Variable Forgetting Factor RLS Research Articles

A New Local Polynomial Modeling-Based Variable Forgetting Factor RLS Algorithm and Its Acoustic Applications

This paper proposes a new class of local polynomial modeling (LPM)-based variable forgetting factor (VFF) recursive least squares (RLS) algorithms called the LPM-based VFF RLS (LVFF-RLS) algorithms. It models the time-varying channel coefficients as local polynomials so as to obtain the expressions of the bias and variance terms in the mean square error (MSE) of the RLS algorithm. A new locally optimal VFF (LOVFF) is then derived by minimizing the resulting MSE and the theoretical analysis is found to be in good agreement with experimental results. Methods for estimating the parameters involved in this LOVFF are also developed, resulting in an improved RLS algorithm with VFF. The algorithm is further extended to include variable regularization and a QR decomposition (QRD) version which is numerically more stable and amenable to multiplier-less implementation using coordinate rotation digital computer (CORDIC) algorithm. Applications of these algorithms to frequency estimation and adaptive beamforming in time-varying speech and audio signals are also presented to illustrate the effectiveness of the proposed algorithms. Simulations show that the convergence and tracking performance of the proposed algorithms compare favorably with conventional algorithms.

Variable Forgetting Factor Recursive Least Square Control Algorithm for DSTATCOM

This paper proposes an implementation of distribution static compensator (DSTATCOM) for the three-phase distribution system. The functions of DSTATCOM are harmonics elimination, compensation of reactive power, and load balancing in power factor correction and voltage regulation modes. A variable forgetting factor recursive least square (VFFRLS)-based control algorithm is proposed for effective operation of DSTATCOM to estimate weighted values of active and reactive power components of load current. The proposed control algorithm is fast in convergence and has quick response. A prototype of DSTATCOM is developed in the laboratory using a digital signal processor (dSPACE 1104) and voltage-source converter and tested for various operating conditions under nonlinear load. The performance of DSTATCOM is found satisfactory in mitigating various power-quality problems with the VFFRLS-based control algorithm.

Modeling and state of charge estimation of lithium-ion battery

Modeling and state of charge (SOC) estimation of lithium-ion (Li-ion) battery are the key techniques of battery pack management system (BMS) and critical to its reliability and safety operation. An auto-regressive with exogenous input (ARX) model is derived from RC equivalent circuit model (ECM) due to the discrete-time characteristics of BMS. For the time-varying environmental factors and the actual battery operating conditions, a variable forgetting factor recursive least square (VFFRLS) algorithm is adopted as an adaptive parameter identification method. Based on the designed model, a SOC estimator using cubature Kalman filter (CKF) algorithm is then employed to improve estimation performance and guarantee numerical stability in the computational procedure. In the battery tests, experimental results show that CKF SOC estimator has a more accuracy estimation than extended Kalman filter (EKF) algorithm, which is widely used for Li-ion battery SOC estimation, and the maximum estimation error is about 2.3%.

Generalized Predictive Control Algorithm with Real-Time Simultaneous Modeling and Tuning

A dual adaptation generalized predictive controller (DA-GPC) capable of synchronous modeling and tuning in real time is proposed. The distinctive feature of the proposed controller is the utilization of a second order discrete time controlled auto-regressive integrated moving average (CARIMA) model to capture all possible variations of numerator dynamics via the variable forgetting factor recursive least squares (VFF-RLS) algorithm. The evolving dynamic model is then used to tune, in real time, the value of the move suppression weight. Simulation studies on two benchmark process control problems, i.e., the Van de Vusse nonlinear isothermal stirred tank reactor (with dead time) and the strong acid–base neutralization process, revealed the improved performance of the DA-GPC scheme over the traditional model adaptive GPC and the conventional GPC schemes in terms of set point tracking. Further simulation studies revealed that the DA-GPC scheme was adept in rejecting load disturbances in the process. The good performance attained was attributed to the judicious adaptation of the move suppression weight. Moreover, a user-friendly aspect of the DA-GPC scheme is that the only parameter requiring manual tuning is the VFF-RLS parameter (σ), which may span a large range without materially affecting the efficacy of this adaptive controller.

An Improved Harmonic Current Detection Method Based on Variable Forgetting Factor RLS Algorithm

The effects of compensating and restraining the power system harmonics using active power filter are determined by the detection precision and its dynamic response characters. To improve both of them, a harmonic current detection algorithm based on variable forgetting factor Recursive Least-Squares algorithm is presented. The occurrence of the dynamic process is identified firstly by the judgment condition which is given by the algorithm, and then the forgetting factor is assigned dynamically, so that the convergent speed is significantly improved. The algorithm overcomes the impact of low-pass filter of traditional p-q or ip-iq algorithm, and releases the contradiction cased by the conflicting requirements of forgetting factor value between steady process and dynamic process. So it has better dynamic performance. Simulation and experiments prove the validity and feasibility of the approaches.

Generalized predictive control with dual adaptation

In this work, the recursive least squares (RLS) algorithm, which traditionally was used in the generalized predictive controller (GPC) framework solely for model adaptation purposes, was extended to cater for auto-tuning of the controller. This new combination which eases the task of controller tuning, contains both model adaptation and auto-tuning capabilities within the same controller structure. Hereafter this scheme will be referred to as the adaptive-model based self-tuning generalized predictive control (AS-GPC). The variable forgetting factor recursive least squares (VFF-RLS) algorithm was selected to capture the dynamics of the process online for the purpose of model adaptation in the controller. Based on the evolution of the process dynamics given by the VFF-RLS algorithm in the form of first order plus dead time (FOPDT) model parameters, the move suppression weight for the AS-GPC was recalculated automatically at every time step based on existing single input single output (SISO) analytical tuning expressions originally used for offline tuning of constraint-free predictive controllers. Closed loop simulation on a validated transesterification reactor model, known for inherent nonlinearities, revealed the superiority of the proposed constrained control scheme in terms of servo and regulatory control as compared to the GPC with model adaptation only, the conventional GPC as well as the conventional PID controller. The tuning expressions used, although intended for constraint-free predictive controllers, yielded good results even in the constrained case.

Run-to-Run Fault Prediction for Semiconductor Manufacturing Process Based on Variable Forgetting Factor RLS

Run-to-Run (RtR) control has been widely used in semiconductor manufacturing industry nowadays. Recursive Least Squares (RLS) approach is an important data-driven algorithm. Both the drift disturbance and the fault will seriously affect the product quality in semiconductor manufacturing process. This paper proposes a variable forgetting factor RLS output prediction method for semiconductor manufacturing process with a fault and the disturbance without and with a drift. In order to improve the accuracy of predictions, we propose a kind of method which adjusts the forgetting factor online, making predicted values track the wide fluctuations of actual output values well, and predict the output 's future fluctuations in case of faults. It has both a strong ability to track parameters, and a small convergence estimate error. Simulation results show that this prediction method works well, has better portability and higher accuracy, can provide effective information for fault warning, fault-tolerant control and health management, furthermore will avoid huge losses in actual production.

A Robust Variable Forgetting Factor Recursive Least-Squares Algorithm for System Identification

The performance of the recursive least-squares (RLS) algorithm is governed by the forgetting factor. This parameter leads to a compromise between (1) the tracking capabilities and (2) the misadjustment and stability. In this letter, a variable forgetting factor RLS (VFF-RLS) algorithm is proposed for system identification. In general, the output of the unknown system is corrupted by a noise-like signal. This signal should be recovered in the error signal of the adaptive filter after this one converges to the true solution. This condition is used to control the value of the forgetting factor. The simulation results indicate the good performance and the robustness of the proposed algorithm.

Space-Time Joint Interference Cancellation Using Fuzzy-Inference-Based Adaptive Filtering Techniques in Frequency-Selective Multipath Channels

An adaptive minimum mean-square error (MMSE) array receiver based on the fuzzy-logic recursive least-squares (RLS) algorithm is developed for asynchronous DS-CDMA interference suppression in the presence of frequency-selective multipath fading. This receiver employs a fuzzy-logic control mechanism to perform the nonlinear mapping of the squared error and squared error variation, denoted by (,), into a forgetting factor. For the real-time applicability, a computationally efficient version of the proposed receiver is derived based on the least-mean-square (LMS) algorithm using the fuzzy-inference-controlled step-size. This receiver is capable of providing both fast convergence/tracking capability as well as small steady-state misadjustment as compared with conventional LMS- and RLS-based MMSE DS-CDMA receivers. Simulations show that the fuzzy-logic LMS and RLS algorithms outperform, respectively, other variable step-size LMS (VSS-LMS) and variable forgetting factor RLS (VFF-RLS) algorithms at least 3 dB and 1.5 dB in bit-error-rate (BER) for multipath fading channels.

Gradient-based variable forgetting factor RLS algorithm in time-varying environments

In this paper, a new control mechanism for the variable forgetting factor (VFF) of the recursive least square (RLS) adaptive algorithm is presented. The control algorithm is basically a gradient-based method of which the gradient is derived from an improved mean square error analysis of RLS. The new mean square error analysis exploits the correlation of the inverse of the correlation matrix with itself that yields improved theoretical results, especially in the transient and steady-state mean square error. It is shown that the theoretical analysis is close to simulation results for different forgetting factors and different model orders. The analysis yields a dynamic equation of mean square error that can be used to derive a dynamic equation of the gradient of mean square error to control the forgetting factor. The dynamic equation can produce a positive gradient when the error is large and a negative gradient when the error is in the steady state. Compared with other variable forgetting factor algorithms, the new control algorithm gives fast tracking and small mean square model error for different signal-to-noise ratios (SNRs).

Gradient based variable forgetting factor RLS algorithm

An accurate new variable forgetting factor recursive least-square adaptive algorithm is derived. An improved mean square behaviour analysis is presented, which shows that the theoretical analysis and the simulation results are close to each other. The control of the forgetting factor is based on the dynamic equation of the gradient of mean square error. Compared with other variable forgetting factor algorithms, the new algorithm provides fast tracking and small mean square model error, and its performance will not be degraded much even in low signal-to-noise ratios.

A variable forgetting factor RLS algorithm with application to fuzzy time-varying systems identification

A modified RLS-type adaptive algorithm with variable forgetting factor is introduced. The forgetting factor decreases or increases as the square of prediction error increases or decreases, and it converges to a constant less than one in the steady state, allowing the adaptive algorithm to track changes in the system automatically as well as produce a small steady-slate error. The convergence of the algorithm is analysed. Simulation results are provided to support the analysis and to compare the performance of the modified algorithm with the constant forgetting factor RLS algorithm of Salgado et al. (1988) and the original variable forgetting factor RLS algorithm of Fortescue et al. (1981). It is shown that the performance of the new algorithm compares favorably with these existing algorithms. Finally, the algorithm is applied to the identification of fuzzy time-varying systems, with satisfactory results