A SVM controller for the stable walking of biped robots based on small sample sizes

Conventional machine learning methods such as neural network (NN) uses empirical risk minimization (ERM) based on infinite samples, which is disadvantageous to the gait learning control based on small sample sizes for biped robots walking in unstructured, uncertain and dynamic environments. Aiming at the stable walking control problem in the dynamic environments for biped robots, this paper puts forward a method of gait control based on support vector machines (SVM), which provides a solution for the learning control issue based on small sample sizes. The SVM is equipped with a mixed kernel function for the gait learning. Using ankle trajectory and hip trajectory as inputs, and the corresponding trunk trajectory as outputs, the SVM is trained based on small sample sizes to learn the dynamic kinematics relationships between the legs and the trunk of the biped robots. Robustness of the gait control is enhanced, which is propitious to realize the stable biped walking, and the proposed method shows superior performance when compared to SVM with radial basis function (RBF) kernels and polynomial kernels, respectively. Simulation results demonstrate the superiority of the proposed methods.     

An insect robot controlled by the emergence of gait patterns

A major problem with walking robots is how to control their walking under unpredictably changing environments. Most walking robots proposed to date can walk in limited environments in which gait patterns are kinematically but not dynamically determined in advance. This means that such robots cannot walk and adapt to changes in the world, while animals can walk flexibly and efficiently in the real world. It has been considered that flexibility and efficiency in animals originate in the pattern of emergence of control information. We have already clarified the mechanism of flexible and efficient generation of gait patterns in animals, so we have tried to make an insect robot based on these mechanisms which can walk and adapt to unpredictable changes in the environment. Since these mechanisms are quite new and are also applicable to other artificial systems, we discuss the emergence system as the control mechanism attaining the target state under the constraints of the real world. This work was presented, in part, at the Third International Symposium on Artificial Life and Robotics, Oita, Japan, January 19–21, 1998     

Demonstration and Analysis of Quadrupedal Passive Dynamic Walking

Animals, including human beings, can travel in a variety of environments adaptively. Legged locomotion makes this possible. However, legged locomotion is temporarily unstable and finding out the principle of walking is an important matter for optimum locomotion strategy or engineering applications. As one of the challenges, passive dynamic walking has been studied on this. Passive dynamic walking is a walking phenomenon in which a biped walking robot with no actuator walks down a gentle slope. The gait is very smooth (like a human) and much research has been conducted on this. Passive dynamic walking is mainly about bipedalism. Considering that there are more quadruped animals than bipeds and a four-legged robot is easier to control than a two-legged robot, quadrupedal passive dynamic walking must exist. Based on the above, we studied saggital plane quadrupedal passive dynamic walking simulation. However, it was not enough to attribute the result to the existence of quadrupedal passive dynamic walking. In this research, quadrupedal passive dynamic walking is experimentally demonstrated by the four-legged walking robot 'Quartet 4'. Furthermore, changing the type of body joint, slope angle, leg length and variety of gaits (characteristics in four-legged animals) was observed passively. Experimental data could not have enough walking time and could not change parameters continuously. Then, each gait was analyzed quantitatively by the experiment and three-dimensional simulation.     

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．     

Energy-optimal gait analysis of quadruped robots

It is important for walking robots such as quadruped robots to have an efficient gait. Since animals and insects are the basic models for most walking robots, their walking patterns are good examples. In this study, the walking energy consumption of a quadruped robot is analyzed and compared with natural animal gaits. Genetic algorithms have been applied to obtain the energy-optimal gait when the quadruped robot is walking with a set velocity. In this method, an individual in a population represents the walking pattern of the quadruped robot. The gait (individual) which consumes the least energy is considered to be the best gait (individual) in this study. The energy-optimal gait is analyzed at several walking velocities, since the amount of walking energy consumption changes if the walking velocity of the robot is changed. The results of this study can be used to decide what type of gait should be generated for a quadruped robot as its walking velocity changes. This work was presented, in part, at the Sixth International Symposium on Artificial Life and Robotics, Tokyo, Japan, January 15–17, 2001.     

四足机器人溜蹄步态动步行的研究

本文应用角速度补偿法实现了四足机器人溜蹄步态的动步行。该四足机器人是本研究室刚刚研制成功的，具有十二个自由度，关节采用滑块摆杆机构，由带有轴角编码器的伺服电机驱动，身体和脚上装有传感器，步行由计算机进行控制。     

Adaptive splitbelt treadmill walking of a biped robot using nonlinear oscillators with phase resetting

To investigate the adaptability of a biped robot controlled by nonlinear oscillators with phase resetting based on central pattern generators, we examined the walking behavior of a biped robot on a splitbelt treadmill that has two parallel belts controlled independently. In an experiment, we demonstrated the dynamic interactions among the robot mechanical system, the oscillator control system, and the environment. The robot produced stable walking on the splitbelt treadmill at various belt speeds without changing the control strategy and parameters, despite a large discrepancy between the belt speeds. This is due to modulation of the locomotor rhythm and its phase through the phase resetting mechanism, which induces the relative phase between leg movements to shift from antiphase, and causes the duty factors to be autonomously modulated depending on the speed discrepancy between the belts. Such shifts of the relative phase and modulations of the duty factors are observed during human splitbelt treadmill walking. Clarifying the mechanisms producing such adaptive splitbelt treadmill walking will lead to a better understanding of the phase resetting mechanism in the generation of adaptive locomotion in biological systems and consequently to a guiding principle for designing control systems for legged robots.     

仿蟹机器人步态分析与仿真

在仿蟹机器人的行走控制中,步态的选择对机器人的稳定快速行走具有至关重要的作用。本文对仿蟹八足机器人的基本步态进行了分类,并进一步对八足波形步态进行分析,得出八足步行机器人在采用双四足步态的行走方式时,既可以满足速度的要求,又可以保证机器人的稳定性。通过计算机软件ADAMS对所选步态进行全局仿真,结果验证了步态规划的合理性,同时得到了机器人相关物理量的变化曲线,为进一步选择电机,分析机器人系统的动态特性提供了依据。     

步行机器人的步态选择与静态平衡

本文的研究范围包括步行机器人的步态选择和四足步行机的静态行走平衡,首先,根据“有限状态理论”,引入了一个关于步态的新定义,在此基础上,研究了步态的选择准则,两个基本特性及其综合方法.最后.导出了使四足步行机保持静态稳定行走的充分和必要条件.     

基于中枢神经模式产生器机器人的控制网络

分析了基于中枢神经模式产生器(Central Pattern Generator,CPG)的仿人机器人控制网络系统结构的特点,介绍了振荡器的数学模型。研究了CPG网络中各神经元的刺激方式,采用Hopf非线性振荡器构造神经元,模仿人类的行走步态,设计一种6关节控制网络。计算仿真中该网络输出信号稳定,运动节奏符合设计要求。最后,应用一仿人机器人完成了实验,提高了其行走的速度和稳定性,验证了该网络的有效性。     

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。     

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．     

Wave CPG model for autonomous decentralized multi-legged robot: Gait generation and walking speed control

In this paper, we propose a method to control gait generation and walking speed control for an autonomous decentralized multi-legged robot by using a wave Central Pattern Generator (CPG) model. The wave CPG model is a mathematical model of nonlinear oscillators and generates rhythmic movements of the legs. The gait generation and the walking speed control are achieved by controlling the virtual energy of the oscillators (Hamiltonian). A real robot experiment showed the relationship to the Hamiltonian, the actual energy consumption and the walking speed, and the effectiveness of the proposed method was verified.     

Steady Crawl Gait Generation Algorithm for Quadruped Robots

The capability of stable walking on irregular terrain is the primary advantage of legged robots over wheeled mobile robots. However, the traditional foothold selection-based gait generation algorithms are not suitable at some points for blind robots which cannot obtain the exact terrain information. A velocity-based gait generation algorithm with real-time adaptation rules which are necessary for steady walking is suggested. Particularly, we have developed a steady crawl gait with duty factor β = 0.75. The main feature of the suggested algorithm is that it is not based on foothold selection and it can be used for the walking of blind robots on more realistic irregular terrain. The adaptation rules are the translational velocity modification to satisfy the steady gait requirement and the swing period modification to avoid the kinematic limitation. The suggested gait generation algorithm has been implemented in a simple quadruped robot that has a total of eight actuated joints on the legs. Using PD controllers for each actuated joint for the trajectory following and the adaptation algorithm of gait parameters, the steady periodic crawl gait on irregular terrain has been demonstrated.     

Crab walking of quadruped robots with a locked joint failure

Fault tolerance is an important aspect in the development of control systems for multi-legged robots since a failure in a leg may lead to a severe loss of static stability of a gait. In this paper, an algorithm for tolerating a locked joint failure is described in gait planning for a quadruped robot with crab walking. A locked joint failure is one for which a joint cannot move and is locked in place. If a failed joint is locked, the workspace of the resulting leg is constrained, but legged robots have fault tolerance capability to continue walking maintaining static stability. A strategy for fault-tolerant gaits is described and, especially, a periodic gait is presented for crab walking of a quadruped. The leg sequence and the formula of the stride length are analytically driven based on gait study and robot kinematics. The adjustment procedure from a normal gait to the proposed fault-tolerant crab gait is shown to demonstrate the applicability of the proposed scheme.     

基于中枢模式发生器的仿生机器狗步态控制的应用研究

由于人工规划产生的步态是比较僵硬的、缓慢的,缺乏灵活的自组织能力,与真正的动物步态存在很大差别;文章提出了机器狗生物步态的概念;以生物的中枢模式发生器CPG模型为核心建立仿生四足机器狗运动控制系统;根据哺乳动物的肢体运动关系,建立机器狗膝髋关节运动关系方程,并设计系统软硬件;设计的控制器能够有效地克服机器狗关节轨迹跟踪控制中耦合、力矩非线性等因素的影响,且具有自适应能力;通过仿真验证了应用于机器狗的生物CPG控制机理的控制方法是有效的。     

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.     

Central pattern generators based on Matsuoka oscillators for the locomotion of biped robots

Biologically inspired control approaches based on central pattern generators (CPGs) with neural oscillators have been drawing much attention for the purpose of generating rhythmic motion for biped robots with human-like locomotion. This article describes the design of a neural-oscillator-based gait-rhythm generator using a network of Matsuoka oscillators to generate a walking pattern for biped robots. This includes the proper consideration of the oscillator’s parameters, such as a time constant for the adaptation rate, coupling factors for mutual inhibitory connections, etc., to obtain a stable and desirable response from the network. The article examines the characteristics of a CPG network with six oscillators, and the effect of assigning symmetrical and asymmetrical coupling coefficients among oscillators within the network structure under different possible inhibitions and excitations. The kinematics and dynamics of a five-link biped robot have been modeled, and its joints are actuated through simulation by the torques output from the neural rhythm generator to generate the trajectories for hip, knee, and ankle joints. The parameters of the neural oscillators are tuned to achieve flexible trajectories. The CPG-based control strategy is implemented and tested through a simulation. This work was presented in part at the 12th International Symposium on Artificial Life and Robotics, Oita, Japan, January 25–27, 2007     

人体步行规律与仿人机器人步态规划

从仿生学角度分析了人体的步行运动规律,提出了一种基于人体运动规律的仿人机器人步态参数设定方法.首先对人体步行运动数据进行捕捉并分析,得出人体各步态参数间的函数关系,以人体步行相似性作为评价指标,提出仿人机器人步态参数的设定方法.其次,通过分析人体在步行过程中的补偿支撑脚偏航力矩的基本原理,提出了基于双臂及腰关节协调运动的仿人机器人偏航力矩补偿算法,以提高仿人机器人行走的稳定性.最后通过仿真及实验验证了所提出的步态规划方法的正确性及有效性.     

Locomotion Control of a Biped Robot Using Nonlinear Oscillators

Recently, many experiments and analyses with biped robots have been carried out. Steady walking of a biped robot implies a stable limit cycle in the state space of the robot. In the design of a locomotion control system, there are primarily three problems associated with achieving such a stable limit cycle: the design of the motion of each limb, interlimb coordination, and posture control. In addition to these problems, when environmental conditions change or disturbances are added to the robot, there is the added problem of obtaining robust walking against them. In this paper we attempt to solve these problems and propose a locomotion control system for a biped robot to achieve robust walking by the robot using nonlinear oscillators, each of which has a stable limit cycle. The nominal trajectories of each limb's joints are designed by the phases of the oscillators, and the interlimb coordination is designed by the phase relation between the oscillators. The phases of the oscillators are reset and the nominal trajectories are modified using sensory feedbacks that depend on the posture and motion of the robot to achieve stable and robust walking. We verify the effectiveness of the proposed locomotion control system, analyzing the dynamic properties of the walking motion by numerical simulations and hardware experiments. Shinya Aoi received the B.E. and M.E. degrees from the Department of Aeronautics and Astronautics, Kyoto University, Kyoto, Japan in 2001 and 2003, respectively. He is a Ph.D. candidate in the Department of Aeronautics and Astronautics, Kyoto University. Since 2003, he has been a research fellow of the Japan Society for the Promotion of Science (JSPS). His research interests include dynamics and control of robotic systems, especially legged robots. He is a member of IEEE, SICE, and RSJ. Kazuo Tsuchiya received the B.S., M.S., and Ph.D. degrees in engineering from Kyoto University, Kyoto, Japan in 1966, 1968, and 1975, respectively. From 1968 to 1990, he was a research member of Central Research Laboratory in Mitsubishi Electric Corporation, Amagasaki, Japan. From 1990 to 1995, he was a professor at the Department of Computer Controlled Machinery, Osaka University, Osaka, Japan. Since 1995, he has been a professor at the Department of Aeronautics and Astronautics, Kyoto University. His fields of research include dynamic analysis, guidance, and control of space vehicles, and nonlinear system theory for distributed autonomous systems. He is currently the principal investigator of “Research and Education on Complex Functional Mechanical Systems” under the 21st Century Center of Excellence Program (COE program of the Ministry of Education, Culture, Sports, Science and Technology, Japan).