Design and control of a quadruped robot walking in unstructured terrain

Abstract

All quadrupeds have to deal with the same challenge of motion generation. Here, a general architecture of a motion generator for statically stable gaits is proposed, and, because of its modular structure, it can be easily adapted to the needs of various four legged walking machines. Encapsulated modules are arranged in a hierarchical structure, where each one carries out a certain task and communicates with the connected modules. The structure can be divided into four main parts: interface to the robot, motion generation, model-based controller and sensor data processing, which uses fuzzy inference to carry out the obstacle avoidance task.