Applications of neural networks for coordinate transformations in robotics

The use of artificial neural networks is investigated for application to trajectory control problems in robotics. The relative merits of position versus velocity control is considered and a control scheme is proposed in which neural networks are used as static maps (trained off-line) to compute the inverse of the manipulator Jacobian matrix. A proof of the stability of this approach is offered, assuming bounded errors in the static map. A representative two-link robot is investigated using an artificial neural network which has been trained to compute the components of the inverse of the Jacobian matrix. The controller is implemented in the laboratory and its performance compared to a similar controller with the analytical inverse Jacobian matrix.     

Li-function activated ZNN with finite-time convergence applied to redundant-manipulator kinematic control via time-varying Jacobian matrix pseudoinversion

This paper presents and investigates the application of Zhang neural network (ZNN) activated by Li function to kinematic control of redundant robot manipulators via time-varying Jacobian matrix pseudoinversion. That is, by using Li activation function and by computing the time-varying pseudoinverse of the Jacobian matrix (of the robot manipulator), the resultant ZNN model is applied to redundant-manipulator kinematic control. Note that there are nine novelties and differences of ZNN from the conventional gradient neural network in the research methodology. More importantly, such a Li-function activated ZNN (LFAZNN) model has the property of finite-time convergence (showing its feasibility to redundant-manipulator kinematic control). Simulation results based on a four-link planar robot manipulator and a PA10 robot manipulator further demonstrate the effectiveness of the presented LFAZNN model, as well as show the LFAZNN application prospect.     

Stable neurovisual servoing for robot manipulators

In this paper, we propose a stable neurovisual servoing algorithm for set-point control of planar robot manipulators in a fixed-camera configuration an show that all the closed-loop signals are uniformly ultimately bounded (UUB) and converge exponentially to a small compact set. We assume that the gravity term and Jacobian matrix are unknown. Radial basis function neural networks (RBFNNs) with online real-time learning are proposed for compensating both gravitational forces and errors in the robot Jacobian matrix. The learning rule for updating the neural network weights, similar to a back propagation algorithm, is obtained from a Lyapunov stability analysis. Experimental results on a two degrees of freedom manipulator are presented to evaluate the proposed controller.     

Visual servoing of redundant manipulator with Jacobian matrix estimation using self-organizing map

Vision based redundant manipulator control with a neural network based learning strategy is discussed in this paper. The manipulator is visually controlled with stereo vision in an eye-to-hand configuration. A novel Kohonen’s self-organizing map (KSOM) based visual servoing scheme has been proposed for a redundant manipulator with 7 degrees of freedom (DOF). The inverse kinematic relationship of the manipulator is learned using a Kohonen’s self-organizing map. This learned map is shown to be an approximate estimate of the inverse Jacobian, which can then be used in conjunction with the proportional controller to achieve closed loop servoing in real-time. It is shown through Lyapunov stability analysis that the proposed learning based servoing scheme ensures global stability. A generalized weight update law is proposed for KSOM based inverse kinematic control, to resolve the redundancy during the learning phase. Unlike the existing visual servoing schemes, the proposed KSOM based scheme eliminates the computation of the pseudo-inverse of the Jacobian matrix in real-time. This makes the proposed algorithm computationally more efficient. The proposed scheme has been implemented on a 7 DOF PowerCube? robot manipulator with visual feedback from two cameras.     

A dual neural network for redundancy resolution of kinematically redundant manipulators subject to joint limits and joint velocity limits

In this paper, a recurrent neural network called the dual neural network is proposed for online redundancy resolution of kinematically redundant manipulators. Physical constraints such as joint limits and joint velocity limits, together with the drift-free criterion as a secondary task, are incorporated into the problem formulation of redundancy resolution. Compared to other recurrent neural networks, the dual neural network is piecewise linear and has much simpler architecture with only one layer of neurons. The dual neural network is shown to be globally (exponentially) convergent to optimal solutions. The dual neural network is simulated to control the PA10 robot manipulator with effectiveness demonstrated.     

Visual Motor Control of a 7 DOF Robot Manipulator Using Function Decomposition and Sub-Clustering in Configuration Space

This paper deals with real-time implementation of visual-motor control of a 7 degree of freedom (DOF) robot manipulator using self-organized map (SOM) based learning approach. The robot manipulator considered here is a 7 DOF PowerCube manipulator from Amtec Robotics. The primary objective is to reach a target point in the task space using only a single step movement from any arbitrary initial configuration of the robot manipulator. A new clustering algorithm using Kohonen SOM lattice has been proposed that maintains the fidelity of training data. Two different approaches have been proposed to find an inverse kinematic solution without using any orientation feedback. In the first approach, the inverse Jacobian matrices are learnt from the training data using function decomposition. It is shown that function decomposition leads to significant improvement in accuracy of inverse kinematic solution. In the second approach, a concept called sub-clustering in configuration space is suggested to provide multiple solutions for the inverse kinematic problem. Redundancy is resolved at position level using several criteria. A redundant manipulator is dexterous owing to the availability of multiple configurations for a given end-effector position. However, existing visual motor coordination schemes provide only one inverse kinematic solution for every target position even when the manipulator is kinematically redundant. Thus, the second approach provides a learning architecture that can capture redundancy from the training data. The training data are generated using explicit kinematic model of the combined robot manipulator and camera configuration. The training is carried out off-line and the trained network is used on-line to compute the joint angle vector to reach a target position in a single step only. The accuracy attained is better than the current state of art.     

Singularity-robust decoupled control of dual-elbow manipulators

The ability of a robot manipulator to move inside its workspace is inhibited by the presence of joint limits and obstacles and by the existence of singular positions in the configuration space of the manipulator. Several kinematic control strategies have been proposed to ameliorate these problems and to control the motion of the manipulator inside its workspace. The common base of these strategies is the manipulability measure which has been used to: (i) avoid singularities at the task-planning level; and (ii) to develop a singularity-robust inverse Jacobian matrix for continuous kinematic control. In this paper, a singularity-robust resolved-rate control strategy is presented for decoupled robot geometries and implemented for the dual-elbow manipulator. The proposed approach exploits the decoupled geometry of the dual-elbow manipulator to control independently the shoulder and the arm subsystems, for any desired end-effector motion, thus incurring a significantly lower computational cost compared to existing schemes.     

Design philosophy of an isotropic six-axis serial manipulator

Presented in this paper is the design philosophy employed for the constructtion of DIESTRO, an isotropic, six-axis, serial manipulator. The kinematic criteria applied so far in manipulator design have been based largely on kinematic solvability, in the sense of allowing for closed-form inverse kinematic solutions. As opposed to this rather limiting criterion, DIESTRO was designed kinematically so as to having a set of configurations in which its Jacobian matrix allows its inversion without roundoff error amplification. Although the basic kinematic chain is of the serial type, this design criterion led to an architecture not admitting closed-form inverse kinematic solutions. The central task was to produce an accurate robot under the prescribed specifications. It is believed that, under similar workspace and load specifications, the particularly challenging design of many other serial manipulators with complex architectures can benefit from the design guidelines given here.     

Indirect iterative learning control for a discrete visual servo without a camera-robot model.

This paper presents a discrete learning controller for vision-guided robot trajectory imitation with no prior knowledge of the camera-robot model. A teacher demonstrates a desired movement in front of a camera, and then, the robot is tasked to replay it by repetitive tracking. The imitation procedure is considered as a discrete tracking control problem in the image plane, with an unknown and time-varying image Jacobian matrix. Instead of updating the control signal directly, as is usually done in iterative learning control (ILC), a series of neural networks are used to approximate the unknown Jacobian matrix around every sample point in the demonstrated trajectory, and the time-varying weights of local neural networks are identified through repetitive tracking, i.e., indirect ILC. This makes repetitive segmented training possible, and a segmented training strategy is presented to retain the training trajectories solely within the effective region for neural network approximation. However, a singularity problem may occur if an unmodified neural-network-based Jacobian estimation is used to calculate the robot end-effector velocity. A new weight modification algorithm is proposed which ensures invertibility of the estimation, thus circumventing the problem. Stability is further discussed, and the relationship between the approximation capability of the neural network and the tracking accuracy is obtained. Simulations and experiments are carried out to illustrate the validity of the proposed controller for trajectory imitation of robot manipulators with unknown time-varying Jacobian matrices.     

A real-time planning algorithm for obstacle avoidance of redundant robots

A computationally efficient, obstacle avoidance algorithm for redundant robots is presented in this paper. This algorithm incorporates the neural networks and pseudodistance function D _p in the framework of resolved motion rate control. Thus, it is well suited for real-time implementation. Robot arm kinematic control is carried out by the Hopfield network. The connection weights of the network can be determined from the current value of Jacobian matrix at each sampling time, and joint velocity commands can be generated from the outputs of the network. The obstacle avoidance task is achieved by formulating the performance criterion as D _p>d _min (d _min represents the minimal distance between the redundant robot and obstacles). Its calculation is only related to some vertices which are used to model the robot and obstacles, and the computational times are nearly linear in the total number of vertices. Several simulation cases for a four-link planar manipulator are given to prove that the proposed collision-free trajectory planning scheme is efficient and practical.     

Neural networks and the inverse kinematics problem

Inverse kinematics is a fundamental problem in robotics. Past solutions for this problem have been realized through the use of various algebraic or algorithmic procedures. In this paper the use of feedforward neural networks to solve the inverse kinematics problem is examined for three different cases. A closed kinematic linkage is used for mapping input joint angles to output joint angles. A three-degree-of-freedom manipulator in 3D space is used to test mappings from both cartesian and spherical coordinates to manipulator joint coordinates. A majority of the results have average errors which fall below 1% of the robot workspace. The accuracy indicates that neural networks are an alternate method for performing the inverse kinematics estimation, thus introducing the fault-tolerant and high-speed advantages of neural networks to the inverse kinematics problem.This paper also shows the use of a new technique which reduces neural network mapping errors with the use of error compensation networks. The results of the work are put in perspective with a survey of current applications of neural networks in robotics.     

Formal analysis of the kinematic Jacobian in screw theory

As robotic systems flourish, reliability has become a topic of paramount importance in the human–robot relationship. The Jacobian matrix in screw theory underpins the design and optimization of robotic manipulators. Kernel properties of robotic manipulators, including dexterity and singularity, are characterized with the Jacobian matrix. The accurate specification and the rigorous analysis of the Jacobian matrix are indispensable in guaranteeing correct evaluation of the kinematics performance of manipulators. In this paper, a formal method for analyzing the Jacobian matrix in screw theory is presented using the higher-order logic theorem prover HOL4. Formalizations of twists and the forward kinematics are performed using the product of exponentials formula and the theory of functional matrices. To the best of our knowledge, this work is the first to formally analyze the kinematic Jacobian using theorem proving. The formal modeling and analysis of the Stanford manipulator demonstrate the effectiveness and applicability of the proposed approach to the formal verification of the kinematic properties of robotic manipulators.     

基于神经网络的冗余度TT-VGT机器人的运动学求解

应用BP神经网络对冗余度TT-VGT机器人的位姿正解进行训练学习,进而求解机器人 的位姿反解问题．根据网络模型求得机器人的一、二阶影响系数,应用神经网络求解雅可比 矩阵的伪逆．并对七重四面体的变几何桁架机器人进行了仿真计算．     

A Lagrangian network for kinematic control of redundant robotmanipulators

A recurrent neural network, called the Lagrangian network, is presented for the kinematic control of redundant robot manipulators. The optimal redundancy resolution is determined by the Lagrangian network through real-time solution to the inverse kinematics problem formulated as a quadratic optimization problem. While the signal for a desired velocity of the end-effector is fed into the inputs of the Lagrangian network, it generates the joint velocity vector of the manipulator in its outputs along with the associated Lagrange multipliers. The proposed Lagrangian network is shown to be capable of asymptotic tracking for the motion control of kinematically redundant manipulators.     

Dynamic neural controllers for induction motor

The paper reports application of recently developed adaptive control techniques based on neural networks to the induction motor control. This case study represents one of the more difficult control problems due to the complex, nonlinear, and time-varying dynamics of the motor and unavailability of full-state measurements. A partial solution is first presented based on a single input-single output (SISO) algorithm employing static multilayer perceptron (MLP) networks. A novel technique is subsequently described which is based on a recurrent neural network employed as a dynamical model of the plant. Recent stability results for this algorithm are reported. The technique is applied to multiinput-multioutput (MIMO) control of the motor. A simulation study of both methods is presented. It is argued that appropriately structured recurrent neural networks can provide conveniently parameterized dynamic models for many nonlinear systems for use in adaptive control.     

Obstacle avoidance of redundant manipulators using neural networks based reinforcement learning

This paper proposes a new approach for solving the problem of obstacle avoidance during manipulation tasks performed by redundant manipulators. The developed solution is based on a double neural network that uses Q-learning reinforcement technique. Q-learning has been applied in robotics for attaining obstacle free navigation or computing path planning problems. Most studies solve inverse kinematics and obstacle avoidance problems using variations of the classical Jacobian matrix approach, or by minimizing redundancy resolution of manipulators operating in known environments. Researchers who tried to use neural networks for solving inverse kinematics often dealt with only one obstacle present in the working field. This paper focuses on calculating inverse kinematics and obstacle avoidance for complex unknown environments, with multiple obstacles in the working field. Q-learning is used together with neural networks in order to plan and execute arm movements at each time instant. The algorithm developed for general redundant kinematic link chains has been tested on the particular case of PowerCube manipulator. Before implementing the solution on the real robot, the simulation was integrated in an immersive virtual environment for better movement analysis and safer testing. The study results show that the proposed approach has a good average speed and a satisfying target reaching success rate.     

Transputer arrays for the on-line computation of robot jacobians

On-line computation of forward and inverse Jacobian matrices is essential in robot manipulator controllers, where high-speed robot motion is required. The complexity of Jacobian calculation is such that the computational burden is large, and parallel processing is necessary if on-line computation is to be achieved. Various algorithms and parallel-processing networks suitable for this are considered. All algorithms have been implemented on transputer networks and computation times measured. The paper emphasises the importance of including communication overheads in comparisons of the computational efficiency of alternative algorithms and processor networks. Theoretical processing times based on computer cycle times and arithmetic operation counts are shown to be a false basis for comparison. Whilst considering the specific case of computation of Jacobian matrices for a robot manipulator, the paper provides a useful example of the considerations and constraints involved in distributing any algorithm across a multi-processor network.     

A neural framework for adaptive robot control

This paper investigates how dynamics in recurrent neural networks can be used to solve some specific mobile robot problems such as motion control and behavior generation. We have designed an adaptive motion control approach based on a novel recurrent neural network, called Echo state networks. The advantage is that no knowledge about the dynamic model is required, and no synaptic weight changing is needed in presence of time varying parameters in the robot. To generate the robot behavior over time, we adopted a biologically inspired approach called neural fields. Due to its dynamical properties, a neural field produces only one localized peak that indicates the optimum movement direction, which navigates a mobile robot to its goal in an unknown environment without any collisions with static or moving obstacles.     

Fault detection and isolation in robotic manipulators via neural networks: A comparison among three architectures for residual analysis

In this article we discuss artificial neural networks‐based fault detection and isolation (FDI) applications for robotic manipulators. The artificial neural networks (ANNs) are used for both residual generation and residual analysis. A multilayer perceptron (MLP) is employed to reproduce the dynamics of the robotic manipulator. Its outputs are compared with actual position and velocity measurements, generating the so‐called residual vector. The residuals, when properly analyzed, provides an indication of the status of the robot (normal or faulty operation). Three ANNs architectures are employed in the residual analysis. The first is a radial basis function network (RBFN) which uses the residuals of position and velocity to perform fault identification. The second is again an RBFN, except that it uses only the velocity residuals. The third is an MLP which also performs fault identification utilizing only the velocity residuals. The MLP is trained with the classical back‐propagation algorithm and the RBFN is trained with a Kohonen self‐organizing map (KSOM). We validate the concepts discussed in a thorough simulation study of a Puma 560 and with experimental results with a 3‐joint planar manipulator. © 2001 John Wiley & Sons, Inc.     

Asymptotically stable visual servoing of manipulators via neural networks

In this article we present a class of position control schemes for robot manipulators based on feedback of visual information processed through artificial neural networks. We exploit the approximation capabilities of neural networks to avoid the computation of the robot inverse kinematics as well as the inverse task space–camera mapping which involves tedious calibration procedures. Our main stability result establishes rigorously that in spite of the neural network giving an approximation of these mappings, the closed‐loop system including the robot nonlinear dynamics is locally asymptotically stable provided that the Jacobian of the neural network is nonsingular. The feasibility of the proposed neural controller is illustrated through experiments on a planar robot. © 2000 John Wiley & Sons, Inc.