Bilateral Control System of Flexible Master-Slave Arms with Random Delay Using Kalman Filter-猫眼课题宝

DOI：

10.5687/sss.2016.88

发表时间：

2016-05

期刊：

影响因子：

通讯作者：

中科院分区：

其他

文献类型：

作者：

M. Yagi;Kengo Kimura;Y. Sawada

机器人系统的远程操作技术对于向远程地点提供人类技能非常重要。近年来，这些技术引起了人们的广泛关注。此外，这些技术预计将在一般环境中使用，如家庭和办公室，而不需要特殊设备。因此，遥操作技术应该通过使用现有的通信网络，如局域网（LAN），广域网（WAN）和无线局域网实现。在这项研究中，双边控制系统作为一种类型的遥操作技术进行了研究。该系统由一个刚性的主臂，一个灵活的从臂，和一个通信网络，这将导致随机延迟。柔性从臂跟踪刚性主臂，为此，通过通信网络传输诸如状态信号和观测信号的信号。因此，具有通信网络的双边控制系统被认为是具有通信网络的反馈系统的类型。此外，由于通信网络引起随机延迟，通过该网络传输的信号变得有噪声，因此不能可靠地获得精确的观测信号。由此产生的随机延迟可能会导致系统不稳定。许多研究者对具有时滞的遥操作系统、双向控制系统和柔性机械臂进行了研究。纳梅里川讨论了一种简单的比例微分（PD）型控制方法和具有时变时延的遥操作系统的控制策略[1]，[2]。Mori等人和Hoshino等人证明了使用对称双边控制器控制的柔性主从机械手的无源性[3]，[4]。Matsuno等人设计了一个比例微分和应变（PDS）反馈控制器来控制两连杆柔性梁，并证明了闭环系统的稳定性[5]。然而，这些研究没有考虑随机延迟。对于更广泛的随机时延，具有随机时延的网络系统的状态估计问题已经被许多研究者所研究。Liu等人研究了闭环网络控制系统的模型预测控制问题。本文将随机时延定义为马尔可夫链，然后将闭环网络控制系统表示为马尔可夫跳变系统[6]。Wu等人证明了具有长随机时滞的网络系统的均方指数稳定性[7]，Guo等人研究了一类具有随机时滞的离散时间系统的指数稳定性（均方意义下）[8]。这些研究人员提出了滤波器来估计网络控制系统的状态。Schenato研究了具有随机延迟和分组丢失的最优估计器的设计[9]，[10]。然而，由于这些研究中的随机延迟不是随机产生的（例如，作为高斯噪声），认为这些随机延迟模型不能很好地模拟实际随机延迟。在我们以前的研究中，双边控制系统的时变时滞进行了研究。通过使用李雅普诺夫定理证明了稳定性和无源性，并通过数值模拟评估了所提出的双边控制系统的性能[14]。在此基础上，研究了具有随机时延的网络化系统的状态估计问题。针对时不变线性系统和受随机时滞影响的观测系统，设计了卡尔曼滤波器。通过数值模拟证实了所提出的卡尔曼滤波器的有效性[15]。本文讨论了具有随机延迟的柔性主从机械臂双向控制系统。利用汉密尔顿原理推导了主、从臂的线性化数学模型，将随机时延定义为平均时延和高斯噪声之和。为了估计受随机时延影响的状态和观测信号，设计了一种新的卡尔曼滤波器。PD控制器和PDS控制器被设计用于产生刚性主臂的反作用扭矩，并且参考Proceedings of the 47 th ISCIE International Symposium on Stochastic Systems Theory and Its Applications檀香山，Dec. 5-8，2015

Teleoperation technologies for robotic systems are important for enabling human skills to be provided to remote locations. In recent years, these technologies have attracted much attention. Furthermore, these technologies are expected to be useful in general environments such as houses and offices without requiring special equipment. Therefore, teleoperation technologies should be realized by using existing communication networks, such as local area networks (LAN), wide area networks (WAN), and wireless LANs. In this study, a bilateral control system is investigated as one type of teleoperation technology. This system consists of a rigid master arm, a flexible slave arm, and a communication network, which causes random delay. The flexible slave arm tracks the rigid master arm. For this purpose, signals such as state signals and observation signals are transmitted through the communication network. Therefore, the bilateral control system with the communication network is considered a type of feedback system with a communication network. Furthermore, because the communication network causes random delay, the signals transmitted through this network become noisy, and so accurate observation signals are not reliable obtained. The random delay that results might cause system instabilities. Many researchers have studied teleoperation systems with time delay, bilateral control systems, and flexible manipulators. Namerikawa discussed a simple proportional derivative (PD)-type control method and a control strategy for teleoperation systems with timevarying delay [1],[2]. Mori et al. and Hoshino et al. proved the passivity of a flexible master-slave manipulator that is controlled using a symmetric bilateral controller [3],[4]. Matsuno et al. designed a proportional derivative and strain (PDS) feedback controller to control the two-link flexible beams and proved the stability of the closed-loop system [5]. However, these studies did not consider random delay. For random delay more broadly, state estimation problems for a networked system with random delay have been studied by many researchers. Liu et al. investigated the model predictive control problem for closed-loop networked control systems. In this study, a random delay is defined as a Markov chain, and then the closed-loop networked control system is expressed as a Markovian jump system [6]. Wu et al. proved the mean-square exponential stability for a networked system with long random delay [7], and Guo et al. studied exponential stability (in the mean-square sense) for a type of discrete-time system with random delay [8]. These researchers proposed filters to estimate the state of networked control systems. Schenato investigated the design of optimal estimators with random delay and packet loss [9],[10]. However, because the random delay in these studies was not generated stochastically (e.g., as Gaussian noise), it is considered that these models of random delay do not well model actual random delay. In our previous studies, a bilateral control system with time-varying delay was investigated. The stability and passivity were proved by using the Lyapunov theorem, and the performance of the proposed bilateral control system was evaluated through numerical simulations [14]. Furthermore, state estimation problems for a networked system with random delay were considered. A Kalman filter was designed for a time-invariant linear system and the observation system that was affected by the random delay. The effectiveness of the proposed Kalman filter was confirmed through numerical simulation [15]. In this paper, a bilateral control system for flexible master–slave arms with random delay is discussed. The linearized mathematical models of the master and slave arms are derived by using Hamilton’s principle, and the random delay is defined as the sum of the average time delay and a Gaussian noise. To estimate the state and observation signals that are affected by the random delay, a novel Kalman filter is designed. The PD controller and PDS controller are designed for generating the reaction torque of the rigid master arm and the refProceedings of the 47th ISCIE International Symposium on Stochastic Systems Theory and Its Applications Honolulu, Dec. 5-8, 2015

Bilateral Control System of Flexible Master-Slave Arms with Random Delay Using Kalman Filter

文献摘要

相似文献