Search

In a typical surveillance situation the number and the trajectories of targets are a priori unknown. Each measurement has an unknown source; either clutter, a target being tracked or a new target. The tracks are initialised and updated using measurements, thus both true and false tracks exist at any given time. The authors present a particle filter approach which recursively calculates the probability of target existence, which may be used as a track quality measure for the false track discrimination. Single target, joint multitarget and linear multitarget versions are presented. The algorithms assume manoeuvring (multiple trajectory propagation models) targets and state dependent probability of target detection.

Joint integrated probabilistic data association (JIPDA) is effective in automatic multitarget tracking in an environment of false and missing measurements. However, JIPDA suffers from coalescence of closely spaced tracks. This paper develops a coalescence-avoiding version of JIPDA, which is named JIPDA. To accomplish this, the descriptor system approach is used that has previously been used for the development of JPDA, i.e., a coalescence-avoiding JPDA version. The effectiveness of JIPDA is demonstrated through simulations.

Local tracking in clutter initialises and updates true and false tracks. Local false track discrimination uses a track quality measure to confirm most of the true tracks, and to terminate most of the false tracks. Confirmed tracks are transmitted for track-to-track fusion. The sets of tracks being considered for fusion may contain both true and false tracks. The authors assume that each track information also includes the track quality measure in the form of the probability of target existence information. This information is used for additional false track discrimination at the fusion centre. They also use this information to enhance the track-to-track association. They propose three different strategies for track fusion with the target existence information: the `single target', the `joint multitarget' and the `linear multitarget'.

In situations with a significant number of targets in mutual proximity (close to each other), optimal multitarget data association approach suffers from the numerical explosion. This severely limits the applicability, i.e., the number of close targets that may be reliably tracked.We propose an iterative implementation of the joint integrated probabilistic data association (JIPDA) which allows a performance/computation resources tradeoff. This approach can also be incorporated into joint integrated track splitting (JITS). The iterations start with the single target integrated probabilistic data association (IPDA) and each subsequent iteration improves the approximation towards JIPDA, reaching the optimal multitarget solution within a finite number of iterations.

We consider target tracking using a network of sensors, where some (or all) sensors may be nonlinear or even nonobservable.With nonlinear/nonobservable sensors we may locally track the measurement state. We show that increasing the local measurement state derivative (partially) compensates the nonlinearity of the local measurement state propagation. For example, given a uniform target motion, one should also track the measurement state acceleration. We also simplify the distributed fusion by avoiding the explicit track-to-track (T2T) association. The equivalent measurements (EMs) are used to both update all existing and initialize new global tracks. This approach allows distributed false track discrimination (FTD) and trajectory estimation. We apply this material to distributed tracking in clutter using the time/frequency difference of arrival (TFDOA) sensors.

Target tracking determines the number and the trajectories of targets in the surveillance area. The sensors provide an imperfect target detection and measurement environment, complicated by clutter measurements. The vast majority of published target tracking algorithms assume that the probability of target detection is independent of the trajectory state. This assumption is often incorrect, with subsequent degradation in the target tracking performance. The authors present two multitarget trackers which incorporate the state-dependent probability of target detection, each of which improves upon the state of the art capabilities of target tracking in such environments.

The authors extend the smoothing integrated probabilistic data association algorithm to multi-target tracking in clutter, or alternatively, use smoothing to improve the joint integrated probabilistic data association (JIPDA) algorithm. The predictions of forward and backward JIPDA are fused to form the smoothing prediction, which is used for smoothing multi-target data association.

A majority of target tracking algorithms are derived assuming the infinite resolution sensor(s), where each measurement has only one source. This assumption is an approximation as the sensors (e.g., radars and sonars) have finite resolution, where close measurements from multiple sources may merge into one. This disparity between assumptions and reality may significantly degrade target tracking performance. This paper proposes a multitarget tracking algorithm that assumes finite resolution sensor(s). A finite resolution variant of linear multitarget tracking is proposed to limit the computational complexity. The proposed algorithm provides both the trajectory estimate and the track quality measure enabling automatic track initiation and termination (false track discrimination) for a complete target tracking solution in nonhomogeneous clutter. A simulation study vindicates the approach in a difficult situation with 10 crossing targets in clutter that may merge measurements.

This paper presents a fixed interval smoothing multi-scan algorithm for target tracking in clutter. Both the probability of target existence and the target trajectory probability density function are calculated using all available measurements. This improves both the false track discrimination and the target trajectory estimate. The fixed interval smoothing fuses the forward and the backward multi-scan predictions, to obtain the smoothing predictions and smoothing innovations. Both trajectory estimates and the data association probabilities are calculated using the smoothing innovations. An overlapping batch procedure is described which limits the smoothing delay.

In a target tracking situation, with clutter measurements and uncertain detections, true tracks often get lost due to temporal occlusions or unfavourable detection sequences, as well as the effects of measurement noise, clutter measurements, or any combination thereof. After a detection delay a new track is then initialized and subsequently confirmed for the same target, and the track becomes segmented. The Markov chain two (MC2) model for target existence propagation includes this temporal loss of detectability and often allows the track to continue through adverse conditions. This paper integrates the MC2 model with the integrated track splitting (ITS) algorithm resulting in a significant reduction of track segmentation.

In a surveillance situation, the origin of each measurement is uncertain. Each measurement may be a clutter measurement or a target detection. In addition to the position [range and angle(s)], measurements often include Doppler velocity. The Doppler velocity measurements are often (e.g. with the linearly modulated frequency transmitted signal) correlated with the range measurements. This study deals with data association (via measurement likelihoods) and estimation in this situation, with a special consideration of the Gaussian mixture approaches. A simulation study shows the benefit of this approach.