Model-Independent Control of a Flexible-Joint Robot Manipulator

2009 ◽  
Vol 131 (4) ◽  
Author(s):  
Withit Chatlatanagulchai ◽  
Peter H. Meckl
Keyword(s):  
Neural Networks ◽  
Open Problem ◽  
High Performance ◽  
Robot Manipulator ◽  
Intrinsic Property ◽  
Variable Structure ◽  
System Model ◽  
Control Input ◽  
Control Laws ◽  
Flexible Joint

Flexibility at the joint of a manipulator is an intrinsic property. Even “rigid-joint” robots, in fact, possess a certain amount of flexibility. Previous experiments confirmed that joint flexibility should be explicitly included in the model when designing a high-performance controller for a manipulator because the flexibility, if not dealt with, can excite system natural frequencies and cause severe damage. However, control design for a flexible-joint robot manipulator is still an open problem. Besides being described by a complicated system model for which the passivity property does not hold, the manipulator is also underactuated, that is, the control input does not drive the link directly, but through the flexible dynamics. Our work offers another possible solution to this open problem. We use three-layer neural networks to represent the system model. Their weights are adapted in real time and from scratch, which means we do not need the mathematical model of the robot in our control algorithm. All uncertainties are handled by variable-structure control. Backstepping structure allows input efforts to be applied to each subsystem where they are needed. Control laws to adjust all adjustable parameters are devised using Lyapunov’s second method to ensure that error trajectories are globally uniformly ultimately bounded. We present two state-feedback schemes: first, when neural networks are used to represent the unknown plant, and second, when neural networks are used to represent the unknown parts of the control laws. In the former case, we also design an observer to enable us to design a control law using only output signals—the link positions. We use simulations to compare our algorithms with some other well-known techniques. We use experiments to demonstrate the practicality of our algorithms.

Towards Autonomic Computing

2011 ◽  
pp. 200-219
Author(s):  
Amar Ramdane-Cherif
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Autonomic Computing ◽  
Robot Manipulator ◽  
Joint Space ◽  
Geometrical Model ◽  
Cognitive Approach ◽  
Control Laws ◽  
Position Controller ◽  
Robotic Devices

Cognitive approach through the neural network (NN) paradigm is a critical discipline that will help bring about autonomic computing (AC). NN-related research, some involving new ways to apply control theory and control laws, can provide insight into how to run complex systems that optimize to their environments. NN is one kind of AC systems that can embody human cognitive powers and can adapt, learn, and take over certain functions previously performed by humans. In recent years, artificial neural networks have received a great deal of attention for their ability to perform nonlinear mappings. In trajectory control of robotic devices, neural networks provide a fast method of autonomously learning the relation between a set of output states and a set of input states. In this chapter, we apply the cognitive approach to solve position controller problems using an inverse geometrical model. In order to control a robot manipulator in the accomplishment of a task, trajectory planning is required in advance or in real time. The desired trajectory is usually described in Cartesian coordinates and needs to be converted to joint space for the purpose of analyzing and controlling the system behavior. In this chapter, we use a memory neural network (MNN) to solve the optimization problem concerning the inverse of the direct geometrical model of the redundant manipulator when subject to constraints. Our approach offers substantially better accuracy, avoids the computation of the inverse or pseudoinverse Jacobian matrix, and does not produce problems such as singularity, redundancy, and considerably increased computational complexity.

Model-Free Trajectory Control of a Two-Link Flexible-Joint Robot: Theory and Experiment

2005 ◽  
Author(s):  
Withit Chatlatanagulchai ◽  
Peter H. Meckl
Keyword(s):  
Control Algorithm ◽  
Variable Structure ◽  
State Feedback Control ◽  
External Disturbances ◽  
Control Laws ◽  
Flexible Joint ◽  
Model Free ◽  
The Neural Network ◽  
Lyapunov’S Second Method ◽  
The Mathematical Model

We present a state-feedback control of a two-link flexible-joint robot. First, we obtain desired control laws from Lyapunov’s second method. Then, we use three-layer neural networks to learn the unknown parts of the desired control laws. In this way, the control algorithm does not require the mathematical model representing the robot. We use a smooth variable structure controller to handle uncertainties from the neural network approximation and external disturbances. To show the effectiveness and practicality of this control algorithm, we performed an experiment on one of the robots in our laboratory.

scholarly journals Computational Complexity Reduction of Neural Networks of Brain Tumor Image Segmentation by Introducing Fermi–Dirac Correction Functions

Entropy ◽  
2021 ◽  
Vol 23 (2) ◽  
pp. 223
Author(s):  
Yen-Ling Tai ◽  
Shin-Jhe Huang ◽  
Chien-Chang Chen ◽  
Henry Horng-Shing Lu
Keyword(s):  
Neural Networks ◽  
Deep Learning ◽  
Computational Complexity ◽  
High Performance ◽  
Low Cost ◽  
Structural Complexity ◽  
Correction Function ◽  
Computational Time ◽  
Learning Methods ◽  
Band Theory

Nowadays, deep learning methods with high structural complexity and flexibility inevitably lean on the computational capability of the hardware. A platform with high-performance GPUs and large amounts of memory could support neural networks having large numbers of layers and kernels. However, naively pursuing high-cost hardware would probably drag the technical development of deep learning methods. In the article, we thus establish a new preprocessing method to reduce the computational complexity of the neural networks. Inspired by the band theory of solids in physics, we map the image space into a noninteraction physical system isomorphically and then treat image voxels as particle-like clusters. Then, we reconstruct the Fermi–Dirac distribution to be a correction function for the normalization of the voxel intensity and as a filter of insignificant cluster components. The filtered clusters at the circumstance can delineate the morphological heterogeneity of the image voxels. We used the BraTS 2019 datasets and the dimensional fusion U-net for the algorithmic validation, and the proposed Fermi–Dirac correction function exhibited comparable performance to other employed preprocessing methods. By comparing to the conventional z-score normalization function and the Gamma correction function, the proposed algorithm can save at least 38% of computational time cost under a low-cost hardware architecture. Even though the correction function of global histogram equalization has the lowest computational time among the employed correction functions, the proposed Fermi–Dirac correction function exhibits better capabilities of image augmentation and segmentation.

scholarly journals Communication Failure Resilient Distributed Neural Network for Edge Devices

Electronics ◽  
2021 ◽  
Vol 10 (14) ◽  
pp. 1614
Author(s):  
Jonghun Jeong ◽  
Jong Sung Park ◽  
Hoeseok Yang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
High Performance ◽  
State Of The Art ◽  
Wearable Devices ◽  
Communication Failure ◽  
Canadian Institute ◽  
Multiple Devices ◽  
Knowledge Distillation ◽  
Partitioning Technique

Recently, the necessity to run high-performance neural networks (NN) is increasing even in resource-constrained embedded systems such as wearable devices. However, due to the high computational and memory requirements of the NN applications, it is typically infeasible to execute them on a single device. Instead, it has been proposed to run a single NN application cooperatively on top of multiple devices, a so-called distributed neural network. In the distributed neural network, workloads of a single big NN application are distributed over multiple tiny devices. While the computation overhead could effectively be alleviated by this approach, the existing distributed NN techniques, such as MoDNN, still suffer from large traffics between the devices and vulnerability to communication failures. In order to get rid of such big communication overheads, a knowledge distillation based distributed NN, called Network of Neural Networks (NoNN), was proposed, which partitions the filters in the final convolutional layer of the original NN into multiple independent subsets and derives smaller NNs out of each subset. However, NoNN also has limitations in that the partitioning result may be unbalanced and it considerably compromises the correlation between filters in the original NN, which may result in an unacceptable accuracy degradation in case of communication failure. In this paper, in order to overcome these issues, we propose to enhance the partitioning strategy of NoNN in two aspects. First, we enhance the redundancy of the filters that are used to derive multiple smaller NNs by means of averaging to increase the immunity of the distributed NN to communication failure. Second, we propose a novel partitioning technique, modified from Eigenvector-based partitioning, to preserve the correlation between filters as much as possible while keeping the consistent number of filters distributed to each device. Throughout extensive experiments with the CIFAR-100 (Canadian Institute For Advanced Research-100) dataset, it has been observed that the proposed approach maintains high inference accuracy (over 70%, 1.53× improvement over the state-of-the-art approach), on average, even when a half of eight devices in a distributed NN fail to deliver their partial inference results.

A Unified Approach to the Design of Adaptive and Repetitive Controllers for Robotic Manipulators

1990 ◽  
Vol 112 (4) ◽  
pp. 618-629 ◽  
Author(s):  
Nader Sadegh ◽  
Roberto Horowitz ◽  
Wei-Wen Kao ◽  
Masayoshi Tomizuka
Keyword(s):  
Lyapunov Theory ◽  
Control Input ◽  
Robot Arm ◽  
Unified Approach ◽  
Parameter Adaptation ◽  
Control Laws ◽  
Position Errors ◽  
Adaptive Compensator ◽  
Mechanical Manipulators ◽  
Repetitive Controller

A unified approach, based on Lyapunov theory, for synthesis and stability analysis of adaptive and repetitive controllers for mechanical manipulators is presented. This approach utilizes the passivity properties of the manipulator dynamics to derive control laws which guarantee asymptotic trajectory following, without requiring exact knowledge of the manipulator dynamic parameters. The manipulator overall controller consists of a fixed PD action and an adaptive and/or repetitive action for feed-forward compensations. The nonlinear feedforward compensation is adjusted utilizing a linear combination of the tracking velocity and position errors. The repetitive compensator is recommended for tasks in which the desired trajectory is periodic. The repetitive control input is adjusted periodically without requiring knowledge of the explicit structure of the manipulator model. The adaptive compensator, on the other hand, may be used for more general trajectories. However, explicit information regarding the dynamic model structure is required in the parameter adaptation. For discrete time implementations, a hybrid version of the repetitive controller is derived and its global stability is proven. A simulation study is conducted to evaluate the performance of the repetitive controller, and its hybrid version. The hybrid repetitive controller is also implemented in the Berkeley/NSK SCARA type robot arm.

Control of Fractional-Order Systems Using Chatter-Free Sliding Mode Approach

2014 ◽  
Vol 9 (3) ◽  
Author(s):  
Mohammad Pourmahmood Aghababa
Keyword(s):  
Fractional Derivative ◽  
Sliding Mode ◽  
Variable Structure ◽  
Stability And Convergence ◽  
Control Input ◽  
Fractional Systems ◽  
Sliding Motion ◽  
Efficient Performance ◽  
Sliding Manifold ◽  
Chattering Free

The problem of stabilization of nonlinear fractional systems in spite of system uncertainties is investigated in this paper. First, a proper fractional derivative type sliding manifold with desired stability and convergence properties is designed. Then, the fractional stability theory is adopted to derive a robust sliding control law to force the system trajectories to attain the proposed sliding manifold and remain on it evermore. The existence of the sliding motion is mathematically proven. Furthermore, the sign function in the control input, which is responsible to the being of harmful chattering, is transferred into the fractional derivative of the control input. Therefore, the resulted control input becomes smooth and free of the chattering. Some numerical simulations are presented to illustrate the efficient performance of the proposed chattering-free fractional variable structure controller.

A Path-Planning Strategy for Overhead Cranes With High Hoisting Speed

2002 ◽  
Author(s):  
Ho-Hoon Lee
Keyword(s):  
Path Planning ◽  
High Performance ◽  
Stability Theorem ◽  
Lyapunov Stability Theorem ◽  
Control Laws ◽  
Planning Strategy ◽  
Overhead Cranes ◽  
Time Control ◽  
Small Load ◽  
Load Swing

This paper proposes a path planning strategy for high-performance anti-swing control of overhead cranes, where the anti-swing control problem is solved as a kinematic problem. First, two anti-swing control laws, one for hoisting up and the other for hoisting down, are proposed based on the Lyapunov stability theorem. Then a new path-planning strategy is proposed based on the concept of minimum-time control and the proposed anti-swing control laws. The proposed path planning is free from the usual constraints of small load swing, slow hoisting speed, and small hoisting distance. The effectiveness of the proposed path planning is shown by computer simulation with high hoisting speed and hoisting ratio.

Sign in / Sign up

Export Citation Format

Share Document