Abstract
Humanoid robots are expected to have broad applications due to their biped mobility and human-like shape. To increase the walking speed, it is necessary to increase the power for driving the joints of legs. However, the resulting mass increasing of the legs leads to a rotational slip when a robot is walking fast. In this paper, a 3D three-mass model is proposed, in which both the trunk and thighs are regarded as an inverted pendulum, and the shanks and feet are considered as mass-points under no constraints with the trunk. Then based on the model, a friction constraint method is proposed to plan the trajectory of the swing leg in order to achieve the fastest walking speed without any rotational slip. Furthermore, the compensation for zero-moment point (ZMP) is calculated based on the 3D three-mass model, and the hip trajectory is obtained based on the compensated ZMP trajectory by using the preview control method, thus improving the robot’s overall ZMP follow-up effect. This planning method involves simple calculations but reliable results. Finally, simulations confirm that the rotational slip is avoided while stable and fast walking is realized, with free joints of the waist and arms, which then could be planned for other tasks.
Highlights
Humanoid robots can enter non-structural environments that wheeled and crawler robots find difficulties to reach or pass through by virtue of their biped mobility advantages [1,2,3]
Sato et al further proposed a three-mass model for the humanoid robot, which divided the robot into a body, support leg and swing leg, and the adopted mass of these three parts to calculate the zero-moment point (ZMP) of the robot, which resulted in a smaller ZMP error and a walk mode with higher efficiency [23,24]
This paper aims to overcome power of the leg joints, which in turn leads to an increase of leg weight
Summary
Humanoid robots can enter non-structural environments that wheeled and crawler robots find difficulties to reach or pass through by virtue of their biped mobility advantages [1,2,3]. Kajita directly takes the zero-moment point (ZMP) Equation as a constraint condition, and obtains a smooth trajectory of CoM that takes into account both the robot’s stability and the CoM’s acceleration increment by means of quadratic optimization [21] It relies heavily on the accuracy of modeling, so it is very effective on humanoid robot with light legs but has poor performance in robots with heavy legs. The rest of this paper is organized as follows: Section 2 proposes the 3D three-mass model; in Section 3, friction constraint is calculated and trajectory of the swing leg is planned; Section 4 records the planning of hip trajectory; and the simulations and conclusions are presented in Sections 5 and 6 respectively
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
