This work focuses on the stride-frequency and step-length autonomous adjustment in response to the environment perturbations. Reinforcement learning is assigned to supervise the stride-frequency. A simple momentum estimation further promised the adjustment. In the learning agent, a sorted action-choose table instructed the learning to find out the proper action in a straightforward way. Incorporating the step-length real-time adjustment mode, the biped is able to smoothly transit motions and walk adaptively to the environment. Dynamic simulation results showed that the supervision is effective. © 2007 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Yang, L., Chew, C. M., Poo, A. N., & Zielinska, T. (2007). Autonomous stride-frequency and step-length adjustment for bipedal walking control. Studies in Computational Intelligence, 76, 189–198. https://doi.org/10.1007/978-3-540-73424-6_22
Mendeley helps you to discover research relevant for your work.