For the multisensor multi-channel autoregressive moving average (ARMA) signals with unknown parameters and noise variances, using the modern time series analysis method, based on the on-line identification of the local ARMA innovation models and fused moving average (MA) innovation model, a class of self-tuning weighted measurement fusion filter and smoother are presented. By using the dynamic error system analysis (DESA) method, it is rigorously proved that the self-tuning signal fusers converge to the optimal signal fusers in a realization. They can reduce the computational burden, and have asymptotic global optimality. A simulation example shows its effectiveness.
For the multisensor system with different measurement matrices, correlated measurement noises and unknown noise variances, by correlated method, the online identifiers of the noise variances are obtained. Based on ARMA innovation model, a self-tuning weighted measurement fusion Kalman filter is presented, which avoids Lyapunov and Riccati equations, reduces the computational burden and is suitable for real time application. By dynamic error system analysis (DESA) method, it is rigorously proved that the proposed self-tuning fused Kalman filter converges to the corresponding optimal fused Kalman filter with probability one or in a realization, i.e. it has asymptotical global optimality. A simulation example for a target tracking systems with 3 sensors shows its effectiveness.
It is troublesome to obtain the fused state estimation for multi-sensor multi-delay systems with correlated noises. The previous conventional fused estimation method uses the information, sent by several different filter, smoothers and predictors, to get one-step estimation, which increases the complexity of the method, so that it is not suitable for real application. In order to get more convenient estimators, the augmented state equation is introduced, and then an augmented steady-state Kalman estimator can be got, which conceals the time delays. Extracting the partial component of that augmented steady-state estimator yields the suboptimal estimator, which ignores the correlation between the components of the augmented estimator, but possesses more excellent rapidity and convenience compared with the previous fused estimator. Then by Sequential Covariance Intersection (CI) fusion method, a fast fusion steady-state suboptimal Kalman filter is obtained. Simulation examples show that although the proposed fusion steady-state estimator is suboptimal, its accuracy is higher than each of the local estimators and approximate to the optimal information fusion estimator.
Using the modern time series analysis method, by the left-coprime factorization, the autoregressive moving average (ARMA) innovation model is constructed, by which two measurement fusion steady-state Kalman filtering algorithms are presented. They have asymptotically global optimality. A numerical simulation example for threesensor tracking system verifies their functional equivalence to the centralized fusion steady-state Kalman filtering algorithms based on the ARMA innovation model and based on the Riccati equation by the classical Kalman filtering method.
This paper develops a new state prediction algorithm for the multisensor linear stochastic descriptor system with same measurement matrix and with correlated noises. Firstly, the fused measurement is obtained based on the least square method. And the fused descriptor system is transformed to two reduced-order non-descriptor subsystem by the singular value decomposition (SVD) method. Finally, for the fused reduced-order non-descriptor subsystem, the weighted measurement fusion(WMF) Kalman predictor based on the information matrix method is presented, which can avoid solving the Riccati equation in the classical Kalman prediction method. Then, the WMF Kalman predictor and its prediction error variance for the original multisensor descriptor system are presented, according to the relationship between the original descriptor system and the reduced-order nondescriptor subsystem. The accuracy of the presented predictor is higher than that of the local predictors or state fusion predictor. A simulation example verifies the effectiveness.
For the autoregressive (AR) signals with multisensor, unknown model parameters and unknown noise variances, using the recursive extended least square (RELS) and the correlation method, the strong consistent information fusion estimators of model parameters and noise variances are presented, and then by substituting them into the optimal weighted measurement fusion Wiener filter based on the autoregressive moving average (ARMA) innovation model, a self-tuning weighted measurement fusion Wiener signal filter is presented. Further, applying the dynamic error system analysis (DESA) method, it is proved that the self-tuning fused Wiener filter converges to the optimal fused Wiener filter in a realization, so that it has asymptotically global optimality. A simulation example shows its effectiveness.
For the multisensor single channel autoregressive moving average (ARMA) signals with unknown model parameters and noise variances, using the recursive instrumental variable (RIV) and the correlated method, the strong consistent information fusion estimators of model parameters and noise variances are presented, and then substituting them into the optimal weighted measurement fusion Wiener signal filter, a self-tuning weighted measurement fusion Wiener signal filter is presented. Further, applying the dynamic error system analysis (DESA) method, it is rigorously proved that the self-tuning fused Wiener filter converges to the optimal fused Wiener filter in a realization, so that it has asymptotically global optimality. A simulation example shows its effectiveness.
Discrete-time NCSs in multiple-packet transmission with random delays were discussed. The random delays from the sensors to the controllers and from the controllers to the actuators were considered. A novel random delay model in multiple-packet transmission was proposed and the relationship of the delays and the system performance was studied. The state feedback H ∞ controller can be constructed via solving a linear matrix inequality, such that the closed-loop system with random delays exponentially mean-square stable and with a prescribed H ∞ performance bound. The result shows that an example presented here illustrates the effectiveness of the proposed method.
The revolution rule of AS in Internet is analyzed and Internet topology model is proposed. Parameters of the model according to the statistics of Internet are calculatad. Simulation results show that the topology generated by the model has characteristics that match those of Internet more closely than the topology generated by BA model.
For the multisensor linear stochastic descriptor system with correlated measurement noises, the fused measurement can be obtained based on the weighted least square (WLS) method, and the reduced-order state components are obtained applying singular value decomposition method. Then, the multisensor descriptor system is transformed to a fused reduced-order non-descriptor system with correlated noise. And the weighted measurement fusion (WMF) Kalman estimator of this reduced-order subsystem is presented. According to the relationship of the presented non-descriptor system and the original descriptor system, the WMF Kalman estimator and its estimation error variance matrix of the original multisensor descriptor system are presented. The presented WMF Kalman estimator has global optimality, and can avoid computing these cross-variances of the local Kalman estimator, compared with the state fusion method. A simulation example about three-sensors stochastic dynamic input and output systems in economy verifies the effectiveness.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)