一种小样本支持向量机控制器在两足机器人步态控制的研究

神经网络等传统的机器学习方法是基于样本数目无穷大的经验风险最小化原则,这对非确定环境下有限样本的步态学习控制非常不利．针对两足机器人面临的非确定环境适应性难题,提出了一种基于支持向量机（SVM）的两足机器人步态控制方法,解决了小样本条件下的步态学习控制问题．提出了一种基于混合核的步态回归方法,仿真研究表明了这种方法比全局核和局部核分别单独用于步态学习时有优越性．SVM以踝关节及髋关节的轨迹作为输入,相应的满足ZMP判据的上体轨迹作为输出,利用有限的理想步态样本对机器人上体轨迹与腿部轨迹之间的动态运动关系进行学习,然后将训练好的SVM置入机器人控制系统,从而增强了步态控制的鲁棒性,有利于实现两足机器人在非结构环境下的稳定步行．仿真结果表明了所提方法的优越性．     

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。     

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.     

面向离散地形的欠驱动双足机器人平衡控制方法

欠驱动双足机器人在行走中为保持自身的平衡,双脚需要不间断运动.但在仅有特定立足点的离散地形上很难实现调整后的落脚点,从而导致欠驱动双足机器人在复杂环境中的适应能力下降.提出了基于虚拟约束(Virtual constraint,VC)的变步长调节与控制方法,根据欠驱动双足机器人当前状态与参考落脚点设计了非时变尺度缩放因子,能够实时重构适应当前环境的步态轨迹;同时构建了全身动力学模型,采用反馈线性化的模型预测控制(Model predictive control,MPC)滚动优化产生力矩控制量,实现准确的轨迹跟踪控制.最终进行了欠驱动双足机器人的随机离散地形稳定行走的仿真实验,验证了所提方法的有效性与鲁棒性.     

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.     

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.     

动态双足机器人的控制与优化研究进展

对动态双足机器人的可控周期步态的稳定性、鲁棒性和优化控制策略的国内外研究现状与发展趋势进行了探讨.首先,介绍动态双足机器人的动力学数学模型,进一步,提出动态双足机器人运动步态和控制系统原理;其次,讨论动态双足机器人可控周期步态稳定性现有的研究方法,分析这些方法中存在的缺点与不足;再次,研究动态双足机器人的可控周期步态优化控制策略,阐明各种策略的优缺点;最后,给出动态双足机器人研究领域的难点问题和未来工作,展望动态双足机器人可控周期步态与鲁棒稳定性及其应用的研究思路.     

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。     

基于能效优化的两足机器人步态控制方法

针对高能耗导致的两足机器人实用化障碍,提出了一种全新的、系统化的步态能效优化控制方法.基于两足机器人运动的重要能耗指标(平均功率、平均功率偏差、平均力矩损耗),提出了能耗预估策略和能效优化算法,获取了零力矩点(ZMP)稳定区域内的能耗极小值.沿着能耗极小值所对应的上体轨迹对机器人步态实施能效优化控制,最终获得满足ZMP稳定判据的低能耗步态.仿真结果证明,该方法能够有效降低机器人能耗并保持其稳定性.     

基于深度强化学习的双足机器人斜坡步态控制方法

为提高准被动双足机器人斜坡步行稳定性, 本文提出了一种基于深度强化学习的准被动双足机器人步态控制方法. 通过分析准被动双足机器人的混合动力学模型与稳定行走过程, 建立了状态空间、动作空间、episode过程与奖励函数. 在利用基于DDPG改进的Ape-X DPG算法持续学习后, 准被动双足机器人能在较大斜坡范围内实现稳定行走. 仿真实验表明, Ape-X DPG无论是学习能力还是收敛速度均优于基于PER的DDPG. 同时, 相较于能量成型控制, 使用Ape-X DPG的准被动双足机器人步态收敛更迅速、步态收敛域更大, 证明Ape-X DPG可有效提高准被动双足机器人的步行稳定性.     

半被动双足机器人的准开环控制

目前的被动行走机器人还只能完成单一的步态,且非常容易摔倒,为此对半被动双足机器人的稳定行走控制问题进行了研究．通过结合被动行走和主动控制两种原理的优点,提出了一种准开环的行走控制方法．通过检测安装在机器人足底和髌骨处的接触开关信号,在髋关节处施加一个间断的、微小的开环振荡力矩,进而实现高效的稳定步行．仿真结果表明,当控制参数在较大范围内变化时,双足机器人仍可实现稳定行走,且步行能耗特性与人类相似;通过调节振荡力矩的参数,机器人可实现稳定的行走模式转换．     

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。     

基于足底压力与支持向量机的步态相位识别研究

随着两足机器人、人工假肢技术以及为行走困难病人康复设计的康复训练机器人的发展,在线的步态相位识别方法越来越重要。本文提出的基于足底压力与支持向量机(SVM)的步态相位识别算法主要由五部分组成,即数据采集、数据预处理、特征提取、训练分类器和分类识别。实验表明：该方法能够对运动中的步态相位进行准确的判断。     

Interval type-2 fuzzy weighted support vector machine learning for energy efficient biped walking

An interval type-2 fuzzy weighted support vector machine (IT2FW-SVM) is proposed to address the problem of high energy consumption for biped walking robots. Different from the traditional machine learning method of ‘copy learning’, the proposed IT2FW-SVM obtains lower energy cost and larger zero moment point (ZMP) stability margin using a novel strategy of ‘selective learning’, which is similar to human selections based on experience. To handle the uncertainty of the experience, the learning weights in the IT2FW-SVM are deduced using an interval type-2 fuzzy logic system (IT2FLS), which is an extension of the previous weighted SVM. Simulation studies show that the existing biped walking which generates the original walking samples is improved remarkably in terms of both energy efficiency and biped dynamic balance using the proposed IT2FW-SVM.     

基于能量成型的双足机器人解耦控制

研究带有膝关节和髋关节的双足机器人在3D (three-dimensional)空间稳定行走的控制器设计.通过构建概循环拉格朗日函数,将双足机器人的3D动态系统解耦成前向和侧向两部分,对前向部分设计势能成型控制器,使前向获得稳定行走步态;用输出零动态控制器控制侧向,满足系统的动态解耦条件.仿真结果表明了所提出方法的有效性.     

Gait generation and control for biped robots with underactuation degree one

The paper develops a unified feedback control law for n degree-of-freedom biped robots with one degree of underactuation so as to generate periodic orbits on different slopes. The periodic orbits on different slopes are produced from an original periodic orbit, which is either a natural passive limit cycle on a specific slope or a stable periodic walking gait on level ground generated with active control. First, inspired by the controlled symmetries approach, a general result on gait generation on different slopes based on a periodic orbit on a specific slope is obtained. Second, the time-scaling control approach is integrated to reproduce geometrically same periodic orbits for biped robots with one degree of underactuation. The degree of underactuation is compensated by one degree-of-freedom in the temporal evolution that scales the original periodic orbit. Necessary and sufficient conditions are investigated for the existence and stability properties of periodic orbits on different slopes with the proposed control law. Finally, the proposed approach is illustrated by two kinds of underactuated biped robots: one has a passive gait on a specific ground slope and the other does not have a natural passive gait.     

Walking Control Algorithm of Biped Humanoid Robot on Uneven and Inclined Floor

This paper describes walking control algorithm for the stable walking of a biped humanoid robot on an uneven and inclined floor. Many walking control techniques have been developed based on the assumption that the walking surface is perfectly flat with no inclination. Accordingly, most biped humanoid robots have performed dynamic walking on well designed flat floors. In reality, however, a typical room floor that appears to be flat has local and global inclinations of about 2°. It is important to note that even slight unevenness of a floor can cause serious instability in biped walking robots. In this paper, the authors propose an online control algorithm that considers local and global inclinations of the floor by which a biped humanoid robot can adapt to the floor conditions. For walking motions, a suitable walking pattern was designed first. Online controllers were then developed and activated in suitable periods during a walking cycle. The walking control algorithm was successfully tested and proved through walking experiments on an uneven and inclined floor using KHR-2 (KAIST Humanoid robot-2), a test robot platform of our biped humanoid robot, HUBO.     

双足机器人自然ZMP轨迹生成方法研究

为了实现双足机器人类人行走,提出了一种基于自然ZMP轨迹的双足机器人步行模式生成方法。在单腿支撑相,根据基于三维线性倒立摆模型,在设定从脚跟到脚趾移动的自然ZMP轨迹后,得到质心轨迹方程;在双腿支撑相采用线性摆模型生成质心轨迹方程。同时给出了在统一坐标系中的多步规划质心轨迹方程。在RoboCup 3D仿真平台实现了采用自然ZMP轨迹的双足机器人类人稳定步行,实验和竞赛结果都验证了该方法的有效性。     

Observer-based dynamic walking control for biped robots

This article presents a novel observer-based control system to achieve reactive motion generation for dynamic biped walking. The proposed approach combines a feedback controller with an online generated feet pattern to assure a stable gait. Using the desired speed of the robot, a preview control system derives the dynamics of the robot’s body, and thereby the trajectory of its center of mass, to ensure a zero moment point (ZMP) movement, which results in a stable execution of the calculated step pattern. Extending the control system by an observer, based on this knowledge and the measured sensor values, compensates for errors in the model parameters and disturbances encountering while walking.     

两足步行机器人动态步行姿态稳定性及姿态控制

本文提出了一种新型两足步行机器人动态步行实时控制方法，引入了姿态稳定性和步态稳定性的概念，把两足动态步行控制分为姿态稳定性控制和步态稳定性控制两个相互联系的子系统，并从姿态稳定性分析出发，重点研究了各关节轨迹跟踪控制和系统整体动态平衡等问题，提出了姿态控制器的结构及设计方法．