Increasingly, for many application areas, it is becoming important to include elements of nonlinearity and non-Gaussianity in order to model accurately the underlying dynamics of a physical system. Moreover, it is typically crucial to process data on-line as it arrives, both from the point of view of storage costs as well as for rapid adaptation to changing signal characteristics. In this paper, we review both optimal and suboptimal Bayesian algorithms for nonlinear/non-Gaussian tracking problems, with a focus on particle filters. Particle filters are sequential Monte Carlo methods based on point mass (or “particle”) representations of probability densities, which can be applied to any state-space model and which generalize the traditional Kalman filtering methods. Several variants of the particle filter such as SIR, ASIR, and RPF are introduced within a generic framework of the sequential importance sampling (SIS) algorithm. These are discussed and compared with the standard EKF through an illustrative example.

This work addresses the problem of tracking multiple moving targets by recursively estimating the joint multitarget probability density (JMPD). Estimation of the JMPD is done in a Bayesian framework and provides a method for tracking multiple targets which allows nonlinear target motion and measurement to state coupling as well as non-Gaussian target state densities. The JMPD technique simultaneously estimates both the target states and the number of targets in the surveillance region based on the set of measurements made. We give an implementation of the JMPD method based on particle filtering techniques and provide an adaptive sampling scheme which explicitly models the multitarget nature of the problem. We show that this implementation of the JMPD technique provides a natural way to track a collection of targets, is computationally tractable, and performs well under difficult conditions such as target crossing, convoy movement, and low measurement signal-to-noise ratio (SNR).

Positioning of moving platforms has been a technical driver for real-time applications of the particle filter (PF) in both the signal processing and the robotics communities. For this reason, we will spend some time to explain several such applications in detail, and to summarize the experiences of using the PF in

This paper addresses the problem of tracking multiple moving targets by recursively estimating the joint multitarget probability density (JMPD). Estimation of the JMPD is done in a Bayesian framework and provides a method for tracking multiple targets which allows nonlinear target motion and measure-ment to state coupling as well as non-Gaussian target state densities. The JMPD technique simultaneously estimates both the target states and the number of targets in the surveillance region based on the set of measurements made. In this paper, we give an implementation of the JMPD method based on particle filtering techniques and provide an adaptive sampling scheme which explicitly models the multitarget nature of the problem. We show that this implementation of the JMPD technique provides a natural way to track a collection of targets, is computationally tractable, and performs well under difficult conditions such as target crossing, convoy movement, and low measurement SNR.

the two great men of science that I have had the good to fortune to have known,

In the visual tracking domain, Particle Filtering (PF) can become quite inefficient when being applied into high dimensional state space. Rao-Blackwellisation [1] has been shown to be an effective method to reduce the size of the state space by marginalizing out some of the variables analytically [2]. In this paper based on our previous work [3] we propose an RBPF tracking algorithm with adaptive system noise model. Experiments using both simulation data and real data show that the proposed RBPF algorithm with adaptive noise variance improves its performance significantly over conventional Particle Filter tracking algorithm. The improvements manifest in three aspects: increased estimation accuracy, reduced variance for estimates and reduced particle numbers are needed to achieve the same level of accuracy. The last two performance improvements are evaluated in this paper using simulation data. 1.

Abstract not found

Tracking or localizing a moving target is a difficult task in a distributed sensor network, due to the lack of knowledge of the target's motion and signal noises. Most existing approaches to the problem use only sensory information and may require accurate target's motion models. In this paper, we present a MarkovJan approach that combines dynamically estimated target's motion models with received sensory information. Without a given motion model, this approach localizes a target in two steps, a location prediction step using dynamically generated motion models and a location correction step integrating sensory readings. Our experimental analysis shows that our approach leads to substantially more accurate and robust location estimations than the previous approaches using only sensory information. In addition, we characterize probabilistic conditions under which the estimation accuracy increases if more sensors are used, and the estimations converge to the target's real position asymptotically. We show an interesting relationship between the steps of location prediction and location correction, i.e., as more sensors are used, belief correction guarantees the estimations to converge to the target's real position at each step, and belief prediction accelerates the convergence.