HUMAN-LIKE CONTROL FOR SEMI-PASSIVE BIPEDAL ROBOT WALKING WITH VARIABLE LEG STIFFNESS1)

研究了变刚度半被动双足机器人行走控制问题。采用仿人的行走控制策略,使用变刚度双足弹簧负载倒立摆模型,利用模型自稳定性,在双支撑阶段调整后腿刚度使机器人的能量保持在期望能量附近,在单支撑阶段调整摆动腿落地位置控制质点的高度和前向速度。仿真结果表明：本文采用的控制策略可以实现双足机器人在水平面上的稳定行走,无扰动时可以使机器人实现零输入的被动周期行走,有外部扰动时腿部变刚度控制可使机器人总能量恢复平衡并重新进入稳态行走,控制系统具有鲁棒性。     

考虑摆动腿动态的变长度柔性双足机器人步态切换控制研究

弹簧负载倒立摆模型是一种典型的双足行走模型,已经成为研究机器人类人行走的基础。本文在此模型的基础上进行了扩展,通过添加刚性躯干、脚质量及采用变长度伸缩腿,充分考虑了躯干及摆动腿动力学对机器人行走步态的影响。首先,利用欧拉–拉格朗日法推导了动力学方程。其次,设计了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。第三,提出了步态切换策略,通过控制腿部长度和髋关节力矩来实现步态切换,从而改变平均行走速度。最后,通过计算机仿真验证了该方法的有效性。仿真结果表明：该控制策略能够有效地跟踪系统的期望轨迹及实现两种自然步态之间的切换,并形成稳定的极限环,实现机器人的稳定行走。     

单足机器人周期跳跃控制的虚拟约束方法

 研究单足机器人周期跳跃控制问题．弹簧支撑倒立摆模型可以比较准确地描述动物的跳跃行为，但无控制的自然跳跃抗干扰能力较差，一般采用轨迹跟踪控制方法实现单足机器人周期跳跃．当系统存在比较大的误差时，传统的时间轨迹跟踪控制方法存在明显的不足．引入虚拟约束技术，采用基于空间路径跟踪的控制方法可以克服时间轨迹跟踪的不足．采用点足机器人模型，并通过控制腿伸缩的方式为系统提供动力，将跳跃过程分为地面摆动和腾空飞行两个阶段，并通过起飞和着陆两个事件完成两个阶段之间的转换，整个系统模型属于欠驱动非光滑动力学系统．根据简化的动力学方程获得系统的虚拟约束解析表达式，并采用部分反馈线性化方法结合PD 控制设计系统的控制律．分析了系统的混合零动力学方程，并证明了闭环系统的临界稳定性．仿真结果表明，提出的控制方法可以实现单足机器人的周期跳跃控制，并且对外部干扰具有较强的鲁棒性．     

高自由度双足机器人数学模型及步行控制研究

基于ZMP理论及位姿矩阵,采用了相对重心以及相对坐标系的数学描述方法,对17自由度双足机器人步行稳定性控制进行数学建模研究．所建立的数学模型能够描述机器人稳定性控制．并经过Matlab数学计算和仿真研究,验证了所建立的模型可以描述双足机器人的步行规律,在理论上达到了双足机器人步行稳定性的控制目的．最后通过VC 把整个模型封装成一个可视化系统,便于将所研究的模型应用到实际的控制中,为未来机器人的实时控制的研究提供了技术支持．     

弹性关节空间机械臂级联智能滑模控制

谐波减速器和力矩传感器等柔性元件因其独特性能而广泛应用在空间机器人关节系统中，以获取高减速比．但同时这些柔性元件的存在为空间机械臂系统引入了关节柔性，使得对其稳定控制变得更为复杂．基于此，文中讨论了基于自适应回归小波神经网络(Self-Recurrent Wavelet Neural Networks, SRWNN)的弹性关节空间机械臂系统动力学建模及级联智能滑模控制．首先，利用级联系统理论及第二类拉格朗日方法推导出了由外环刚性臂子系统和内环关节电机转子子系统组成的系统级联动力学模型；其次，为两个子系统分别设计了内、外环自适应滑模回归小波神经网络控制．外环控制算法以期望轨迹为控制量，而其控制信号作为抑制弹性关节振动的内环控制算法的控制量，整个控制系统由内、外环控制系统叠加而成；而后，基于Lyapunov稳定性理论证明了整个控制系统的稳定性并设计了自适应回归小波神经网络的各权值参数在线学习算法．所提的控制算法有效地消除了模型不确定的影响，避免了复杂的求导计算和角加速度可测的要求，同时，控制系统设计过程中未涉及惯常奇异摄动双时标分解操作，在理论上适合任意大小的关节柔性刚度．最后，系统对比仿真试验证明了所提的级联智能控制算法优于惯常基于奇异摄动法和基于柔性铰补偿奇异摄动法的控制方案．     

简谐载荷下嵌入式单壁碳纳米管动态响应的Magnus方法

基于两端固支的弹性梁模型,研究嵌入式单壁碳纳米管在横向简谐载荷作用下的非线性振动问题。利用Galerkin方法对运动微分方程进行近似处理,将原方程从非线性动力学系统转化到二阶动力学系统,对于二阶动力学方程采用Magnus级数方法进行求解。通过数值实验,分析了嵌入式单壁碳纳米管非线性振动幅频特性,根据非线性动力学理论分析了碳纳米管动态响应,结果表明倍周期分岔产生混沌。     

机器人在轨组装结构的耦合动力学与步态优化

机器人在轨移动组装空间结构是建造大型航天器最有潜力的方式之一,但机器人在结构表面作业时两者存在严重的动力学耦合效应,给空间结构的建造带来了新挑战.针对三分支机器人行走在空间柔性结构上形成的耦合动力学问题,提出一种机器人-结构耦合动力学建模与步态优化方法.首先,基于拉格朗日方程和欧拉-伯努利梁模型建立机器人-结构耦合动力学模型,该模型可用于预测机器人在结构表面行走时的耦合动力学响应.然后,基于耦合动力学方程推导出机器人运动与结构振动的关系,以降低结构振动响应为目标开展了机器人行走步态的优化研究.最后,对机器人不同蠕动步态运动方式下的空间结构动力学响应进行了数值仿真,重点分析了机器人以不同步频、不同步长以及不同抬起高度行走移动时对空间结构动力学响应的影响规律.仿真结果表明,空间结构的动力学响应与机器人的运动方式密切相关.因此在设计行走移动组装机器人的运动步态步频时应避免为空间结构固有频率的两倍,同时在保障机器人组装安全稳定的前提下应尽可能减小运动步长和抬起高度.并且通过对机器人运动步态进行优化调整可以有效抑制空间结构的振动.     

柔性关节-柔性臂空间机器人的神经网络自适应反演控制及双重柔性振动抑制

空间机器人系统的柔性主要体现在空间机器人的臂杆和连接各臂杆之间的铰关节。由于空间机器人系统结构的复杂性,以往研究人员对同时具有柔性关节和柔性臂的系统关注不够。为此探讨了参数未知柔性关节-柔性臂空间机器人系统的动力学模拟、轨迹跟踪控制算法设计和关节、臂杆双重柔性振动的主动抑制问题。首先,采用多体动力学建模方法并结合漂浮基空间机器人固有的线动量和角动量守恒动力学特性,推导了系统的动力学方程。以此为基础,考虑到空间机器人实际应用中各关节铰具有较强柔性的情况,引入一种关节柔性补偿控制器解决了传统奇异摄动法应用受关节柔性限制问题,导出了适用于控制系统算法设计的数学模型。然后,利用该模型,基于反演思想在慢时标子系统中设计神经网络自适应控制算法来补偿系统参数未知和柔性关节引起的转动误差,实现系统运动轨迹跟踪性能;针对快时标子系统,设计了鲁棒最优控制算法抑制因柔性关节及柔性臂引起的系统双重弹性振动,保证系统的稳定性。最后,通过仿真对比实验验证了所设计控制算法的有效性。     

单杆柔性机械臂的控制仿真研究

本文应用模态控制理论，对柔性机械臂的主动控制问题中的动力学模型进行了研究。在小变形假设的前提下，考虑由于横向变形而引起的轴向位称位的影响，采用拉格朗日方程建立了计及动力刚化项的动力学模型，并将ＰＤ控制理论和方法应用于刻模型。最后，对一单杆柔性机械臂的振动控制进行了计算机仿真．     

漂浮基空间机器人系统基于增广变量思想的改进变结构滑模控制

讨论载体位置与姿态均不受控制的漂浮基空间机器人系统的控制问题.首先导出了空间机器人欠驱动形式的系统动力学方程.之后借助于增广变量法,证明可以将上述系统动力学方程及系统增广广义Jacobi矩阵分别表示为一组适当选择的组合惯性参数的线性函数.以此为基础,根据具有较强鲁棒性的变结构滑模控制理论,设计了一种空间机器人惯性空间期望轨迹跟踪的改进变结构滑模控制方案.与传统变结构滑模控制相比,所提控制方案通过一次离线预估控制律中相关矩阵元素的上下限,从而避免了实时控制过程中重复计算系统动力学方程中科氏力、离心力项的麻烦,因此有效减少了计算量,更适用于机载计算机运算能力有限的空间机器人控制系统实时应用.仿真运算,证实了方法的有效性.     

Stabilization of the passive walking dynamics of the compass-gait biped robot by developing the analytical expression of the controlled Poincaré map

Nonlinear Dynamics - The compass-gait biped robot is a two-DoF legged mechanical system that has been known by its passive dynamic walking. This kind of passive biped robot is modeled by an...     

Efficient parametric excitation walking with delayed feedback control

In passive dynamic walking proposed by McGeer, mechanical energy lost by heel strike is restored by transporting potential energy to kinetic energy as walking down a slope. When energy input is larger as a slope is steeper, the bifurcation of a walking cycle occurs. In the parametric excitation walking, which is to realize passive dynamic-like walking on the level ground, the bifurcation of a walking cycle has also been observed when walking speed is fast. Recently, Asano et al. have shown that bifurcation exerts an adverse influence upon walking performance by using a rimless wheel model. In this paper, we apply the delayed feedback control (DFC), originally used in chaos control, to parametric excitation walking to suppress bifurcation. We show in numerical simulation that the proposed method makes period-two walking to period-one walking, and improves energy efficiency. In addition, the proposed method can generate a sustainable gait in the region where a biped robot cannot walk without DFC. The analyses using a Poincaré map reveal that period-one walking with DFC corresponds to an unstable periodic orbit and reveal that a robot model in this paper satisfies the sufficient condition of applicability of DFC.     

Input torque sensitivity to uncertain parameters in biped robot

Input torque is the main power to maintain bipedal walking of robot, and can be calculated from trajectory planning and dynamic modeling on biped robot. During bipedal walking, the input torque is usually required to be adjusted due to some uncertain parameters arising from objective or subjective factors in the dynamical model to maintain the pre-planned stable trajectory. Here, a planar 5-link biped robot is used as an illustrating example to investigate the effects of uncertain parameters on the input torques. Kine-matic equations of the biped robot are firstly established by the third-order spline curves based on the trajectory planning method, and the dynamic modeling is accomplished by taking both the certain and uncertain parameters into account. Next, several evaluation indices on input torques are intro-duced to perform sensitivity analysis of the input torque with respect to the uncertain parameters. Finally, based on the Monte Carlo simulation, the values of evaluation indices on input torques are presented, from which all the robot param-eters are classified into three categories, i.e., strongly sensi-tive, sensitive and almost insensitive parameters.     

A Note on a Walking Machine Model

The paper continues studies intended to find out whether it is possible to create a prototype walking machine with relatively simple components. In this connection, the control problem is solved for a two-dimensional model of biped machine. It has a torso and two telescopic legs. Each leg includes a ponderable section of constant length and an imponderable section of variable length. The machine, regarded as a system with variable constraints, implements a single-stance gait (one stance leg at a time) with a step of constant duration. The contact of the swing leg with the ground is analyzed within the framework of Carnot's theorem (perfectly inelastic impact). It is assumed that the force developed in the stance leg is due to the deformation of the leg's spring and that this deformation can be controlled. An algorithm is proposed to synthesize a control system that takes into account collisions occurring at reverse of the roles of the legs. This algorithm is based on methods of optimizing periodic systems. The algorithm is compared with approaches used by other authors     

Impactless biped walking on a slope

Walking without impacts has been considered in dynamics as a motion/force control problem. In order to avoid impacts, an approach for both the specified motion of the biped and its ground reaction forces was presented yielding a combined motion and force control problem. As an application, a walker on a horizontal plane has been considered. In this paper, it is shown how the control of the ground reaction forces and the energy consumption depend on the gradient of a slope. The biped dynamics and the constraints within the biped system and on the ground are discussed. A motion control synthesis is developed using the inverse dynamics principle proven to be most efficient for human walking research, too. The impactless walking with controlled legs is illustrated by a seven-link biped. The “flying” biped has nine degrees of freedom, with six control inputs. During locomotion, the standing leg has three scleronomic constraints, and the trunk has three rheonomic constraints. However, there are three rheonomic constraints for the prescribed leg motion or three scleronomic constraints for reaction forces of the trailing leg, respectively. The nominal control action for impactless walking can be precomputed and stored. The model proposed allows the investigation of several problems: uphill and downhill walking, optimization of step length, stiction of the feet on the slope and many more. All these findings are also of interest in biomechanics.     

Treadmill walking of the pneumatic biped Lucy: Walking at different speeds and step-lengths

Actuators with adaptable compliance are gaining interest in the field of legged robotics due to their capability to store motion energy and to exploit the natural dynamics of the system to reduce energy consumption while walking and running. To perform research on compliant actuators we have built the planar biped Lucy. The robot has six actuated joints, the ankle, knee and hip of both legs with each joint powered by two pleated pneumatic artificial muscles in an antagonistic setup. This makes it possible to control both the torque and the stiffness of the joint. Such compliant actuators are used in passive walkers to overcome friction when walking over level ground and to improve stability. Typically, this kind of robots is only designed to walk with a constant walking speed and step-length, determined by the mechanical design of the mechanism and the properties of the ground. In this paper, we show that by an appropriate control, the robot Lucy is able to walk at different speeds and step-lengths and that adding and releasing weights does not affect the stability of the robot. To perform these experiments, an automated treadmill was built Published in Prikladnaya Mekhanika, Vol. 44, No. 7, pp. 134–142, July 2008.     

A proposed hybrid neural network for position control of a walking robot

The use of a proposed recurrent neural network control system to control a four-legged walking robot is presented in this paper. The control system consists of a neural controller, a standard PD controller, and the walking robot. The robot is a planar four-legged walking robot. The proposed Neural Network (NN) is employed as an inverse controller of the robot. The NN has three layers, which are input, hybrid hidden and output layers. In addition to feedforward connections from the input layer to the hidden layer and from the hidden layer to the output layer, there is also a feedback connection from the output layer to the hidden layer and from the hidden layer to itself. The reason to use a hybrid layer is that the robot’s dynamics consists of linear and nonlinear parts. The results show that the neural-network controller can efficiently control the prescribed positions of the stance and swing legs during the double stance phase of the gait cycle after sufficient training periods. The goal of the use of this proposed neural network is to increase the robustness of the control of the dynamic walking gait of this robot in the case of external disturbances. Also, the PD controller alone and Computed Torque Method (CTM) control system are used to control the walking robot’s position for comparison.     

Self-stability of a simple walking model driven by a rhythmic signal

In this paper, we analyzed the dynamic properties of a simple walking model of a biped robot driven by a rhythmic signal from an oscillator. The oscillator receives no sensory feedback and the rhythmic signal is an open loop. The simple model consists of a hip and two legs that are connected at the hip. The leg motion is generated by a rhythmic signal. In particular, we analytically examined the stability of a periodic walking motion. We obtained approximate periodic solutions and the Jacobian matrix of a Poincaré map by the power-series expansion using a small parameter. Although the analysis was inconclusive when we used only the first order expansion, by employing the second order expansion it clarified the stability, revealing that the periodic walking motion is asymptotically stable and the simple model possesses self-stability as an inherent dynamic characteristic in walking. We also clarified the stability region with respect to model parameters such as mass ratio and walking speed.