Fuzzy control of a quadruped robot foot trajectory

Abstract

The progressive deployment of robotic industry and the capabilities of quadruped robots led us to introduce an approach to use fuzzy finite state machine for determining gait parameters of a quadruped robot. In this method, each step of robot- in a trot gait- is represented by two parameters: the maximum length (L) and height (H) of the foot tip in a step. Generated fuzzy rules were used for calculating those parameters by which the robot could pass unknown routes. To apply this algorithm, a simple robot is simulated in WEBOTS and finite surfaces with different slopes and friction coefficients were designed. After obtaining suitable parameters by which the robot could pass the surfaces, the rule base was generated. By implementing the resulted fuzzy if-then rules, the robot could pass three distinct unknown routes. The results convinced us that fuzzy approach can make the control of quadruped robot easier and further experiments can be performed to give the ability of passing rougher terrains to a quadruped robot which may decrease the complexity and cost of the robot.