Gait planning and experimental study of biped robot based on 3-D walking sequence

CHEN Jianfang,DING Jiatao,XIAO Xiaohui

DOI: https://doi.org/10.11817/j.issn.1672-7207.2018.04.009

2018-01-01

Abstract:In order to improve the adaptability of biped walking in the real environment, a novel gait planning method based on 3-D walking sequence was proposed. By imitating human walking behaviour and taking the change of ground height into consideration, 3D walking sequence (WS), which also considers the rotational motion of ankle joint and the vertical height of the robot, was firstly defined. Then, during single support phase (SSP), optimal control (OC) theory and linear inverted pendulum model (LIMP) were adopted for generation of zero moment point (ZMP) trajectories. The 5th order spline interpolation was used to generate ZMP trajectories and to produce the center-of-mass (COM) height variation during the double support phase (DSP). By employing motion parameters of the foot, the foot trajectory was also obtained by piecewise continuous cubic Hermite interpolation. After that, the Newton-Raphson method was used to solve the desired joint angle. Finally, dynamic simulation and hardware experiments of bipedal walking on flat and slope ground were carried out by employing the NAO robot. The results show that the proposed method is effective in practice and that the robot walks more stably in the real environment. Fluctuation range of the foot contact force of NAO robot in flat and slope conditions decrease by 37.01% and 59.94% respectively compared with those in simulation test, and fluctuation range of the hip joint decrease by 69.13% and 62.75%, respectively.