Ratnasingham Tharmarasa

Large-Scale Optimal Sensor Array Management for Multitarget Tracking

In this paper, we are concerned with the problem of utilizing a large network of sensors in order to track multiple targets. Large-scale sensor array management has applications in a number of target tracking domains. For example, in ground target tracking, hundreds or even thousands of unattended ground sensors may be dropped over a large surveillance area. At any one time, it may then only be possible to utilize a very small number of the available sensors at the fusion center because of physical limitations, such as available communications bandwidth. A similar situation may arise in tracking sea-surface or underwater targets using a large network of sonobuoys. The general problem is then to select a small subset of the available sensors in order to optimize tracking performance. In a practical scenario with hundreds of sensors, the number of possible sensor combinations would make it infeasible to use enumeration in order to find the optimal solution. Motivated by this consideration, in this paper we use an efficient search technique in order to determine near-optimal sensor utilization strategies in real-time. This search technique consists of convex optimization followed by greedy local search. We consider several problem formulations and the posterior Cramer-Rao lower bound is used as the basis for network management. Simulation results illustrate the performance of the algorithms, both in terms of their real-time capability and the resulting estimation accuracy. Furthermore, in comparisons it can also be seen that the proposed solutions are near-optimal.

A Multiple-Detection Joint Probabilistic Data Association Filter

Most conventional target tracking algorithms assume that a target can generate at most one measurement per scan. However, there are tracking problems where this assumption is not valid. For example, multiple detections from a target in a scan can arise due to multipath propagation effects as in the over-the-horizon radar (OTHR). A conventional multitarget tracking algorithm will fail in these scenarios, since it cannot handle multiple target-originated measurements per scan. The Joint Probabilistic Data Association Filter (JPDAF) uses multiple measurements from a single target per scan through a weighted measurement-to-track association. However, its fundamental assumption is still one-to-one. In order to rectify this shortcoming, this paper proposes a new algorithm, called the Multiple-Detection Joint Probabilistic Data Association Filter (MD-JPDAF) for multitarget tracking, which is capable of handling multiple detections from targets per scan in the presence of clutter and missed detection. The multiple-detection pattern, which can account for many-to-one measurement set-to-track association rather than one-to-one measurement-to-track association, is used to generate multiple detection association events. The proposed algorithm exploits all the available information from measurements by combinatorial association of events that are formed to handle the possibility of multiple measurements per scan originating from a target. The MD-JPDAF is applied to a multitarget tracking scenario with an OTHR, where multiple detections occur due to different propagation paths as a result of scattering from different ionospheric layers. Experimental results show that multiple-detection pattern based probabilistic data association improves the state estimation accuracy. Furthermore, the tracking performance of the proposed filter is compared against the Posterior Cramér-Rao Lower Bound (PCRLB), which is explicitly derived for the multiple-detection scenario with a single target.

Decentralized Sensor Selection for Large-Scale Multisensor-Multitarget Tracking

The problem of sensor resource management for multitarget tracking in decentralized tracking systems is considered here. Inexpensive sensors available today are used in large numbers to monitor wide surveillance regions. However, due to frequency, power, and bandwidth limitations, there is an upper limit on the number of sensors that can be used by a fusion center (FC) at any one time. The problem is then to select the sensor subsets to be used at each sampling time in order to optimize the tracking performance (i.e., maximize the tracking accuracy of existing tracks and detect new targets as quickly as possible) under the given constraints. The architecture considered in this paper is decentralized in which there is no central fusion center (CFC); each FC communicates only with the neighboring FCs, as a result communication is restricted. In such cases each FC has to decide which sensors should be used by itself at each sampling time by considering which sensors can be used by neighboring FCs. An efficient optimization-based algorithm is proposed here to address this problem in real time. Simulation results illustrating the performance of the proposed algorithms are also presented to support its efficiency. The novelty lies in the discrete optimization formulation for large-scale sensor selection in decentralized networks.

Optimization-Based Dynamic Sensor Management for Distributed Multitarget Tracking

In this paper, the general problem of dynamic assignment of sensors to local fusion centers (LFCs) in a distributed tracking framework is considered. With technological advances, a large number of sensors can be deployed for multitarget tracking purposes. However, due to physical limitations such as frequency, power, bandwidth, and fusion center capacity, only a limited number of them can be used by each LFC. The transmission power of future sensors is anticipated to be software controllable within certain lower and upper limits. Thus, the frequency reusability and the sensor reachability can be improved by controlling transmission powers. Then, the problem is to select the sensor subsets that should be used by each LFC and to find their transmission frequencies and powers in order to maximize the tracking accuracies and minimize the total power consumption. The frequency channel limitation and the advantage of variable transmitting power have not been discussed in the literature. In this paper, the optimal formulation for the aforementioned sensor management problem is provided based on the posterior Cramer-Rao lower bound. Finding the optimal solution to the aforementioned NP-hard multiobjective mixed-integer optimization problem in real time is difficult in large-scale scenarios. An algorithm is presented to find a suboptimal solution in real time by decomposing the original problem into subproblems, which are easier to solve, without using simplistic clustering algorithms that are typically used. Simulation results illustrating the performance of sensor array manager are also presented.

Multiple Unresolved Target Localization and Tracking using Colocated MIMO Radars

In this paper localization and tracking of multiple unresolved targets using a colocated multiple input multiple output (MIMO) radar is addressed. The commonly-used model for colocated MIMO radars is modified in order to guarantee the observability in received measurements. Then, a maximum likelihood estimator is derived for localizing multiple targets falling within a certain resolution cell. The Cramer-Rao lower bound (CRLB) for localization with the new model is also derived. For the tracking part, a multiple-hypothesis-based approach is used to deal with the uncertainty in target state estimation. In addition, an unscented Kalman filter (UFK) based estimator is used to tackle the nonlinearity in the measurement model. Finally, the posterior CRLB (PCRLB) is derived to evaluate the consistency of tracking results. Simulation results confirm the superiority of the proposed approach in resolving multiple targets over using the standard localization results for tracking.

Integrated Clutter Estimation and Target Tracking using Poisson Point Processes

In this paper, based on Poisson point processes, two new methods for joint nonhomogeneous clutter background estimation and multitarget tracking are presented. In many scenarios, after the signal detection process, measurement points provided by the sensor (e.g., sonar, infrared sensor, radar) are not distributed uniformly in the surveillance region as assumed by most tracking algorithms. On the other hand, in order to obtain accurate results, the target tracking filter requires information about clutters spatial intensity. Thus, nonhomogeneous clutter spatial intensity has to be estimated from the measurement set and the tracking filters output. Also, in order to take advantage of existing tracking algorithms, it is desirable for the clutter estimation method to be integrated into the tracker itself. Nonhomogeneous Poisson point processes, whose intensity function are assumed to be a mixture of Gaussian functions, are used to model clutter points here. Based on this model, a recursive maximum likelihood (ML) method and an approximated Bayesian method are proposed to estimate the nonhomogeneous clutter spatial intensity. Both clutter estimation methods are integrated into the probability hypothesis density (PHD) filter, which itself also uses the Poisson point process assumption. The mean and the covariance of each Gaussian function are estimated and used to calculate the clutter density in the update equation of the PHD filter. Simulation results show that both methods are able to improve the performance of the PHD filter in the presence of slowly time-varying nonhomogeneous clutter background.

Widely Separated MIMO versus Multistatic Radars for Target Localization and Tracking

The detection, localization, and tracking performance of multiple input-multiple output (MIMO) radars with widely separated antennas is investigated and compared with that of multistatic radar systems. A multiple-hypothesis (MH)-based algorithm is proposed for multitarget localization for the case where extended targets with multiple spatial reflections become unobservable in certain transmitter-receiver pairs. A particle filter (PF)-based algorithm is then proposed to handle dynamic multitarget tracking. Finally, simulation results are provided to demonstrate the relative capability of MIMO radars in localizing and tracking extended targets under various signal-to-noise ratio (SNR) conditions compared with multistatic radars.

PHD filter based track-before-detect for MIMO radars

In this paper a Probability Hypothesis Density (PHD) filter based track-before-detect (TBD) algorithm is proposed for Multiple-Input-Multiple-Output (MIMO) radars. The PHD filter, which propagates only the first-order statistical moment of the full target posterior, is a computationally efficient solution to multitarget tracking problems with varying number of targets. The proposed algorithm avoids any assumption on the maximum number of targets as a result of estimating the number of targets together with target states. With widely separated transmitter and receiver pairs, the algorithm utilizes the Radar Cross Section (RCS) diversity as a result of target illumination from ideally uncorrelated aspects. Furthermore, a multiple sensor TBD is proposed in order to process the received signals from different transmitter-receiver pairs in the MIMO radar system. In this model, the target observability to the sensor as a result of target RCS diversity is taken in to consideration in the likelihood calculation. In order to quantify the performance of the proposed algorithm, the Posterior Cramer-Rao Lower Bound (PCRLB) for widely separated MIMO radars is also derived. Simulation results show that the new algorithm meets the PCRLB and provides better results compared with standard Maximum Likelihood (ML) based localizations.

A Multiple-Detection Probability Hypothesis Density Filter

Most conventional target tracking algorithms assume that one target can generate at most one detection per scan. However, in many practical target tracking applications, one target may generate multiple detections in one scan, because of multipath propagation, or high sensor resolution or some other reason. If the multiple detections from the same target can be effectively utilized, the performance of the multitarget tracking system can be improved. However, the challenge is that the uncertainty in the number of targets and the measurement set-to-target association will increase the complexity of tracking algorithms. To solve this problem, the random finite set (RFS) modeling and the random finite set statistics (FISST) are used in this paper. Without any extra approximation beyond those made in the standard probability hypothesis density (PHD) filter, a general multi-detection PHD (MD-PHD) update formulation is derived. It is also established in this paper that, with certain reasonable assumptions, the proposed MD-PHD recursion can function as a generalized extended target PHD or multisensor PHD filter. Furthermore, a Gaussian Mixture (GM) implementation of the proposed MD-PHD formulation, called the MD-GM-PHD filter, is presented. The proposed MD-GM-PHD filter is demonstrated on a simulated over-the-horizon radar (OTHR) scenario.

IMM Forward Filtering and Backward Smoothing for Maneuvering Target Tracking

The interacting multiple model (IMM) estimator has been proven to be effective in tracking agile targets. Smoothing or retrodiction, which uses measurements beyond the current estimation time, provides better estimates of target states. Various methods have been proposed for multiple model (MM) smoothing in the literature. A new smoothing method is presented here which involves forward filtering followed by backward smoothing while maintaining the fundamental spirit of the IMM. The forward filtering is performed using the standard IMM recursion, while the backward smoothing is performed using a novel interacting smoothing recursion. This backward recursion mimics the IMM estimator in the backward direction, where each mode-conditioned smoother uses standard Kalman smoothing recursion. The resulting algorithm provides improved but delayed estimates of target states. Simulation studies are performed to demonstrate the improved performance with a maneuvering target scenario. Results of the new method are compared with existing methods, namely, the augmented state IMM filter and the generalized pseudo-Bayesian estimator of order 2 smoothing. Specifically, the proposed IMM smoother operates just like the IMM estimator, which approximates N2 state transitions using N filters, where N is the number of motion models. In contrast, previous approaches require N2 smoothers or an augmented state.