Robotic soccer games represent the most significant form of research in artificial intelligence. Using the humanoid soccer robot’s basic movement and strategic actions, the robot takes part in a dynamic and unpredictable contest and must recognize its own position on the field at all times. Therefore, the localization system for the soccer robot represents the key technology for improving its performance. This work proposes a new approach for self-localization, an Adaptive Vision-Based Self-Localization System (AVBSLS), which allows the humanoid robot to integrate the information from the pan/tilt motors and a single camera to achieve self-localization. The proposed approach uses a measuring artificial neural network technique to adjust the position of the humanoid robot. A systematic method to measure the intrinsic parameters is proposed for the CCD camera adjustment. Using this approach, any type of CCD camera can be used to precisely calculate the robot’s position. The experimental results indicate that the average accuracy of the localization is 92.3% for a frame rate of 15 frames per second (FPS).
Relation:
International Journal of Innovative Computing, Information and Control 9(3), pp.991-1012
