Planning and Control of Forward Jumping Movement of Humanoid Robot

Humanoid robot's stable walking and crawling movement patterns can help them adapt to a relatively flat road, but these movement patterns are difficult to deal with conditions such as ditches and platforms. Adding jumping motion to the robot can greatly enhance the robot's ability to adapt to the environment. This paper explores the forward jumping of the humanoid robot. A simplified three-link model of the humanoid robot in the sagittal plane is established and the jumping process is decomposed into three phases: take-off phase, flight phase, and landing phase. The mathematical model for each phase of the forward jumping of the humanoid robot is developed. In the whole process, motors provide torques for the three joints of the robot's hip joint, knee joint and ankle joint. Then the kinematics and dynamics are constrained for each phase, and the cost function is established. Furthermore, the pseudo-spectral method is used to optimize the design of the jumping trajectory to ensure the forward jumping of the humanoid robot. Finally, the validity of the proposed method for the whole forward jumping process is verified by the simulation experiment. The simulation result shows that the model of the humanoid robot can achieve 0.85 meters forward jumping.

I. Introduction

In the past few decades, research in the field of humanoid robots has made great progress. Humanoid robots' locomotion ability has been greatly improved. Most of them have the stable walking and crawling movement patterns which can accommodate relatively flat roads [1]–[7]. However, when faced with conditions such as ditches and platforms, these movement patterns can hardly deal with the problem. Inspired by the human jumping movement, adding jumping motion to the robots can greatly enhance their environment adaptability [8]. Therefore, achieving the jumping motion is a crucial goal for the research of intelligent robot.