A hybrid walk controller for resource-constrained humanoid robots

摘要: Zero moment point (ZMP) preview controller is a widely adopted method for bipedal locomotion. However, robots which are resource constrained or working in dynamic environments, simple reactive walk controllers still favored as ZMP have more control latency and computationally demanding. In this work, we present hybrid that dynamically switches between based on the analytic solution of linear inverted pendulum model uses future foothold positions demanding tasks. The boundary conditions center mass (COM) state considered optimization process to ensure seamless transition two controllers. We demonstrate physically realistic simulations, well experimentally small humanoid robot platform.