Information fusion identification method for ARMA signal with colored measurement noise

In this paper, an innovative optimal information fusion methodology based on adaptive and robust unscented Kalman filter (UKF) for multi-sensor nonlinear stochastic systems is proposed. Based on the linear minimum variance criterion, this multi-sensor information fusion method has a two-layer architecture: at the first layer, a new adaptive UKF scheme for the time-varying noise covariance is developed and serves as a local filter to improve the adaptability together with the estimated measurement noise covariance by applying the redundant measurement noise covariance estimation, which is isolated from the state estimation; the second layer is the fusion structure to calculate the optimal matrix weights and gives the final optimal state estimations. Based on the hypothesis testing theory with the Mahalanobis distance, the new adaptive UKF scheme utilizes both the innovation and the residual sequences to adapt the process noise covariance timely. The results of the target tracking simulations indicate that the proposed method is effective under the condition of time-varying process-error and measurement noise covariance.

Download Full-text

Self-Tuning Weighted Measurement Fusion Kalman Signal Filter

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.274.579 ◽

2013 ◽

Vol 274 ◽

pp. 579-582

Author(s):

Wen Qiang Liu ◽

Gui Li Tao ◽

Ze Yuan Gu ◽

Song Li

Keyword(s):

Signal Processing ◽

Single Channel ◽

Moving Average ◽

Measurement Noise ◽

Autoregressive Moving Average ◽

Model Parameters ◽

Dead Band ◽

Identification Method ◽

Self Tuning ◽

Signal Filter

For the single channel autoregressive moving average (ARMA) signals with multisensor and a colored measurement noise, when the model parameters and noise variances are partially unknown, based on identification method and Gevers-Wouters algorithm with a dead band, a self-tuning weighted measurement fusion Kalman signal filter is presented. A simulation example applied to signal processing shows its effectiveness.

Download Full-text

Identification Method for Multichannel Multisensor ARMA Signal with Colored Measurement Noises

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.220-223.1922 ◽

2012 ◽

Vol 220-223 ◽

pp. 1922-1928

Author(s):

Ming Bo Zhang ◽

Xiao Jun Sun

Keyword(s):

Information Fusion ◽

Dead Zone ◽

Correlation Method ◽

Model Parameters ◽

Second Stage ◽

Identification Method ◽

Multi Stage ◽

Third Stage ◽

On Line ◽

Measurement Noises

For the multisensor multichannel ARMA signal with ARMA colored measurement noises and unknown model parameters and noise variances, this paper presents a kind of multi-stage identification method. At the first stage, the on-line information fusion estimator for the unknown model parameters is presented based on the Recursive Instrumental Variable (RIV) algorithm and the Recursive Extended Least Squares (RELS) algorithm, which is realized by computing the average of local estimators for model parameter. At the second stage, the on-line information fusion estimator for the unknown variances is obtained using the correlation method, which is realized by computing the average of the local estimators for noise variances. At the third stage, the information fusion parameter estimator of MA model is presented using the correlation method and the dead zone Gevers-Wouters and LS algorithms.

Download Full-text

Parameter Estimation of the FitzHugh–Nagumo Neuron Model Using Integrals Over Finite Time Periods

Journal of Computational and Nonlinear Dynamics ◽

10.1115/1.4028601 ◽

2015 ◽

Vol 10 (2) ◽

Cited By ~ 4

Author(s):

Antonio Concha ◽

Rubén Garrido

Keyword(s):

Parameter Estimation ◽

Membrane Potential ◽

Finite Time ◽

Neuron Model ◽

Wavelet Denoising ◽

Measurement Noise ◽

Identification Method ◽

Identification Technique ◽

Time Periods ◽

Derivatives Of

This paper proposes two methodologies for estimating the parameters of the FitzHugh–Nagumo (FHN) neuron model. The identification procedures use only measurements of the membrane potential. The first technique is named the identification method based on integrals and wavelets (IMIW), which combines a parameterization based on integrals over finite time periods and a wavelet denoising technique for removing the measurement noise. The second technique, termed as the identification method based only on integrals (IMOI), does not use any wavelet denoising technique and attenuates the measurement noise by integrating the IMIW parameterization two times more over finite time periods. Both procedures use the least squares algorithm for estimating the FHN parameters. Integrating the FHN model over finite time periods allows eliminating the unmeasurable recovery variable of this model, thus obtaining a parameterization based on integrals of the measurable membrane potential variable. Unlike an identification technique recently published, the proposed methods do not rely on the time derivatives of the membrane potential and are not limited to continuously differentiable input current stimulus. Numerical simulations show that both the IMIW and IMOI have a good and a similar performance, however, the implementation of the latter is simpler than the implementation of the former.

Download Full-text