Il Young Song

Distributed fusion receding horizon filtering for uncertain linear stochastic systems with time-delay sensors

Abstract A new distributed fusion receding horizon filtering problem is investigated for uncertain linear stochastic systems with time-delay sensors. First, we construct a local receding horizon Kalman filter having time delays (LRHKFTDs) in both the system and measurement models. The key technique is the derivation of recursive error cross-covariance equations between LRHKFTDs in order to compute the optimal matrix fusion weights. It is the first time to present distributed fusion receding horizon filter for linear discrete-time systems with delayed sensors. It has a parallel structure that enables processing of multisensory time-delay measurements, so the calculation burden can be reduced and it is more reliable than the centralized version if some sensors turn faulty. Simulations for a multiple time-delays system show the effectiveness of the proposed filter in comparison with centralized receding horizon filter and non-receding versions.

Simultaneous pedestrian and multiple mobile robots localization using distributed extended Kalman filter

This paper is concerned with distributed extended Kalman filtering (DEKF) for simultaneous pedestrian and multiple mobile robots localization. Here, extended Kalman filter (EKF) is applied to the multiple robots for the pedestrian localization. The estimate from each robot is fused by distributed algorithm to improve the accuracy. Furthermore, we used multiple robots formation control to keep a triangle formation at the same time. The focus of this paper is to investigate the effect of the proposed algorithm on simultaneous localization accuracy. A Monte Carlo simulation result is presented to demonstrate the efficiency in localization accuracy of the distributed fusion of EKFs.

Distributed fusion receding horizon filtering in linear stochastic systems

This paper presents a distributed receding horizon filtering algorithm for multisensor continuous-time linear stochastic systems. Distributed fusion with a weighted sum structure is applied to local receding horizon Kalman filters having different horizon lengths. The fusion estimate of the state of a dynamic system represents the optimal linear fusion by weighting matrices under the minimum mean square error criterion. The key contribution of this paper lies in the derivation of the differential equations for determining the error cross-covariances between the local receding horizon Kalman filters. The subsequent application of the proposed distributed filter to a linear dynamic system within a multisensor environment demonstrates its effectiveness.

Classification of road surface status using a 94 GHz dual-channel polarimetric radiometer

In this article, we classify road surface statuses using a Bayesian classification method. This article uses principal component analysis (PCA) that combines a 94 GHz dual-channel polarimetric radiometer. The radiometer is used to investigate the behaviour of the brightness temperature (BT) of different road surface statuses in an open-air laboratory. The aim of this investigation is to characterize four different road surface classes (dry, wet, snowy and icy). Here, the BT (radiothermal emission) characteristics are measured at horizontal and vertical polarizations. For a given database of weather information (including BT, road surface temperature, wind speed, etc.), a PCA subspace is constructed, and the score vectors are classified by solving the Bayesian classification method. As a result, the road surface statuses were found to be well classified by the proposed method in real time.

Multisensory prediction fusion of nonlinear functions of the state vector in discrete-time systems

We propose two new multisensory fusion predictors for an arbitrary nonlinear function of the state vector in a discrete-time linear dynamic system. Nonlinear function of the state (NFS) represents a nonlinear multivariate functional of state variables, which can indicate useful information of the target system for automatic control. To estimate the NFS using multisensory information, we propose centralized and decentralized predictors. For multivariate polynomial NFS, we propose an effective closed-form computation procedure for the predictor design. For general NFS, the most popular procedure for the predictor design is based on the unscented transformation. We demonstrate the effectiveness and estimation accuracy of the fusion predictors on theoretical and numerical examples in multisensory environment.

Estimation fusion of nonlinear cost functions with application to multisensory Kalman filtering

Abstract This paper focuses on four fusion algorithms for the estimation of nonlinear cost function (NCF) in a multisensory environment. In multisensory filtering and control problems, NCF represents a nonlinear multivariate functional of state variables, which can indicate useful information of the target systems for automatic control. To estimate the NCF using multisensory information, we propose one centralized and three decentralized estimation fusion algorithms. For multivariate polynomial NCFs, we propose a simple closed-form computation procedure. For general NCFs, the most popular procedure for the evaluation of their estimates is based on the unscented transformation. The effectiveness and estimation accuracy of the proposed fusion algorithms are demonstrated with theoretical and numerical examples.

Distributed receding horizon filtering for mixed continuous–discrete multisensor linear stochastic systems

A new distributed receding horizon filtering algorithm for mixed continuous–discrete linear systems with different types of observations is proposed. The distributed fusion filter is formed by summation of the local receding horizon Kalman filters (LRHKFs) with matrix weights depending only on time instants. The proposed distributed filter has a parallel structure and allows parallel processing of measurements; thereby, it is more reliable than the centralized version if some sensors become faulty. Also, the selection of the receding horizon strategy makes the proposed distributed filter robust against dynamic model uncertainties. The key contribution of this paper is the derivation of the error cross-covariance equations between the LRHKFs in order to compute the optimal matrix weights. High accuracy and efficiency of the proposed distributed filter are demonstrated on the damper harmonic oscillator motion and the water tank mixing system.

Multisensor Estimation Fusion of Nonlinear Cost Functions in Mixed Continuous-Discrete Stochastic Systems

We propose centralized and distributed fusion algorithms for estimation of nonlinear cost function (NCF) in multisensory mixed continuous-discrete stochastic systems. The NCF represents a nonlinear multivariate functional of state variables. For polynomial NCFs, we propose a closed-form estimation procedure based on recursive formulas for high-order moments for a multivariate normal distribution. In general case, the unscented transformation is used for calculation of nonlinear estimates of a cost functions. To fuse local state estimates, the mixed differential difference equations for error cross-covariance between local estimates are derived. The subsequent application of the proposed fusion estimators for a multisensory environment demonstrates their effectiveness.

Efficient multisensor fusion with sliding window Kalman filtering for discrete-time uncertain systems with delays

In this study, we provide two computationally effective multisensory fusion filtering algorithms for discrete-time linear uncertain systems with state and observation time-delays. The first algorithm is shaped by algebraic forms for multirate sensor systems, and then we propose a matrix form of filtering equations using block matrices. The second algorithm is based on exact cross-covariance matrix equations. These equations are useful to compute matrix weights for fusion estimation in a multidimensional-multisensor environment. Furthermore, our proposed filtering algorithms are based on the sliding window strategy in order to achieve high estimation accuracy and stability under parametric uncertainties. The authors demonstrate the low computational complexities of the proposed fusion filtering algorithms and how the proposed algorithms robust against dynamic model uncertainties comparing with Kalman filtering with time delays.

Distributed receding horizon prediction in linear multisensor stochastic systems

This paper is concerned with distributed receding horizon prediction for continuous-time linear stochastic systems with multiple sensors. A distributed fusion with the weighted sum structure is applied to the optimal local receding horizon predictors. The distributed prediction algorithm represents the optimal linear fusion by weighting matrices under the minimum mean square criterion. The algorithm has the parallel structure and allows parallel processing of observations making it reliable since the rest faultless sensors can continue to the fusion estimation if some sensors occur faulty. The derivation of equations for error cross-covariances between the local predictors is the key of this paper. Example demonstrates effectiveness of the distributed receding horizon predictor.