On-line training and pruning for recursive least square algorithms

1996 ◽  
Vol 32 (23) ◽  
pp. 2152 ◽  
Author(s):  
Chi Sing Leung ◽  
Kwok Wo Wong ◽  
Pui Fai Sum ◽  
Lai Wan Chan
Keyword(s):  
Least Square ◽  
Recursive Least Square ◽  
On Line ◽  
Training And Pruning

Auto-Tuning Digital Control Based on Parameters Identification and Disturbance Observer

2006 ◽  
Vol 505-507 ◽  
pp. 529-534 ◽  
Author(s):  
Chin Sheng Chen ◽  
Yu Reng Lee
Keyword(s):  
Controller Design ◽  
Tracking Error ◽  
Parameters Identification ◽  
Least Square ◽  
Servo Motor ◽  
Recursive Least Square ◽  
Estimated Parameters ◽  
On Line ◽  
Feedforward Controller ◽  
Auto Tuning

This paper presented a digital servo driver that realizes an auto-tuning feedback and feedforward controller design using on-line parameters identification. Firstly, the variant inertia constant, damping constant and the disturbed load torque of the servo motor are estimated by the recursive least square (RLS) estimator, which is composed of an RLS estimator and a disturbance torque compensator. Furthermore, the auto-tuning algorithm of feedback and feedforward controller is realized according to the estimated parameters to match the tracking specification. The proposed auto-tuning digital servo controllers are evaluated and compared experimentally with a traditional controller on a microcomputer-controlled servo motor positioning system. The experimental results show that this auto-tuning digital servo system remarkably reduces the tracking error.

scholarly journals Recursive Least Squares Filtering Algorithms for On-Line Viscoelastic Characterization of Biosamples

Actuators ◽  
2018 ◽  
Vol 7 (4) ◽  
pp. 74 ◽  
Author(s):  
Paolo Di Giamberardino ◽  
Maria Aceto ◽  
Oliviero Giannini ◽  
Matteo Verotti
Keyword(s):  
Least Square ◽  
Constitutive Law ◽  
Recursive Least Squares ◽  
Measurement Tool ◽  
Recursive Least Square ◽  
Novel Approach ◽  
Rigid Body Model ◽  
Gradient Based ◽  
On Line

The mechanical characterization of biological samples is a fundamental issue in biology and related fields, such as tissue and cell mechanics, regenerative medicine and diagnosis of diseases. In this paper, a novel approach for the identification of the stiffness and damping coefficients of biosamples is introduced. According to the proposed method, a MEMS-based microgripper in operational condition is used as a measurement tool. The mechanical model describing the dynamics of the gripper-sample system considers the pseudo-rigid body model for the microgripper, and the Kelvin–Voigt constitutive law of viscoelasticity for the sample. Then, two algorithms based on recursive least square (RLS) methods are implemented for the estimation of the mechanical coefficients, that are the forgetting factor based RLS and the normalised gradient based RLS algorithms. Numerical simulations are performed to verify the effectiveness of the proposed approach. Results confirm the feasibility of the method that enables the ability to perform simultaneously two tasks: sample manipulation and parameters identification.

Direct Adaptive Fuzzy Predictive Control and its Application to CSTR Process

2012 ◽  
Vol 241-244 ◽  
pp. 1191-1194
Author(s):  
Cheng Li Su ◽  
Ma Lina ◽  
Ping Li
Keyword(s):  
Predictive Control ◽  
Fuzzy Model ◽  
Least Square ◽  
Effective Control ◽  
Continuous Stirred Tank Reactor ◽  
Recursive Least Square ◽  
On Line ◽  
S Model ◽  
Takagi Sugeno ◽  
Fuzzy Predictive Control

In order to obtain accurate prediction model and avoid solving nonlinear programming problem, a direct adaptive predictive control (DAPC) method is proposed. Firstly, a nonlinear system was described based on Takagi-Sugeno (T-S) fuzzy models. Assuming that that the antecedent parameters of T-S models were kept, the consequent parameters were identified on-line by using the weighted recursive least square (WRLS) method. Secondly, the identified parameters of fuzzy model were used to directly receive the model predicted output with direct iterative for the T-S model. Finally, the application results for continuous stirred tank reactor (CSTR) process show that the proposed algorithm is an effective control strategy with excellent tracing ability. The proposed algorithm is a good way to resolve the two major problems, modeling and optimization, and provides a guarantee for high-precision control of nonlinear systems.

Parameters Identification of Induction Motor for Electric Vehicle Based on Least Squares Method

2011 ◽  
Vol 383-390 ◽  
pp. 648-653
Author(s):  
Jin Liang Zhang ◽  
Cun Shan Zhang
Keyword(s):  
Induction Motors ◽  
Least Squares Method ◽  
Least Square ◽  
Recursive Least Square ◽  
Output Data ◽  
Leakage Inductance ◽  
Rotor Resistance ◽  
On Line ◽  
Simulation Results ◽  
Rotor Winding

This paper deals with on-line parameter identification of induction motors (IM) by means of least square techniques .Using stator voltages, stator currents and velocity as input-output data. For analytical identification by recursive least square (RLS) algorithms, filtering of experimental data is performed by means of anticausal filters. The simulation results show that error of the parameter estimation of the rotor resistance, self inductance of the rotor winding, as well as the stator leakage inductance are less than 5% .It demonstrate the practical use of the identification method.

Direct Assessment of Effective Renal Plasma Flow from Renoscintigraphy with a Gamma Camera and an On-Line Computer

1981 ◽  
Vol 20 (06) ◽  
pp. 274-278
Author(s):  
J. Liniecki ◽  
J. Bialobrzeski ◽  
Ewa Mlodkowska ◽  
M. J. Surma
Keyword(s):  
Count Rate ◽  
Renal Plasma Flow ◽  
Early Period ◽  
Least Square Method ◽  
Least Square ◽  
Uptake Coefficient ◽  
Positive Linear Correlation ◽  
On Line ◽  
Best Fit ◽  
Plasma Activity

A concept of a kidney uptake coefficient (UC) of 131I-o-hippurate was developed by analogy from the corresponding kidney clearance of blood plasma in the early period after injection of the hippurate. The UC for each kidney was defined as the count-rate over its ROI at a time shorter than the peak in the renoscintigraphic curve divided by the integral of the count-rate curve over the "blood"-ROI. A procedure for normalization of both curves against each other was also developed. The total kidney clearance of the hippurate was determined from the function of plasma activity concentration vs. time after a single injection; the determinations were made at 5, 10, 15, 20, 30, 45, 60, 75 and 90 min after intravenous administration of 131I-o-hippurate and the best-fit curve was obtained by means of the least-square method. When the UC was related to the absolute value of the clearance a positive linear correlation was found (r = 0.922, ρ > 0.99). Using this regression equation the clearance could be estimated in reverse from the uptake coefficient calculated solely on the basis of the renoscintigraphic curves without blood sampling. The errors of the estimate are compatible with the requirement of a fast appraisal of renal function for purposes of clinical diagknosis.

Sign in / Sign up

Export Citation Format

Share Document