Fuzzy Control with Central Pattern Generators for the Locomotion of Quadruped Robotic Systems

The development with CPG of the locomotion of a quadruped robot with three degrees of freedom per leg and its management by fuzzy control is presented. The robot, simulated in Gazebo-ROS, has three proximity sensors, an IMU sensor, and an odometry sensor. The purpose of the robot is the execution of locomotion in different types of movement. To reproduce the motor function of locomotion the central pattern generators use a novel mathematical model. The robot’s environment is a labyrinth on a flat surface. A three-level control architecture operates the robot’s locomotion. At levels I and II, the motor function of locomotion is reproduced with central pattern generators. Level III plans locomotion movement. This level has parallel control of both obstacle avoidance with a fuzzy controller and robot stability with a central pattern generator. The priority in motion planning is the stability of the robot. The proper functioning of the robot control architecture is demonstrated.