Gait generation of a two-legged robot by using adaptive network based fuzzy logic control

In this paper, a control strategy is proposed for gait generation for a two-legged humanoid robot. The two-legged robot is assumed as a 3-dimensional robot with 5-links. Gait generation is performed by assuming motions in the saggital and lateral planes. Dynamic model of the robot is obtained by using MATLAB
^reg
/SimMechanics Toolbox. A fuzzy logic controller (FLC) is established for gait generation. The rule-base of the controller is optimized offline by using artificial neural network (NN). The neural networks are trained by using the reference joint trajectories obtained from clinical gait analysis (CGA) [13].