Vision-Based Kinematic Configuration Recognition for Re-configurable Modular Robots

摘要

Re-configurable modular robot systems are advantageous for their scalability and versatility. However, creating kinematic models for every new assembly is a time-consuming and error-prone task. In this paper, a vision-based kinematic configuration recognition method is proposed. The modular graph is firstly extracted by recognition of Quick Response (QR) Codes placed on each module. String information, obtained by searching kinematic chains on the modular graph, is then compared with that in a preset database, to figure out the current configuration. Experiments are conducted to verify the effectiveness of the proposed method, and to test the accuracy of the recognition. Results show that the intersection angle between the camera and the QR codes should be less than 45 degrees, to achieve the best recognition effect.