Font Size: a A A

Gait Research And Implementation Of Soccer Humanoid Robot Based On CMA-ES Algorithm

Posted on:2015-06-01Degree:MasterType:Thesis
Country:ChinaCandidate:J XueFull Text:PDF
GTID:2308330473456992Subject:Computer technology
Abstract/Summary:PDF Full Text Request
The research of humanoid robot began from 1960s, it’s not a long time from today, but its development process is very fast and it has become one of the main research directions of robotics research. Gait planning has been a hot issue of humanoid robot research, there is still not a universal solution to achieve a stable and fast walking motion in a complex and dynamic environment. In this paper, we study a method of the implementation and optimization of a dribbling gait of a soccer robot on the platform of RoboCup3D.The dribbling gait of a soccer humanoid robot refers to a special gait of in the soccer game environment, its main feature is to ensure the stability when the robot is frequently alternating between walking action and other actions of soccer game as long as improve walking speed. Firstly, we achieve a set of simple gait using keyframe planning methods, then, we comprehensive learn and optimize the parameter of the simple gait in a simplified soccer game environment using CMA-ES algorithm. The experimental results show that the performance of the simple gait in the soccer game is improved and the CMA-ES algorithm is very effective.Three dimensional linear inverted pendulum model is a commonly used simplified model of gait planning of bipedal humanoid robot, its advantages is that we can extract the most important features from the complex dynamic characteristics of the biped robot, and then we can achieve a smooth, stable and adaptable walking motion by analyzing the function of the walking system using these important characteristics. In this paper, we study a method to plan and implement the gait of humanoid robot using three dimensional linear inverted pendulum model on the platform of RoboCup3D. Then, we use the method of layered learning to learn and optimize the gait parameters in the soccer game environment. The gait achieved by this method is more in line with the characteristics of dribbling gait, and further improved the performance of the dribbling gait in the soccer game.
Keywords/Search Tags:Humanoid robot, gait planning, Dribbling gait, CMA-ES
PDF Full Text Request
Related items