Motion synchronization with predefined rhythms for humanoid robot

Abstract

In order to make a humanoid robot to resemble a human, individual behaviours are created based on simple primitive rules. These behaviours will be synchronized together to perform a more useful tasks. In this work, various robot behaviours that are required to play soccer using simple primitive rules are developed and these individual behaviours will be synchronized together so that the humanoid robot is able to play soccer autonomously. The behaviours created are separated into two categories namely object detection behaviours and motion behaviours. In object detection behaviours consists of Red Ball Detection which uses primitive rules of colour blob segmentation and depth estimation through trigonometric properties whereas Goal Post Detection uses the primitive rule of grid-based segmentation pattern learning and recognition. In motion behaviours consists of Scoring, Kip Up and Diving which uses the primitive rule of support polygon and centre of mass. Once these behaviours are designed and created, it is synchronized into two different roles namely player and keeper.