Adaptive network based fuzzy control of a dynamic biped walking robot

In this paper, we proposed a adaptive-network-based fuzzy inference system (ANFIS) control strategy based on a hierarchy of walking planning level, gait generating level and joint control level, which do not require detailed kinematics or dynamic biped models. The ANFIS controller, which enhances Sugeno fuzzy controller with self-learning capability from adaptive network, can combine the qualitative knowledge in fuzzy rules and be fine-tuned by online learning. The effectiveness of the proposed ANFIS joint control was verified through a 5-link biped robot simulation. We demonstrated that the designed hierarchical control system can use the experimental input-output data pairs for the biped robot learning and walking with dynamic balance. It is also shown that the further online self-learning capability of the ANFIS controller can markedly improve the dynamic walking performance of the biped robot.