基于Fuzzy–CPG的双足机器人适应性行走控制

为提高双足机器人的环境适应性,本文提出了一种基于模糊控制与中枢模式发生器(CPG)的混合控制策略,称之为Fuzzy–CPG算法.高层控制中枢串联模糊控制系统,将环境反馈信息映射为行走步态信息和CPG幅值参数.低层控制中枢CPG根据高层输出命令产生节律性信号,作为机器人的关节控制信号.通过机器人运动,获取环境信息并反馈给高层控制中枢,产生下一步的运动命令.在坡度和凹凸程度可变的仿真环境中进行混合控制策略的实验验证,结果表明,本文提出的Fuzzy–CPG控制方法可以使机器人根据环境的变化产生适应的行走步态,提高了双足机器人的环境适应性行走能力.     

双足步行机器人能量成型控制

为了使双足被动行走机器人的行走步态符合仿生规律,且当路面坡度变化后,迅速进入新的稳定步态行走,提出了角度不变能量成型控制策略.研究了欠驱动双足机器人能量匹配条件和能量成型控制器的求解;由于动能相对于旋转变换不具有对称性,通过在能量成型控制中附加一个辅助控制量,实现角度不变控制.仿真结果表明,该算方法可实现仿生控制,既能扩大吸引域,又改善系统的鲁棒性.     

双足机器人稳定行走的仿人预测控制研究

针对双足机器人的稳定行走,提出了一种新的仿人预测控制在线步行模式生成方法。把期望零力矩点（ZMP）分解成离线规划好的参考ZMP和实时变化的可变ZMP之和,通过预测控制和其逆系统共同作用对质心运动进行控制,从而生成具有自适应性的步行模式。但单一的预测控制系统对诸如矩形齿状扰动的可变ZMP的跟踪存在较大的误差,结合仿人智能控制对误差的强抑制能力,设计了与预测控制相结合的仿人预测控制系统。仿真实验验证对矩形齿状扰动的可变ZMP,仿人预测系统也能实现较好的跟踪。     

变长度柔性双足机器人行走控制及稳定性分析

针对传统双足机器人模型缺少脚质量和躯干的问题,提出考虑摆动腿动态及躯干影响的柔性双足机器人模型,并对其行走控制及稳定性进行研究。首先,建立系统的动力学模型并采用欧拉-拉格朗日法推导了系统的动力学方程;同时,在弹簧负载倒立摆（SLIP）模型的基础上添加刚性躯干、脚质量及采用变长度伸缩腿,充分考虑躯干及摆动腿动力学对机器人行走步态的影响;其次,设计基于变长度腿的反馈线性化控制器来跟踪目标轨迹,以及调节摆动腿和躯干的姿态;最后,利用Newton-Raphson迭代法和庞加莱映射分析机器人的不动点及轨道稳定性条件,并在理论分析的基础上进行仿真。仿真结果表明,所提控制器可以实现机器人的周期行走,对外界干扰具有良好的鲁棒性,且雅可比矩阵所有特征值的模均小于1,能形成稳定的极限环,证明系统是轨道稳定的。     

仿人预测控制在双足机器人步行控制中的应用

基于零力矩点(ZMP)的预测控制是目前双足机器人步行控制中最先进的方法,但是预测控制需要比较精确的预测模型,在环境扰动导致模型失配时,预测控制的性能下降较快。为了解决这个问题,利用仿人智能控制对环境误差具有较强抑制的特点改进预测控制。探讨了在步行控制中引入仿人智能控制的必要性和仿人智能控制改进预测控制的可行性,并设计了仿人预测控制器。最后通过仿真实验验证了新的控制器对双足机器人步行控制的有效性。     

Principle and method of speed control for dynamic walking biped robots

In this paper, the method of speed control for 3D biped robots is addressed. First, the primary principle of speed control by regulation of input energy is studied, the feature of which is to regulate the speed and the step length synchronically. The method of Poincaré mapping is used to prove the stability of speed control in the common range. Second, a method of speed control for an 18 DOFs bipedal 3D robot, which is characterized by the two-point-foot, is proposed. The method is developed on the basis of the 3D walking pattern proposed previously, with the new function of speed regulation being added in. The simulations show that the performances of regular walking, acceleration, and deceleration are effective and stable, and therefore verify the feasibility of the proposed method. Furthermore, some walking features, such as the walking efficiency and lateral control, are demonstrated.     

伸缩腿双足机器人半被动行走控制研究

研究半被动伸缩腿双足机器人行走控制和周期解的全局稳定性问题.使用杆长可变的倒立摆机器人模型,以支撑腿的伸缩作为行走动力源,采用庞加莱映射方法分析了双足机器人行走的不动点及其稳定性.当脚与地面冲击时,假设两腿间的夹角保持为常数,设计了腿伸缩长度的支撑腿角度反馈控制率.证明了伸缩腿双足机器人行走过程不动点的全局稳定性.仿真结果表明,本文提出的腿伸缩长度反馈控制可以实现伸缩腿双足机器人在水平面上的稳定行走,并且周期步态对执行器干扰和支撑腿初始角速度干扰具有鲁棒性.     

双足机器人动态步行实时步态生成

针对双足机器人动态步行生成关节运动轨迹复杂问题,提出了一种简单直观的实时步态生成方案。建立了平面五杆双足机器人动力学模型,通过模仿人类步行主要运动特征并根据双足机器人动态步行双腿姿态变化的要求,将动态步行复杂任务分解为顺序执行的四个过程,在关节空间相对坐标系下设计了躯干运动模式、摆动腿和支撑腿动作及步行速度调整模式,结合当前步行控制结果反馈实时产生稳定的关节运动轨迹。仿真实验验证了该方法的有效性,简单易实现。     

基于deep Q-network双足机器人非平整地面行走稳定性控制方法

针对双足机器人在非平整地面行走时容易失去运动稳定性的问题,提出一种基于一种基于价值的深度强化学习算法DQN（Deep Q-Network）的步态控制方法。首先通过机器人步态规划得到针对平整地面环境的离线步态,然后将双足机器人视为一个智能体,建立机器人环境空间、状态空间、动作空间及奖惩机制,该过程与传统控制方法相比无需复杂的动力学建模过程,最后经过多回合训练使双足机器人学会在不平整地面进行姿态调整,保证行走稳定性。在V-Rep仿真环境中进行了算法验证,双足机器人在非平整地面行走过程中,通过DQN步态调整学习算法,姿态角度波动范围在3°以内,结果表明双足机器人行走稳定性得到明显改善,实现了机器人的姿态调整行为学习,证明了该方法的有效性。     

Stability analysis and time-varying walking control for an under-actuated planar biped robot

This paper presents two walking controllers for a planar biped robot with unactuated point feet. The control is based on the tracking of reference motions expressed as a function of time. First, the reference motions are adapted at each step in order to create a hybrid zero dynamic (HZD) system. Next, the stability of the walking gait under closed-loop control is evaluated with the linearization of the restricted Poincaré map of the HZD. When the controlled outputs are selected to be the actuated coordinates, most periodic walking gaits for this robot are unstable, that is, the eigenvalues of the linearized Poincaré map (ELPM) is larger than one. Therefore, two control strategies are explored to produce stable walking. The first strategy uses an event-based feedback controller to modify the ELPM and the second one is based on the choice of controlled outputs. The stability analysis show that, for the same robot and for the same reference trajectory, the stability of the walking (or ELPM) can be modified by some pertinent choices of controlled outputs. Moreover, by studying some walking characteristics of many stable cases, a necessary condition for stable walking is proposed. It is that the height of swing foot is nearly zero at the desired moment of impact. Based on this condition, the duration of the step is almost constant in presence of initial error, so a method for choosing controlled outputs for the second controller is given. By using this method, two stable domains for the controlled outputs selection are obtained.     

基于CPG的双足机器人多层步行控制器设计

针对双足机器人步行控制器的设计问题,基于生物学启发原理,提出一种基于中枢模式发生器(CPG)与稳定性分析相结合的多层次结构控制器设计方法。分析机器人的步行运动,给出双足运动描述。基于CPG理论设计多层次结构的步行控制器,应用极限环理论方法分析运动稳定性。在保证步行稳定的前提下,所提出的控制方法具有结构简单、通用性好、方便在线平滑修正等优点,步行仿真实验验证了所提出算法的有效性。     

Reinforcement learning for quasi-passive dynamic walking of an unstable biped robot

A class of biped locomotion called Passive Dynamic Walking (PDW) has been recognized to be efficient in energy consumption and a key to understand human walking. Although PDW is sensitive to the initial condition and disturbances, studies of Quasi-PDW which incorporates supplemental actuators have been reported to overcome this sensitivity. In this article, we propose a reinforcement learning method designed particularly for Quasi-PDW of a biped robot whose possession of knees makes the system unstable. Simulations show that the learning is quickly accomplished after 1000 episodes, and the obtained controller is robust against variations in the slope gradient and sudden perturbations.     

Stability and control of dynamic walking for a five-link planar biped robot with feet

During dynamic walking of biped robots, the underactuated rotating degree of freedom (DOF) emerges between the support foot and the ground, which makes the biped model hybrid and dimension-variant.This paper addresses the asymptotic orbit stability for criterion for DVHS is also presented, and the result is then used to study dynamic walking for a five-link planar biped robot with feet. Time-invariant gait planning and nonlinear control strategy for dynamic walking with flat feet is also introduced. Simulation results indicate that an asymptotically stable limit cycle of dynamic walking is achieved by the proposed method.     

欠驱动两足步行机器人侧向稳定控制方法研究

以欠驱动两足步行机器人为对象研究其侧向运动稳定控制问题。首先分析引起机器人侧向运动不稳定的原因,然后提出步宽控制和侧向力矩补偿两种控制策略。步宽控制通过控制机器人侧向落脚位置,使其侧向运动周期与前向周期趋于一致实现侧向运动稳定。力矩补偿控制通过在踝关节引入侧向控制力矩,使侧向运动与前向运动协调一致实现侧向运动稳定。仿真实验表明,机器人实现了稳定的3D动态行走,达到了预期的控制效果。     

Stability and control of dynamic walking for a five-link planar biped robot with feet

During dynamic walking of biped robots, the underactuated rotating degree of freedom （DOF） emerges between the support foot and the ground, which makes the biped model hybrid and dimension-variant. This paper addresses the asymptotic orbit stability for dimension-variant hybrid systems （DVHS）. Based on the generalized Poincare map, the stability criterion for DVHS is also presented, and the result is then used to study dynamic walking for a five-link planar biped robot with feet. Time-invariant gait planning and nonlinear control strategy for dynamic walking with fiat feet is also introduced. Simulation results indicate that an asymptotically stable limit cycle of dynamic walking is achieved by the proposed method.     

双足机器人自适应常值驱动与传感反馈结合的仿生行走控制

为了将双足机器人的混沌步态控制收敛到稳定的周期步态,提出一种控制策略。首先用庞卡莱截面法研究斜坡倾角变化对步态的影响,结果表明,坡度增大会导致倍周期步态到混沌步态的产生;然后以人类步行的生物力学为仿生依据,根据延迟反馈控制的基本思路,设计了自适应常值驱动与传感反馈相结合的仿生行走控制策略,并依据当前步和前两步初始状态对控制器参数进行逐步调节,最终将混沌步态控制收敛到周期步态。仿真结果表明了所提出算法的有效性。     

Efference copies in neural control of dynamic biped walking

In the early 1950s, von Holst and Mittelstaedt proposed that motor commands copied within the central nervous system (efference copy) help to distinguish ‘reafference’ activity (afference activity due to self-generated motion) from ‘exafference’ activity (afference activity due to external stimulus). In addition, an efference copy can be also used to compare it with the actual sensory feedback in order to suppress self-generated sensations. Based on these biological findings, we conduct here two experimental studies on our biped “RunBot” where such principles together with neural forward models are applied to RunBot’s dynamic locomotion control. The main purpose of this article is to present the modular design of RunBot’s control architecture and discuss how the inherent dynamic properties of the different modules lead to the required signal processing. We believe that the experimental studies pursued here will sharpen our understanding of how the efference copies influence dynamic locomotion control to the benefit of modern neural control strategies in robots.     

学习人类控制策略的双足机器人步态控制研究

针对双足机器人面临的复杂环境下动态行走的适应性难题,提出了一种基于学习人类控制策略的双足机器人步态控制方法。利用三维线性倒立摆模型构造双足行走系统的状态方程,建立学习人类控制策略的参数化模型,设计了基于SVM的学习型控制器。该方法保证了躯干始终处于与地面近似垂直,增强了步态控制的鲁棒性,提高了双足机器人在复杂环境下行走的动态稳定性。实验验证了该方法的有效性。     

仿人机器人动态步行控制综述

综述了仿人机器人动态步行的研究历史和研究现状。归纳了动态步行的特点,分析了动态步行稳定性判定方法,介绍了基于ZMP的姿态稳定判据和基于庞加莱映射（Poincaré Map）的步态稳定判据。总结了仿人机器人学习适应复杂地面环境步行的方法,概述了动态步行控制实现的典型解决方案,指出了动态步行控制中待解决的问题,并探讨了未来的发展方向。     

Neuro-fuzzy gait synthesis with reinforcement learning for a biped walking robot

 A reinforcement learning-based neuro-fuzzy gait synthesizer, which is based on the GARIC (Generalized Approximate Reasoning for Intelligent Control) architecture, is proposed for the problem of biped dynamic balance. We modify the GARIC architecture to enable it to generate the trunk trajectory in both sagittal and frontal plane. The proposed gait synthesizer is trained by reinforcement learning that uses a multi-valued scalar signal to evaluate the degrees of failure or success for the biped locomotion by means of the ZMP (Zero Moment Point). It can form the initial dynamic balancing gait from linguistic rules, which are obtained from human intuitive balancing knowledge and biomechanics studies, and accumulate dynamic balancing knowledge through reinforcement learning, and thus constantly improve its gait during walking. The feasibility of the proposed method is verified through a 5-link biped robot simulation.