迭代学习神经网络控制在机器人示教学习中的应用

示教学习是机器人运动技能获取的一种高效手段.当采用摄像机作为示教轨迹记录部件时,示教学习涉及如何通过反复尝试获得未知机器人摄像机模型问题.本文力图针对非线性系统重复作业中的可重复不确定性学习,提出一个迭代学习神经网络控制方案,该控制器将保证系统最大跟踪误差维持在神经网络有效近似域内.为此提出了一个适合于重复作业应用的分布式神经网络结构.该神经网络由沿期望轨线分布的一系列局部神经网络构成,每一局部神经网络对对应期望轨迹点邻域进行近似并通过重复作业完成网络训练.由于所设计的局部神经网络相互独立,因此一个全程轨迹可以通过分段训练完成,由起始段到结束段,逐段实现期望轨迹的准确跟踪.该方法在具有未知机器人摄像机模型的轨迹示教模仿中得到验证,显示了它是一种高效的训练方法,同时具有一致的误差限界能力.     

非完整移动机器人的神经网络鲁棒自适应控制

在非完整移动机器人轨迹跟踪问题中,针对机器人运动学与动力学模型的参数和非参数不确定性,提出了一种混合神经网络鲁棒自适应轨迹跟踪控制器,该控制器由运动学控制器和动力学控制器两部分组成;其中,采用了参数自适应的径向基神经网络对运动学模型的未知部分进行了建模,并采用权值在线调整的单层神经网络和自适应鲁棒控制项构成了动力学控制器;基于Lyapunov方法的设计过程保证了系统的稳定性和收敛性,仿真结果证明了算法的有效性。     

Control of a nonholonomic mobile robot using an RBF network

This article deals with handling unknown factors, such as external disturbance and unknown dynamics, for mobile robot control. We propose a radial-basis function (RBF) network-based controller to compensate for these. The stability of the proposed controller is proven using the Lyapunov function. To show the effectiveness of the proposed controller, several simulation results are presented. Through the simulations, we show that the proposed controller can overcome the modelling uncertainty and the disturbances. The proposed RBF controller also outperforms previous work from the viewpoint of computation time, which is a crucial fact for real-time applications.This work was presented in part at the 8th International Symposium on Artificial Life and Robotics, Oita, Japan, January 24–26, 2003     

基于模糊CMAC的移动机器人轨迹跟踪控制

针对受非完整约束的移动机器人的轨迹跟踪问题,提出了一种基于模糊CMAC的轨迹跟踪控制策略。该策略利用模糊CMAC神经网络逼近移动机器人动力学模型的非线性和不确定,同时与速度误差结合起来构成力矩控制器,并用滑模项来补偿不确定性扰动对系统的影响。李亚普诺夫稳定性定理保证了系统的稳定性和跟踪误差的渐近收敛,仿真结果进一步验证了所提方法的有效性。     

Robust adaptive control for a nonholonomic mobile robot with unknown parameters

A robust adaptive controller for a nonholonomic mobile robot with unknown kinematic and dynamic parameters is proposed. A kinematic controller whose output is the input of the relevant dynamic controller is provided by using the concept of backstepping. An adaptive algorithm is developed in the kinematic controller to approximate the unknown kinematic parameters, and a simple single-layer neural network is used to express the highly nonlinear robot dynamics in terms of the known and unknown parameters. In order to attenuate the effects of the uncertainties and disturbances on tracking performance, a sliding mode control term is added to the dynamic controller. In the deterministic design of feedback controllers for the uncertain dynamic systems, upper bounds on the norm of the uncertainties are an important clue to guarantee the stability of the closed-loop system. However, sometimes these upper bounds may not be easily obtained because of the complexity of the structure of the uncertainties. Thereby, simple adaptation laws are proposed to approximate upper bounds on the norm of the uncertainties to address this problem. The stability of the proposed control system is shown through the Lyapunov method. Lastly, a design example for a mobile robot with two actuated wheels is provided and the feasibility of the controller is demonstrated by numerical simulations.     

Neural network Reinforcement Learning for visual control of robot manipulators

It is known that most of the key problems in visual servo control of robots are related to the performance analysis of the system considering measurement and modeling errors. In this paper, the development and performance evaluation of a novel intelligent visual servo controller for a robot manipulator using neural network Reinforcement Learning is presented. By implementing machine learning techniques into the vision based control scheme, the robot is enabled to improve its performance online and to adapt to the changing conditions in the environment. Two different temporal difference algorithms (Q-learning and SARSA) coupled with neural networks are developed and tested through different visual control scenarios. A database of representative learning samples is employed so as to speed up the convergence of the neural network and real-time learning of robot behavior. Moreover, the visual servoing task is divided into two steps in order to ensure the visibility of the features: in the first step centering behavior of the robot is conducted using neural network Reinforcement Learning controller, while the second step involves switching control between the traditional Image Based Visual Servoing and the neural network Reinforcement Learning for enabling approaching behavior of the manipulator. The correction in robot motion is achieved with the definition of the areas of interest for the image features independently in both control steps. Various simulations are developed in order to present the robustness of the developed system regarding calibration error, modeling error, and image noise. In addition, a comparison with the traditional Image Based Visual Servoing is presented. Real world experiments on a robot manipulator with the low cost vision system demonstrate the effectiveness of the proposed approach.     

Feedback control strategies for a nonholonomic mobile robot using a nonlinear oscillator

Among control problems for mobile robots, point‐to‐point stabilization is the most challenging since it does not admit designs with smooth static state feedback laws. Stabilization strategies for mobile robots, and nonholonomic systems generally, are smooth, time‐varying or nonsmooth, time‐invariant. Time‐varying control strategies are designed with umdamped linear oscillators but their fixed structure offer limited flexibility in control design. The central theme of this paper lies in use of nonlinear oscillators for mobile robot control. Large numbers of qualitatively different control strategies can be designed using nonlinear oscillators since stiffness and damping can be functions of robot states. We demonstrate by designing two fundamentally different controllers for two‐wheeled mobile robot using two variants of a particular nonlinear oscillator. First controller is dynamic and generates smooth control action. Second controller is almost‐smooth and time‐invariant. While first controller guarantees global asymptotic stability for any desired posture of robot, second controller is stable, and converges robot from almost any posture to desired posture. The only gap in posture space is unstable equilibrium manifold of measure zero. For both control strategies we mathematically establish stability and convergence of mobile robot to desired posture. Simulation results support theoretical claims. ©1999 John Wiley & Sons, Inc.     

Adaptive output feedback tracking control of a nonholonomic mobile robot

An adaptive output feedback tracking controller for nonholonomic mobile robots is proposed to guarantee that the tracking errors are confined to an arbitrarily small ball. The major difficulties are caused by simultaneous existence of nonholonomic constraints, unknown system parameters and a quadratic term of unmeasurable states in the mobile robot dynamic system as well as their couplings. To overcome these difficulties, we propose a new adaptive control scheme including designing a new adaptive state feedback controller and two high-gain observers to estimate the unknown linear and angular velocities respectively. It is shown that the closed loop adaptive system is stable and the tracking errors are guaranteed to be within the pre-specified bounds which can be arbitrarily small. Simulation results also verify the effectiveness of the proposed scheme.     

非完整移动机器人群体的优化蜂拥控制

考虑到非完整移动机器人群体蜂拥运动过程中保持位置拓扑全局连通的性能,提出一类基于局部信息交互的优化蜂拥控制算法.利用趋向局部最小外接圆圆心位置的控制方式维持群体位置拓扑在运动过程中的全局连通性,保证群体位置的聚合;结合角度控制和贝塞尔曲线规划个体的运动路径,在路径长度阈值的限定下,通过粒子群算法求取个体的优化目标方向角,保障群体运动方向的匹配;最后给出了可行的避碰方案.     

Neural network control for a fire-fighting robot

The paper discusses the development of an associative, neural network as an on-line algorithm to train and control a fire-fighting robot. Learning is externally supervised with encoded target actions. The robot acquires basic navigation skills as well as the ability to detect a fire and to extinguish it.     

Finite‐time tracking control of a nonholonomic mobile robot

In this paper, the finite‐time tracking problem is investigated for a nonholonomic wheeled mobile robot in a fifth‐order dynamic model. We consider the whole tracking error system as a cascaded system. Two continuous global finite‐time stabilizing controllers are designed for a second‐order subsystem and a third‐order subsystem respectively. Then finite‐time stability results for cascaded systems are employed to prove that the closed‐loop system satisfies the finite‐time stability. Thus the closed‐loop system can track the reference trajectory in finite‐time when the desired velocities satisfy some conditions. In particular, we discuss the control gains selection for the third‐order finite‐time controller and give sufficient conditions by using Lyapunov and backstepping techniques. Simulation results demonstrate the effectiveness of our method. Copyright © 2009 John Wiley and Sons Asia Pte Ltd and Chinese Automatic Control Society     

Adaptive control of nonlinear PID-based analog neural networks for a nonholonomic mobile robot

An adaptive controller of nonlinear PID-based analog neural networks is developed for the velocity- and orientation-tracking control of a nonholonomic mobile robot. A superb mixture of a conventional PID controller and a neural network, which has powerful capability of continuously online learning, adaptation and tackling nonlinearity, brings us the novel nonlinear PID-based analog neural network controller. It is appropriate for a kind of plant with nonlinearity uncertainties and disturbances. Computer simulation for a differentially driven nonholonomic mobile robot is carried out in the velocity- and orientation-tracking control of the nonholonomic mobile robot. The effectiveness of the proposed control algorithm is demonstrated through the simulation experiment, which shows its superior performance and disturbance rejection.     

Neural network MBPC for mobile robot path tracking

This paper presents a way of implementing a model-based predictive controller (MBPC) for mobile robot path tracking. The method uses a non-linear model of mobile robot dynamics and thus allows an accurate prediction of the future trajectories. Constraints on the maximum attainable speeds are also considered by the algorithm. A multilayer perceptron is used to implement the MBPC. The perceptron has been trained to reproduce the MBPC bahaviour in a supervised way. Experimental results obtained when applying the neural network controller to a TRC labmate mobile platform are given in the paper.     

Distributed extremum seeking and formation control for nonholonomic mobile network

In this paper, an integrated control and optimization problem is studied in the context of formation and coverage of a cluster of nonholonomic mobile robots. In particular, each communication channel is modeled by its outage probability, and hence, connectivity is maintained if the outage probability is less than a certain threshold. The objective of the communication network is to not only maintain resilient communication quality but also extend the network coverage. An information theory based performance index is defined to quantify this control objective. Unlike most of the existing results, the proposed cooperative control design does not assume the knowledge of any gradient (of the performance index). Rather, a distributed extremum seeking algorithm is designed to optimize the connectivity and coverage of the mobile network. The proposed approach retains all the advantages of cooperative control, and it can not only perform extremum seeking individually, but also ensures a consensus of estimates between any pair of connected systems. Simulation results demonstrate effectiveness of the proposed methodology.     

非完整移动机器人鲁棒运动规划

非完整移动机器人利用传感器可以解决不确定性模型和未知环境中的许多问题. 利用移动机器人上配备的传感器的信息组合提出了一种在线视点寻求算法, 结合移动机器人的运动方程和传感器的量测方程采用扩展Kalman估计来对移动机器人的位置进行修正, 以降低运动的不确定性, 从而得到一种鲁棒的规划算法, 仿真的结果证明了上述方法是行之有效的.     

Hybrid trigonometric compound function neural networks for tracking control of a nonholonomic mobile robot

The purpose of this paper is to propose a hybrid trigonometric compound function neural network (NN) to improve the NN-based tracking control performance of a nonholonomic mobile robot with nonlinear disturbances. In the mobile robot control system, two NN controllers embedded in the closed-loop control system have the simple continuous learning and rapid convergence capability without the dynamics information of the mobile robot to realize the tracking control of the mobile robot. The neuron functions of the hidden layer in the three-layer feedforward network structure consist of the compound cosine function and the compound sine function combining a cosine or a sine function with a unipolar sigmoid function. The main advantages of this NN-based mobile robot control system are better real-time control capability and control accuracy by use of the proposed NN controllers for a nonholonomic mobile robot with nonlinear disturbances. Through simulation experiments applied to the nonholonomic mobile robot with the nonlinear disturbances of dynamics uncertainty and external disturbances, the simulation results show that the proposed NN control system of a nonholonomic mobile robot has better real-time control capability and control accuracy than the compound cosine function NN control system of a nonholonomic mobile robot and then verify the effectiveness of the proposed hybrid trigonometric compound function NN controller for improving the tracking control performance of a nonholonomic mobile robot with nonlinear disturbances.     

基于传感器的非完整移动机器人运动规划

针对非完整移动机器人在未知室内环境中提出了一种路径规划方法, 通过利用传感器对周围环境的探测和实时处理传感器数据, 以及所设计的目标寻找函数, 可以有效地完成其运动规划. 该方法能够确保移动机器人在无障碍物区或障碍物对机器人不构成危险时加速前进, 在障碍物区能够慢速绕过, 从而使得移动机器人快速且安全地到达目标位置, 仿真的结果证明了该方法的有效性.     

基于单应性矩阵的移动机器人视觉伺服切换控制

针对具有单目视觉的移动机器人,本文提出了一种基于单应性矩阵的视觉伺服控制算法,在缺乏深度信息的情况下利用视觉反馈实现了移动机器人的控制目标,即给定机器人目标位姿下拍摄得到的图像,通过视觉伺服使机器人从初始位姿准确到达目标位姿.视觉反馈环节采用单应性矩阵中的元素构造状态变量,而非利用常见的单应性分解,此外,考虑到视野约束,本文提出的算法在计算单应性矩阵时结合了单应性的传递特性,从而避免了参考目标的实时可见性.伺服环节设计了切换控制器,在满足非完整约束的同时可驱动机器人到达期望位姿.理论分析及实物仿真验证了该算法的可行性和有效性.     

Saturated stabilization and tracking of a nonholonomic mobile robot

This paper presents a framework to deal with the problem of global stabilization and global tracking control for the kinematic model of a wheeled mobile robot in the presence of input saturations. A model-based control design strategy is developed via a simple application of passivity and normalization. Saturated, Lipschitz continuous, time-varying feedback laws are obtained and illustrated in a number of compelling simulations.     

A review on egomotion by means of differential epipolar geometry applied to the movement of a mobile robot

The estimation of camera egomotion is an old problem in computer vision. Since the 1980s, many approaches based on both the discrete and the differential epipolar constraint have been proposed. The discrete case is used mainly in self-calibrated stereoscopic systems, whereas the differential case deals with a single moving camera. This article surveys several methods for 3D motion estimation unifying the mathematics convention which are then adapted to the common case of a mobile robot moving on a plane. Experimental results are given on synthetic data covering more than 0.5 million estimations. These surveyed algorithms have been programmed and are available on the Internet.