带多丢包和滞后随机不确定系统的最优线性估计

研究了带多丢包和滞后网络化随机不确定系统的最优线性估计问题. 通过白色乘性噪声来描述系统参数的随机不确定性. 通过一组满足Bernoulli分布的随机变量来描述数据传输过程中发生的丢包和滞后现象. 应用新息分析方法, 设计了线性最小方差意义下的最优线性估值器, 包括滤波器, 预报器和平滑器. 给出了稳态估值器存在的一个充分条件. 仿真例子验证了其有效性.     

Optimal linear estimators for systems with multiple random measurement delays and packet dropouts

This article is concerned with the optimal linear estimation problem for linear discrete-time stochastic systems with possible multiple random measurement delays and packet dropouts, where the largest random delay is limited within a known bound and packet dropouts can be infinite. A new model is constructed to describe the phenomena of multiple random delays and packet dropouts by employing some random variables of Bernoulli distribution. By state augmentation, the system with random delays and packet dropouts is transferred to a system with random parameters. Based on the new model, the least mean square optimal linear estimators including filter, predictor and smoother are easily obtained via an innovation analysis approach. The estimators are recursively computed in terms of the solutions of a Riccati difference equation and a Lyapunov difference equation. A sufficient condition for the existence of the steady-state estimators is given. An example shows the effectiveness of the proposed algorithms.     

Centralized filtering and smoothing algorithms from outputs with random parameter matrices transmitted through uncertain communication channels

The least-squares linear centralized estimation problem is addressed for discrete-time signals from measured outputs whose disturbances are modeled by random parameter matrices and correlated noises. These measurements, coming from different sensors, are sent to a processing center to obtain the estimators and, due to random transmission failures, some of the data packet processed for the estimation may either contain only noise (uncertain observations), be delayed (sensor delays) or even be definitely lost (packet dropouts). Different sequences of Bernoulli random variables with known probabilities are employed to describe the multiple random transmission uncertainties of the different sensors. Using the last observation that successfully arrived when a packet is lost, the optimal linear centralized fusion estimators, including filter, multi-step predictors and fixed-point smoothers, are obtained via an innovation approach; this approach is a general and useful tool to find easily implementable recursive algorithms for the optimal linear estimators under the least-squares optimality criterion. The proposed algorithms are obtained without requiring the evolution model of the signal process, but using only the first and second-order moments of the processes involved in the measurement model.     

具有一步随机滞后和多丢包的网络系统的最优线性估计

研究了具有随机时滞和丢包的网络系统的最优线性估计问题.本文通过两个满足 Bernoulli分布的随机变量来描述网络数据传输中可能存在的一步随机滞后和多丢包现象. 并基于新息分析方法,提出了线性最小方差下的最优线性状态滤波器、预报器和平滑器. 它们通过解一个Riccati方程和一个Lyapunov方程得到.最后,给出了稳态估值器存在的一个充分条件. 并通过仿真例子验证其有效性.     

具有随机多丢包的ARMA信号最优线性估计

研究网络环境下具有随机丢包的自回归滑动平均(ARMA)信号的估计问题,其中丢包现象通过一个满足Bernoulli分布的随机变量描述.通过ARMA模型与状态空间模型的转化,将具有丢包的ARMA信号估计问题转化为具有丢包的状态空间模型的状态和白噪声估计问题.利用射影理论分别给出线性最小方差最优线性状态估值器和白噪声估值器,进而获得ARMA信号估值器.仿真结果表明,当存在数据丢失时,所提出的算法与以往基于完整数据的最优估计算法相比具有最优性和有效性.     

具有有限连续丢包网络控制系统的最优线性估值器

本文研究了观测数据和控制输入数据传输具有有限连续丢包的线性离散随机系统的最优估计问题.利用两个满足Bernoulli分布的随机变量来分别描述从传感器到估值器和从控制器到执行器之间的数据丢包现象.通过引入两组新的变量,将原系统转化为一个带有随机参数的系统.利用射影理论,提出了线性最小方差最优线性估值器,包括滤波器、预报器和平滑器.最后研究了稳态线性估值器,并给出了稳态存在的一个充分条件.仿真例子验证了算法的有效性.     

Minimum variance estimation for linear uncertain systems with one-step correlated noises and incomplete measurements

This paper deals with state estimation problem for linear uncertain systems with correlated noises and incomplete measurements. Multiplicative noises enter into state and measurement equations to account for the stochastic uncertainties. And one-step autocorrelated and cross-correlated process noises and measurement noises are taken into consideration. Using the latest received measurement to compensate lost packets, the modified multi-step random delays and packet dropout model is adopted in the present paper. By augmenting system states, measurements and new defined variables, the original system is transformed into the stochastic parameter one. On this basis, the optimal linear estimators in the minimum variance sense are designed via projection theory. They depend on the variances of multiplicative noises, the one-step correlation coefficient matrices together with the probabilities of delays and packet losses. The sufficient condition on the existence of steady-state estimators is then given. Finally, simulation results illustrate the performance of the developed algorithms.     

具有数据包丢失的离散随机不确定系统的线性最优满阶估值器

研究了具有数据包丢失和随机不确定性离散随机线性系统的状态估计问题.其中数据包丢失是随机的,且满足Bernoulli分布,系统矩阵中的随机不确定性由一个白色乘性噪声来描述.首先,通过配方方法,提出了最小均方意义下的无偏最优线性递推满阶滤波器.所提出的滤波器用到了当前时刻和最近时刻接收到的观测来保证线性最优性.与多项式滤波和增广滤波器相比,本文的滤波器具有较小的计算负担.然后,基于所获得的线性滤波器推导了线性最优预报器和平滑器.进一步研究了线性最优估值器的渐近稳定性,给出了稳态特性存在的一个充分条件.最后,通过两个仿真例子验证了所提估计算法的优越性.     

Multi-rate distributed fusion estimation for sensor networks with packet losses

This paper presents a distributed fusion estimation method for estimating states of a dynamical process observed by wireless sensor networks (WSNs) with random packet losses. It is assumed that the dynamical process is not changing too rapidly, and a multi-rate scheme by which the sensors estimate states at a faster time scale and exchange information with neighbors at a slower time scale is proposed to reduce communication costs. The estimation is performed by taking into account the random packet losses in two stages. At the first stage, every sensor in the WSN collects measurements from its neighbors to generate a local estimate, then local estimates in the neighbors are further collected at the second stage to form a fused estimate to improve estimation performance and reduce disagreements among local estimates at different sensors. Local optimal linear estimators are designed by using the orthogonal projection principle, and the fusion estimators are designed by using a fusion rule weighted by matrices in the linear minimum variance sense. Simulations of a target tracking system are given to show that the time scale of information exchange among sensors can be slower while still maintaining satisfactory estimation performance by using the developed estimation method.     

Covariance-based fusion filtering for networked systems with random transmission delays and non-consecutive losses

The distributed and centralized fusion filtering problems for multi-sensor networked systems with transmission random one-step delays and non-consecutive packet losses are addressed. The signal evolution model is not required, as only covariance information is used. The measurements of individual sensors, subject to uncertainties modeled by random matrices and correlated noises, are transmitted to local processors through different communication channels and, due to random transmission failures, some of the data packets may be delayed or even definitely lost. The random transmission delays and non-consecutive packet losses are modeled by sequences of Bernoulli variables with different probabilities. By an innovation approach, local least squares linear filtering estimators are obtained by recursive algorithms; the distributed fusion framework is then used to obtain the optimal matrix-weighted combination of the local filters, using the mean squared error as optimality criterion. Also, a recursive least squares linear estimation algorithm is designed within the centralized fusion framework.     

Optimal linear estimators for systems with random measurement delays

This paper is concerned with the optimal linear estimation problem for linear discrete-time stochastic systems with random measurement delays. A new model that describes the random delays is constructed where possible the largest delay is bounded. Based on this new model, the optimal linear estimators including filter, predictor and smoother are developed via an innovation analysis approach. The estimators are recursively computed in terms of the solutions of a Riccati difference equation and a Lyapunov difference equation. The steady-state estimators are also investigated. A sufficient condition for the convergence of the optimal linear estimators is given. A simulation example shows the effectiveness of the proposed algorithms.     

Optimal and suboptimal minimum-variance filtering in networked systems with mixed uncertainties of random sensor delays,packet dropouts and missing measurements

In this paper the Kalman filtering problem for networked stochastic linear discrete-time systems with random measurement delays, packet dropouts and missing measurements is studied. Based on a quasi Markov-chain approach, a unified/combined model is developed to accommodate random delay, packet dropout and missing measurement. Two approaches for constructing a filter via the linear matrix inequality approach are proposed. Simulation studies are then conducted to evaluate the effectiveness of the constructed estimators.     

Reduced order state estimation for linear systems with exact measurements

Necessary and sufficient conditions are derived for the existence of reduced order, steady state, optimal linear state estimators for continuous, linear, dynamical systems when some output measurements are noise free. These conditions are shown to be dual to the generalized, strengthened Legendre-Clebsch conditions for the extremal controls of a dual singular regulator when the order of singularity is at most two. The optimal estimators are designed without the use of Luenberger observers.     

Information-theoretical optimal smoothing estimators

The optimal smoothing estimators are derived based on information theory. For a continuous-time stochastic linear system it is proved that minimizing the mutual information between the smoothing estimate vector and the smoothing error vector is equivalent to minimizing the entropy of the smoothing estimation error vector. By minimizing the entropy of the smoothing estimation error vector we derive the basic equations for the optimal smoothing estimators. Using the basic equations, three kinds of optimal smoothing estimators are derived. Furthermore, the information structure of optimal-smoothing problems is clarified.     

Optimal linear estimator for discrete-time systems with random delays

In this paper,optimal estimation for discrete-time linear time-varying systems with randomly state and measurement delays is considered.By introducing a set of binary random variables,the system is con...     

基于非策略Q-学习的网络控制系统最优跟踪控制

针对具有数据包丢失的网络化控制系统跟踪控制问题,提出一种非策略Q-学习方法,完全利用可测数据,在系统模型参数未知并且网络通信存在数据丢失的情况下,实现系统以近似最优的方式跟踪目标.首先,刻画具有数据包丢失的网络控制系统,提出线性离散网络控制系统跟踪控制问题;然后,设计一个Smith预测器补偿数据包丢失对网络控制系统性能的影响,构建具有数据包丢失补偿的网络控制系统最优跟踪控制问题;最后,融合动态规划和强化学习方法,提出一种非策略Q-学习算法.算法的优点是:不要求系统模型参数已知,利用网络控制系统可测数据,学习基于预测器状态反馈的最优跟踪控制策略;并且该算法能够保证基于Q-函数的迭代Bellman方程解的无偏性.通过仿真验证所提方法的有效性.     

Multisensor optimal information fusion input white noise deconvolution estimators

The unified multisensor optimal information fusion criterion weighted by matrices is rederived in the linear minimum variance sense, where the assumption of normal distribution is avoided. Based on this fusion criterion, the optimal information fusion input white noise deconvolution estimators are presented for discrete time-varying linear stochastic control system with multiple sensors and correlated noises, which can be applied to seismic data processing in oil exploration. A three-layer fusion structure with fault tolerant property and reliability is given. The first fusion layer and the second fusion layer both have netted parallel structures to determine the first-step prediction error cross-covariance for the state and the estimation error cross-covariance for the input white noise between any two sensors at each time step, respectively. The third fusion layer is the fusion center to determine the optimal matrix weights and obtain the optimal fusion input white noise estimators. The simulation results for Bernoulli-Gaussian input white noise deconvolution estimators show the effectiveness.     

Predictive control for networked systems affected by correlated packet loss

In this paper, we consider the predictive control problem of designing receding horizon controllers for networked linear systems subject to random packet loss in the controller to actuator link. The packet dropouts are temporarily correlated in the sense that they obey a Markovian transition model. Our design task is to solve the optimal controller that minimizes a given receding horizon cost function, using the available packet loss history. Due to the correlated nature of the packet loss, standard linear quadratic regulator methods do not apply. We first present the optimal control law by considering the correlations. This controller turns out to depend on the packet loss history and would typically require a large lookup table for implementation when the Markovian order is high. To address this issue, we present and compare several suboptimal design approaches to reduce the number of control laws.     

Optimal estimators for time-averaged measurement systems

The optimal linear prediction estimator and filter are derived for dynamic systems with time-averaged measurements using the Wiener-Hopf theory. The estimates are compared with other optimal estimators for various measurement systems.     

Distributed fusion filters from uncertain measured outputs in sensor networks with random packet losses

This paper addresses the distributed fusion filtering problem for discrete-time random signals from measured outputs perturbed by random parameter matrices and correlated additive noises. These measurements are obtained by a sensor network with a given topology, where random packet dropouts occur during the data transmission through the different network communication channels. The distributed fusion estimation is accomplished in two phases. Firstly, by an innovation approach and using the last observation that successfully arrived if a packet is lost, a preliminary distributed least-squares estimator is designed at each sensor node using its own measurements and those from its neighbors. Secondly, every sensor collects the preliminary filters that are successfully received from its neighbors and fuses this information with its own one to generate the least-squares linear matrix-weighted distributed fusion estimator. The accuracy of the proposed estimators, which is measured by the estimation error covariances, is examined by a numerical simulation example.