arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1006.4553cs.RO
收藏

基于神经振荡器控制器和和声搜索算法优化的步行进化

Evolution of Biped Walking Using Neural Oscillators Controller and Harmony Search Algorithm Optimizer

Qazvin Azad University(加兹温阿扎德大学)

Ebrahim Yazdi, Abolfazl Toroghi Haghighat

摘要:

在本文中,一个简单的神经网络控制器已被用来实现稳定的步行NAO机器人,与22个自由度,实现在一个虚拟的基于物理的仿真环境中的Robocup足球仿真环境。该算法使用Matsuoka基神经振荡器来产生机器人的控制信号。为了找到最佳的角轨迹和优化网络参数,使用了一种新的基于种群的搜索算法,称为和声搜索(HS)算法。该算法将一组音乐家概念化,试图寻找更好的和声状态。仿真结果表明,由于感觉反馈信号的步进周期和步行运动的修改提高了步行运动的稳定性。

英文摘要:

In this paper, a simple Neural controller has been used to achieve stable walking in a NAO biped robot, with 22 degrees of freedom that implemented in a virtual physics-based simulation environment of Robocup soccer simulation environment. The algorithm uses a Matsuoka base neural oscillator to generate control signal for the biped robot. To find the best angular trajectory and optimize network parameters, a new population-based search algorithm, called the Harmony Search (HS) algorithm, has been used. The algorithm conceptualized a group of musicians together trying to search for better state of harmony. Simulation results demonstrate that the modification of the step period and the walking motion due to the sensory feedback signals improves the stability of the walking motion.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑