Reference model supervisory loop for neural network based adaptive control of a flexible joint with hard nonlinearities

In this paper, we propose a VLSI implementation of a control strategy based on artificial neural networks (ANN) for a positioning system with a flexible transmission element, taking into account Coulomb friction for both motor and load, and using a variable learning rate for adaptation to parameter changes and to accelerate convergence. The control structure includes an ANN that approximates the inverse of the model and a reference model that defines the desired error dynamic. Adaptation of the learning rate of the ANN is used to reduce the sensitivity of the control structure to load and motor inertia variations. The paper presents a VLSI implementation of a supervisor that adapts the neural network learning rate on a Virtex2 Pro 2VP30 Field Programmable Gate Array (FPGA) from Xilinx. A pipelined implementation was used to speed-up the process. Simulation results highlight the performance of the controller and its response.

show abstract

“…Adaptation of the parameters of the reference model (5) has also been evaluated in a paper [10] to increase the stability domain.…”

Section: Discussionmentioning

confidence: 99%

FPGA implementation of learning rate supervisory loop for neural network based adaptive control of a flexible joint

Chaoui¹,

Sicard

Fatine³

2006

2006 American Control Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…In 1997, the authors of [43] disputed the assumption of weak flexibility (or large spring factor), and they showed that, for a highly flexible FJR, the proposed method of backstepping would have better results than Spong's method which causes instability for low stiffness. However, the control of highly flexible manipulators, especially with unknown varying load is still an open trend [44,45].…”

Section: Model Promotionmentioning

confidence: 99%

A Survey on the Control of Flexible Joint Robots

Ozgoli

Taghirad²

2006

Asian Journal of Control

141

View full text Add to dashboard Cite

show abstract

“…Where: ∆ and ∆ ∆ , ∆ , ∆ is the error function [8][9][10][11][12]. When ∆ and ∆ are within a given limit, the robot joint error is an adjustable one; when they exceed this limit, the robot joint error is a nonadjustable one [13][14] .…”

Section: Joint Errors Of Robotmentioning

confidence: 99%

Research on Robot Joint Error Based on Neural Network

Huang¹,

Liao²,

Chen³

et al. 2017

dtetr

View full text Add to dashboard Cite

Since the beginning of the 21 st century, robots have played an important role in both daily life and industrial production. This paper aims to evaluate the types of robot joint errors. First, the types of robot joint errors are described, and the two characteristic indexes, ∆ and ∆ , are identified. Next, robot joint errors are divided into adjustable ones and nonadjustable ones. A robot joint error type model is established using PB neural network model. The areas of adjustable joint errors and nonadjustable joint errors are defined. Finally, reasonable suggestions are given for different types of robot joint errors.Keywords: joint error of robot adjustable and nonadjustable; PB neural network

show abstract

Reference model supervisory loop for neural network based adaptive control of a flexible joint with hard nonlinearities

Cited by 21 publications

References 7 publications

FPGA implementation of learning rate supervisory loop for neural network based adaptive control of a flexible joint

FPGA implementation of learning rate supervisory loop for neural network based adaptive control of a flexible joint

A Survey on the Control of Flexible Joint Robots

Research on Robot Joint Error Based on Neural Network

Contact Info

Product

Resources

About