Development of hybrid system control theory and application to biped robot

混合系统控制理论发展及其在双足机器人中的应用

• 批准号: 18560446
• 负责人: YINGJIE Yie
• 金额: $ 2.53万
• 依托单位: The Institute of Physical and Chemical Research
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2006
• 资助国家: 日本
• 起止时间: 2006 至 2007
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-18560446/
• 关键词: Hybrid system; biped robot; Locomotion planning; Motion control; Complex system; 制御理論; 運動制御システム

Human being can decide his walking or running to auto-adapt to environment changes. Comparing with this, the automation of biped robots is still inefficient. One reason for this is the less of a unified modeling framework for various types of biped motion in varying environment. Another reason is the lack of a systematic theory for control of the complex system. In this research, we at first proposed to recast types of biped motion in the framework of hybrid system, expressed it as a mixed logic dynamical model. Such approach possesses the advantage that it encompasses all the motions and environment change into a unique one, thus allow the control system to be design optimally. For the gait pattern planning of biped motion, we proposed a hybrid external system for the synthesis of periodic orbit which is both robust to disturbance and adaptive to environment. Comparing with traditional motion planning methods where the trajectories were piecewise approximated by polynomials, our proposal is more general and systematic. The structure, the manifold of initial state, and the selection of parameters of the hybrid external system were analytically addressed. The output of the hybrid external system corresponded to the periodic trajectory of joints, which can be designed arbitrary smoothly. Basing on the nonlinear regulation theory, the biped robot was controlled to asymptotically track the hybrid external system.

人可以决定步行或跑步以自动适应环境变化。与此相比，双头机器人的自动化仍然效率低下。造成这种情况的一个原因是，在不同环境中，对于各种类型的双重运动，统一的建模框架较少。另一个原因是缺乏控制复杂系统的系统理论。在这项研究中，我们首先提出在混合系统框架中重新铸造的双头运动类型，将其表示为混合逻辑动力学模型。这种方法具有这样的优点，即它包含所有动作和环境变为唯一的动作，从而使控制系统可以最佳设计。对于双重运动的步态模式计划，我们提出了一个混合外部系统，以合成周期性轨道，这既适合干扰又适应环境。与传统的运动计划方法相比，轨迹被多项式近似近似，我们的建议更加通用和系统性。分析解决了结构，初始状态的多种歧视以及混合外部系统的参数的选择。混合外部系统的输出对应于关节的周期轨迹，可以平稳设计。在非线性调节理论的基础上，将双支化的机器人控制以渐近地跟踪混合外部系统。

项目成果

A Hybrid System Control Approach to Biped Robot Control

双足机器人控制的混合系统控制方法

• 发表时间: 2007
• 作者: Yingjie Yin;Shigeyuki Hosoe;And Zhiwei Luo;Yingjie Yin and Shigeyuki Hosoe
• 通讯作者: Yingjie Yin and Shigeyuki Hosoe

Design and Modelling of a Jumping Machine in Circular Motion

圆周运动跳跃机的设计与建模

• 发表时间: 2007
• 作者: 村上 祐貴;鈴木 修一;大下 英吉;堤 知明;Yingjie Yin;村上 祐貴;Yingjie Yin
• 通讯作者: Yingjie Yin

Humanoid Robots : Human-like Machines, Chapter 16 : Mixed Logic Dynamical Modeling and On Line Optimal Control of Biped Robot

人形机器人：类人机器，第 16 章：双足机器人的混合逻辑动态建模和在线优化控制

• 发表时间: 2007
• 作者: 鈴木 修一;村上 祐貴;大下 英吉;堤 知明;Yingjie Yin and Shigeyuki Hosoe;Yingjie Yin and Shigeyuki Hosoe
• 通讯作者: Yingjie Yin and Shigeyuki Hosoe

A mixed logic dynamical modeling formulation and optimal control of intelligent robots

• DOI: 10.1007/s11081-007-9017-z
• 发表时间: 2007-07
• 期刊: Optimization and Engineering
• 影响因子: 2.1
• 作者: Yingjie Yin;S. Hosoe;Zhiwei Luo

Humanoid Robots:Human-like Machines, Chapter 16: Mixed Logic Dynamical Modeling and On Line Optimal Control of Biped Robot

仿人机器人：类人机器，第 16 章：双足机器人的混合逻辑动力学建模和在线优化控制

• 发表时间: 2007
• 作者: 鈴木 修一;村上 祐貴;大下 英吉;堤 知明;Yingjie Yin and Shigeyuki Hosoe;Yingjie Yin and Shigeyuki Hosoe;Yingjie Yin and Shigeyuki Hosoe
• 通讯作者: Yingjie Yin and Shigeyuki Hosoe