Compliant Bipedal Walking Based on Variable Spring-Loaded Inverted Pendulum Model with Finite-sized Foot*

Traditional Spring-Loaded Inverted Pendulum (SLIP) model is underactuated for ignoring the foot with ankle, which limits the movability and controllability of the robot in bipedal walking. This paper proposes a new walking template called Variable Spring-Loaded Inverted Pendulum Model with Finite-sized Foot (VSLIP-FF) for biped robots. By extending the SLIP model with a finite-sized foot and a 1-DoF ankle joint for each leg and making the leg stiffness adjustable, the VSLIP-FF model can be used to realize compliant bipedal walking in complex environments. Inspired by the characteristics of human walking, an adaptive leg stretching and contracting strategy for gait planning is proposed to play the role of the ankle joint. To ensure walking stability, the Foot-Rotation Indicator (FRI) point is used as the stability criterion to prevent the robot from falling. The differential evolution algorithm is used to generate the desired center of mass (CoM) and foot trajectories of a complete walking cycle. Simulation results suggest that, compared with the SLIP model, the step length range of the VSLIP-FF model with the gait planning method is increased by 19.35%. Based on this method, the complaint walking experiment for the biped robot to step over discrete terrain is realized.