A balanced jumping control algorithm for quadruped robots

The jumping movement of quadruped robots is crucial, so a complete set of balanced jumping algorithms is proposed in this paper due to the flaws of the current jumping algorithms. The proposed algorithm includes trajectory planning of the Center of Mass(CoM) and four jumping phases, illustrating the jumping process in detail. Tasks that a quadruped robot with the height of 0.6 m jumps up a step with the height of 0.3 m, 0.4 m, 0.5 m and 0.6 m are the research object. Optimal Parabola Trajectory of CoM(OPTC) is solved by about ten iterations based on the fastest approaching strategy before jumping. During the first phase of jumping, Ground Reaction Forces(GRFs) are precisely distributed to control six Degrees of Freedom(DoFs) based on symmetric six-dimensional spatial mechanics decoupling solution, controlling the robot to adjust itself to the best ejecting posture. The maximum displacement error is less than 0.005 m. During the second phase, full-leg ejection is implemented to eject, guaranteeing that the robot accurately tracks the OPTC after takeoff by updating proportional virtual forces. The tracking Mean Square Error(MSE) is less than 0.06. During the third phase, the flying attitude is adjusted by swinging leg theory summarized in the paper, with the maximum pitch angle less than 4.5 degrees. Meanwhile, the theoretical landing points of feet are calculated to lead the movement of feet, ensuring a soft landing to reduce touchdown impact. The momentary landing velocities of feet are less than 0.09 m/s. During the fourth phase, the robot buffers and brakes to a static state based on the algorithm used in the first phase. Eventually, the proposed algorithms are verified through simulating experiments on Webots physical engine, and the effectiveness and feasibility are validated by the experimental results.(c) 2022 Elsevier B.V. All rights reserved.