ekorudiawan / Omnidirectional-Walking-ZMP-Analytical

Omnidirectional Walking Pattern Generation for Humanoid Robot using ZMP Analytical Method

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Omnidirectional Walking Pattern Generation for Humanoid Robot using ZMP Analytical Method

By: Eko Rudiawan

Overview

This is an implementation of omnidirectional walking pattern generator for humanoid robot using ZMP analytical approach that was proposed by Harada, Kensuke, et al. The input is defined as motion command $cmd_x$, $cmd_y$, and $cmd_\theta$ that represented next landing position and orientation of swing foot. The source code will generate trajectory of CoM, left foot and right foot written from world coordinate. This code can be implemented on real robot by adding inverse kinematics solver.

Example Results

  1. Forward walking with motion command $cmd_x = 0.05$, $cmd_y = 0.00$, and $cmd_\theta = 0.00$

    alt text

  2. Diagonal walking with motion command $cmd_x = 0.04$, $cmd_y = 0.02$, and $cmd_a = 0.00$

    alt text

  3. Turning with motion command $cmd_x = 0.05$, $cmd_y = 0.00$, and $cmd_a = 10^o$

    alt text

References

[1] Harada, Kensuke, et al. "An analytical method for real-time gait planning for humanoid robots." International Journal of Humanoid Robotics 3.01 (2006): 1-19.

About

Omnidirectional Walking Pattern Generation for Humanoid Robot using ZMP Analytical Method


Languages

Language:Python 100.0%