence is severity for vision system. YUV color space model can minimized the disturbing of the change of illumination, so we take YUV constant threshold approach for accomplishing pixel classifying.2. The soccer robot's vision is a system based on color and multi-target tracking, and it process color video image. The data of image processing is very big, but system real time is an important require for soccer robot system. To save the time of image processing, we take searching in a dynamic window based on connectivity region labeling for object searching. Relative to search the whole image, this method can save approximately 98% time and meet the real time require. 3. Low-level motion control of soccer robot is important to robot soccer system, and it can provide a good basis for higher motion and the whole decision-making subsystem. The paper analyzes the lacks of tradition PID control algorithm, and presents combining fuzzy control and PID control in the low-level motion control of soccer robot. Simulation experiments indicate that this control method can exert respective merits of fuzzy control and PID control.4. In the traditional control, the path of soccer robot is out of control, so we present a path planning based on Bezier curve model in soccer robot path planning. Simulation experiments prove the feasibility of the proposed method.