Track-before-detect Algorithm Research Articles

A group target track‐before‐detect approach using two‐stage strategy with maximum‐likelihood probabilistic data association

AbstractIn tracking scenarios involving groups with dense targets, achieving effective data association is challenging due to mutual occlusion and interference among targets. The complexity of the tracking problem is further exacerbated in low‐observable environments by the increase in false alarm rates. The track‐before‐detect (TBD) is an advanced technology for detecting and tracking low‐observable targets, effectively mitigating data association problems by integrating multi‐frame echo data. However, the existing multi‐target TBD algorithms typically assume that the targets are spatially separated and are not suitable for scenarios involving group targets. A group target maximum‐likelihood probabilistic data association (GT‐ML‐PDA) algorithm, based on the concept of TBD, is proposed to track group targets effectively in low‐observable environments. The proposed algorithm divides group target tracking into two stages: group centre trajectory estimation and individual target trajectory estimation. To enhance the performance of the proposed algorithm, two strategies are suggested: modifying the equivalent measurements and extracting independent measurement sets for individual targets. Simulation results demonstrate that the proposed algorithm is capable of effectively tracking numerous individual targets within a group, even in the presence of heavy clutter.

Convolutional Neural Network Reference for Track-Before-Detect Applications

TBD (Track-Before-Detect) algorithms allow the detection and tracking of objects of which the signal is lost in the background noise. The use of convolutional neural networks (ConvNN) allows to obtain more effective algorithms than the previous, because it is possible to take into account the background as well as the spatial and temporal characteristics of the tracked object signal. The article presents solutions for taking into account the motion with variable trajectory and speed through segmental interpolation and rectification of the trajectory, which allows the effective convolutional implementation of the TBD algorithm. The boundary of object detection was determined depending on the number of pixels of the object in relation to the number of pixels of the image stack and signal strength for the simplest neural network, so it is possible to analyse and compare more complex solutions with the proposed reference.

Optimal maneuver strategy for an autonomous underwater vehicle with bearing-only measurements

Track-before-detect (TBD) algorithms have been proven to circumvent the challenges of measurement-to-track association and have excellent robustness under harsh conditions. These benefits are aptly relevant for developing unmanned passive sonar tracking systems deployed on an autonomous underwater vehicle (AUV). This work considers the optimal maneuver problem exclusively for passive sonar TBD bearing-only localization algorithms, which is essential for an AUV to enhance observability autonomously. To solve this problem, we derive the Fisher information matrix (FIM) of the TBD algorithms, whose determinant reflects the observability. The determinant of the FIM is utilized as a cost function to design an optimal maneuver strategy (OMS). Although the cost function is nonconvex, we demonstrate that the optimal global solution maximizing the cost function can be analytically established. In addition, the designed OMS is extended to the condition with a course constraint to consider physical limitations in practical applications. Finally, simulated and real experiments are performed to verify the effectiveness of the designed OMS.

Candidate-plots-based dynamic programming algorithm for track-before-detect

Compared with the classical detect-before-track (DBT) technique, the track-before-detect (TBD) is an effective method to improve the performance of detecting weak targets. Operating directly on the plot-lists data obtained by the detector and plot-extractor, the classical dynamic programming for the TBD algorithm (DP-TBD) confronts the loss of merit function (MF), which will lead to extra degradation of detection performance. Thus, we analyze the influence of relevant factors on the loss of MF, such as target detection probability and clutter. We present results from the analysis that clutter plots near the positions of target-missed detections contribute to ease the loss of MF. The information of target-missed detections is beneficial to prevent the loss of MF. Then, we propose a candidate-plots-based DP-TBD (CP-DP-TBD) method providing candidate plots carrying the information of target-missed detections by the modified MF transfer procedure. Finally, the computational efficiency and the detection performance of CP-DP-TBD are demonstrated by comparing it with other methods in both simulated data and experimental data.

Dually Supervised Track-Before-Detect Processing of Multichannel Video SAR Data

Track-Before-Detect (TBD) algorithm has been used to track weak moving target shadow in video synthetic aperture radar (SAR), where strong maneuverability may deteriorate tracking performance. Fortunately, the Doppler characteristic of target can additionally provide velocity guidance for tracking. This article presents ground moving target indication (GMTI) results of the multi-channel video SAR raw data, and a tracking approach is proposed based on the Doppler supervision in an improved dynamic programming-based TBD (DP-TBD) framework that uses a dual-domain merit function. Both the sequential high-resolution SAR images and low-resolution Range-Doppler (RD) spectra are used, and the clutter in RD spectrum is suppressed by using an adaptive displaced phase center antenna (ADPCA) technique. Fine states can be searched in dual-domain through resampling from random expansion in state initialization and transition. The inverse shadow amplitude and Doppler energy of the same potential target are simultaneously integrated to determine whether the target exists. The proposed Dual-DP-TBD can deal with the tracking of a time-varying number of targets in successive measurements. Compared to other TBD algorithms, this approach has fewer missing alarms and false alarms thanks to precise velocity estimation constrained from diverse features both in SAR image and RD spectrum.

Track-before-detect for complex extended targets based sequential monte carlo Mb-sub-random matrices filter

Tracking for multiple extended objects with a complex extension is a challenging radar technology; especially for small back-scattering objects such as extended stealth targets (ESTs). This work provides a new approach for ESTs tracking under the non-linear dynamic system based on track-before-detect (TBD) approach. The sequential Monte Carlo multi-Bernoulli (SMC-MB) filter provides a good framework to cope with TBD approach. Recently, the SMC-MB filter with a random matrix model (RMM) has been applied for tracking extended targets by additional state variables. However, SMC-MB-RMM filter is implemented with known detection probability, which is unsuitable for ESTs-TBD scenario. Therefore, we introduce a new SMC-MB-RMM filter hybrid with TBD algorithm, which is effective method to track ESTs. In ESTs-RMM-TBD scenarios, although the extension ellipsoid is effective, it may not be accurate enough due to lacking the useful parameters, such as shape, size and orientation. Therefore, we propose a ESTs-Sub-RMM-TBD composed of sub-ellipses; each one is applied by RMM. Based on such models, a SMC-Sub-RMM-MB-TBD algorithm is applied to estimate extensions and kinematic states for each sub-objects. The simulation results show that the presented filter has a small OSPA errors and more accurate cardinality calculation than the other algorithms.

A Bernoulli Track-Before-Detect Filter for Interacting Targets in Maritime Radar

This article is devoted to the problem of multitarget tracking in the maritime environment using Bernoulli track-before-detect (TBD) filtering. Detection and tracking of weak targets in sea clutter using high-resolution airborne radar is notoriously challenging compared to classical target tracking problems. Furthermore, the measured amplitude in range-azimuth cells that are in the vicinity of closely spaced (interacting) targets represents a superposition of individual target contributions, demanding a special joint treatment. This article proposes a novel solution for multitarget tracking using Bernoulli TBD for maritime radar and covers cases when targets are both far apart and closely spaced. Comparisons with existing TBD algorithms are used to verify the efficacy of the proposed Bernoulli TBD algorithm with realistic sea clutter simulations used to create challenging scenarios with variable target strengths.

Bernoulli filter for tracking maritime targets using point measurements with amplitude

Statistical performance of the Bernoulli Track-Before-Detect (TBD) algorithm for maritime radar is superior to that of the Bernoulli tracking filter which processes, in a conventional manner, the point measurements (PM) produced by a detector. However, the Bernoulli TBD is orders of magnitude computationally more intensive. In this paper we develop a hybrid algorithm that takes advantage of the amplitude information (exploited by the TBD algorithm), but operates in the framework of conventional point-measurement Bayesian tracking. The resulting filter is formulated using the Rao-Blackwell decomposition: a particle filter estimates the target amplitude, while the target position and velocity, given the amplitude, are estimated using a Gaussian mixture filter. Numerical results show that the statistical and computational efficiency of the new hybrid filter is between the two extremes, namely, the Bernoulli TBD and the Bernoulli PM tracker.

A Cramér-Rao Lower Bound Derivation for Passive Sonar Track-Before-Detect Algorithms

Track-before-detect (TkBD) algorithms have been shown to greatly abate measurement-to-track association (MTA) challenges. These simplifications are aptly relevant for reducing operator workload in deployed sonar systems that require a human “in-the-loop.” In a prior manuscript a case study of a passive bearings-only target motion analysis TkBD algorithm was demonstrated in complex sonar scenarios relevant to advanced fielded sonar systems. In this manuscript, a Cramer-Rao Lower Bound (CRLB) is derived for the algorithm previously developed. The approximations used in developing the CRLB are validated with a real data set. The CRLB itself, as a predictor of state estimation error performance, is validated with single- and multi-contact simulated data scenarios. The prior algorithm and CRLB derived herein is applicable to passive sonar, active sonar, radar, and optical applications through a change of point spread functions and Jacobians. The CRLB derived is simple to implement, requires minimal statistical assumptions, and is applicable to similarly implemented TkBD algorithms.

Multiple maneuvering extended targets detection by 3D projection and tracklet association

In this paper, the five fundamental issues of target extension, target maneuvering, closely spaced targets, weak targets, and dense clutter are tackled simultaneously in the framework of a new tracking approach that is capable of effectively detecting multiple closely-spaced maneuvering weak and extended targets in densely cluttered environment. The proposed technique finds its foundation in the track-before-detect (TBD) paradigm, proven to be effective in tracking weak targets in presence of dense clutter. The approach consists of two stages. In the first stage, a 3-dimensional projection based TBD algorithm allows addressing the problem of target extension, weak targets, and dense clutter. Straight line tracklets are obtained at the end of the first stage. In the second stage, tracklets association is performed to obtain the target trajectories. The issue of closely spaced targets and target maneuvering is addressed in this second stage. To illustrate the effectiveness of the proposed approach, synthetic data, real data from a netted radar system, and real data from a high resolution radar are processed both with the proposed algorithm and with several existing algorithms. The obtained results reveal that the proposed method can outperform the competing approaches in various scenarios.

Multi-PRF and multi-frame track-before-detect algorithm in multiple PRF radar system

This paper considers the detection and tracking of dim targets for multiple pulse repetition frequency (PRF) pulsed Doppler radar systems through the use of multi-frame track-before-detect (TBD) methods. The main challenge is that each measurement frame is a mixture with measurements recorded, periodically, at different PRF settings, thus with different range and Doppler ambiguities. To achieve this goal, we propose variable step size multi-PRF and multi-frame (or scan) TBD (MP-MF-TBD) algorithm. Specifically, by incorporating the ambiguity factors related to both the range and velocity ambiguities, MP-MF-TBD first accomplishes the target SNR integration directly using multiple ambiguous measurement frames. After analyzing the properties of the over-threshold tracks, we further propose an efficient detection procedure, which can effectively eliminate the false tracks caused by the measurement ambiguities and accomplish the target kinematic state estimation and multi-PRF ambiguity resolution simultaneously. The proposed MP-MF-TBD does not require certain limitations on the PRF type (i.e., low, medium and high PRF) or the number of PRFs in pulsed Doppler radar systems, and appears to be robust to the target loss when target fall into a blind frequency band of certain PRFs. Extensive numerical simulations demonstrated the efficacy of the proposed MP-MF-TBD by comparing with classical techniques.

Sequential Monte Carlo Cardinalized probability hypothesized density filter based on Track-Before-Detect for fluctuating targets in heavy-tailed clutter

In this paper, a Sequential Monte Carlo Cardinalized Probability Hypothesized Density (SMC-CPHD) filter based on Track-Before-Detect (TBD) algorithm in heavy-tailed clutter is proposed. The CPHD filter, which jointly propagates the first-order statistical moment and cardinality distribution, is an effective and efficient solution to multitarget filter problem in the presence of data association uncertainty, detect uncertainty, and time-varying. The TBD algorithm is extremely sensitive to the presence of clutter spikes in the data as the formulation of the TBD filter is dependent on knowledge of underlying clutter behavior. Therefore, the clutter is modeled in terms of K distribution with sea surface background and G0 distribution with primary forest background, while the popular Swerling targets of type 1 and 3 are considered to capture the target amplitude fluctuation frame-to-frame in this paper. The utilization of K distribution and G0 distribution clutter models instead of conventional Rayleigh distribution are verified through analysis of simulated and realistic data. The relevant results prove that proposed SMC-CPHD-TBD algorithm based on K distribution model and G0 distribution model can improve the detect and tracking performance over the conventional method in heavy-tailed clutter scenario. Additionally, the numerical simulations show that proposed CPHD-TBD algorithm can output the multitarget trajectories in the better precision with considerably less computational complexity compared with DP-TBD algorithm.

Track‐before‐detect algorithm for airborne radar system

This study considers the detection of dim target via track-before-detect (TBD) in airborne radar system. Dim target detection by TBD is well-studied while aircraft ego-motion is not taken into account. Ego-motion of the aircraft (platform) causes traditional TBD method useless in airborne radar system by two reasons. On one hand, the movement of the aircraft causes measurement misaligned among different frames. On the other hand, state space and measurement space is non-one-to-one correspondence. Considering these two issues, the authors proposed a TBD algorithm for airborne radar based on plots (AR-P-TBD). First, they derived coordinate rotation and motion compensation formula, and convert each frame of measurement into a unified coordinate system to make these measurements aligned. Then, TBD processing in measurement space is implemented. It avoids the non-one-to-one correspondence between state space and measurement space. Simulation results demonstrate the effectiveness of AR-P-TBD.

Track-Before-Detect Bearings-Only Localization Performance in Complex Passive Sonar Scenarios: A Case Study

Track-before-detect (TkBD) algorithms have been shown to greatly abate measurement-to-track association (MTA) challenges. These simplifications are aptly relevant for reducing operator workload in deployed sonar systems that require a human “in the loop.” This paper presents a case study of a passive bearings-only target motion analysis TkBD algorithm using two complex data sets: one simulated and one obtained at sea, which are representative of scenario characteristics typically encountered with advanced operational sonar systems. The simulated data set exhibits crossing contacts of differing signal-to-noise ratios and short data gaps. The at-sea data, obtained from an open ocean exercise, are used to assess performance under realistic conditions that exhibit high levels of clutter, varying background noise levels, fading contacts, a high bearing-rate contact, and large data gaps. The TkBD algorithm is designed for the general complexities encountered in the examined data sets and employs a particle filter that defines its likelihood function as the accumulation of raster bin values from the output of the sonar beamformer conditioned on the particle's state trajectory. For the scenarios examined the localization performance of the algorithm is encouraging. The paper briefly reviews the chief motivation for using a TkBD approach, which is to circumvent MTA.

Target Detection and Tracking With Errors in Intensity and Mapping Dimensions

This paper addresses an engineering problem that measurement errors in intensity and mapping dimensions exist in the track-before-detect (TBD) architecture simultaneously. Drawing lessons from the probability data association (PDA) and the basic generalized likelihood ratio test TBD (GLRT-TBD), the joint log-likelihood ratio test TBD (JLLRT-TBD) method is proposed. Unlike conventional TBD algorithms, which consider intensities captured in the cells occupied by physically admissible transitions only, the novel method regards the cells centered at the transitions as effective ones. After that, detection probabilities are deduced from their positions, and then weighting likelihood ratios of their intensities results in metrics of different transitions. Finally, the maximum metric is compared with the designed threshold, and the estimation of the target state is obtained by retracing the maximization process. With this, the incoherent integration gain along consecutive scans is acquired, and the increment in computational complexity is avoided. Monte Carlo numerical results demonstrate that the detection and estimation performances of the JLLRT-TBD are superior significantly to those of the basic GLRT-TBD while the lower averaged execution time is spent.

The performance analysis of three-dimensional track-before-detect algorithm based on Fisher-Tippett-Gnedenko theorem

The dim moving target tracking from the infrared image sequence in the presence of high clutter and noise has been recently under intensive investigation. The track-before-detect (TBD) algorithm processing the image sequence over a number of frames before decisions on the target track and existence is known to be especially attractive in very low SNR environments (⩽3dB). In this paper, we shortly present a three-dimensional (3-D) TBD with dynamic programming (TBD-DP) algorithm using multiple IR image sensors. Since traditional two-dimensional TBD algorithm cannot track and detect the along the viewing direction, we use 3-D TBD with multiple sensors and also strictly analyze the detection performance (false alarm and detection probabilities) based on Fisher-Tippett-Gnedenko theorem. The 3-D TBD-DP algorithm which does not require a separate image registration step uses the pixel intensity values jointly read off from multiple image frames to compute the merit function required in the DP process. Therefore, we also establish the relationship between the pixel coordinates of image frame and the reference coordinates.

Generalized Recursive Track-Before-Detect With Proposal Partitioning for Tracking Varying Number of Multiple Targets in Low SNR

Track-before-detect (TBD) algorithms incorporate unthresholded measurements to track targets under low signal-to-noise ratio (SNR) conditions. In this paper, we generalize a single target recursive TBD tracking algorithm to track a time-varying number of targets in low SNR and cluttered environments. This algorithm is important, for example, in tracking scenes involving multiple, closely spaced targets moving along the same direction such as a convoy of low observable vehicles moving through a forest, or multiple targets moving in a crisscross fashion. The proposed multi-mode, multi-target TBD (MM-MM-TBD) algorithm is based on estimating the probabilities of all possible combinations of target existence scenarios to obtain the joint multi-target posterior probability density function in a recursive Bayesian framework. We implement the algorithm recursively using particle filtering to also incorporate nonlinear/non-Gaussian tracking models. As the number of target existence combinations dynamically increases with the number of targets, we also propose an efficient proposal density function through partitioning of the multiple target space in order to decrease the number of particles and thus improve the approximation accuracy of the particle filter. We employ a heuristic decision-directed based approach to keep the computational complexity as a linear function of the maximum number of possible targets by exploiting the information obtained from the estimated mode probabilities.

Frequency Management for Electromagnetic Continuous Wave Conductivity Meters.

Ground conductivity meters use electromagnetic fields for the mapping of geological variations, like the determination of water amount, depending on ground layers, which is important for the state analysis of embankments. The VLF band is contaminated by numerous natural and artificial electromagnetic interference signals. Prior to the determination of ground conductivity, the meter’s working frequency is not possible, due to the variable frequency of the interferences. Frequency management based on the analysis of the selected band using track-before-detect (TBD) algorithms, which allows dynamical frequency changes of the conductivity of the meter transmitting part, is proposed in the paper. Naive maximum value search, spatio-temporal TBD (ST-TBD), Viterbi TBD and a new algorithm that uses combined ST-TBD and Viterbi TBD are compared. Monte Carlo tests are provided for the numerical analysis of the properties for a single interference signal in the considered band, and a new approach based on combined ST-TBD and Viterbi algorithms shows the best performance. The considered algorithms process spectrogram data for the selected band, so DFT (Discrete Fourier Transform) could be applied for the computation of the spectrogram. Real–time properties, related to the latency, are discussed also, and it is shown that TBD algorithms are feasible for real applications.

Improved multi-target radar TBD algorithm

Considering radar detection for multi-target recognition, a track before detect (TBD) algorithm based on Hough transform is adopted for identifying and tracking multi-target radar. By increasing the dimensions of the target characteristic parameters, the target detection and track accuracy is increased. Also, by multilevel filtering processing, the diverging points of the echo signal are condensed, which improves the performance of identifying and tracking multiple targets. Simulation results show that compared with traditional TBD algorithms, the presented algorithm has better performance in the aspects of multi-target tracking, detecting and distinguishing. c 1990-2011 Beijing Institute of Aerospace Information.

Multi-target track-before-detect algorithm based on joint multi-target probability density model

Concerning the problem of Track-Before-Detect(TBD) in a multi-target environment,in this paper,a TBD algorithm based on Joint Multi-target Probability Density(JMPD) model was proposed.The JMPD was a single probabilistic entity that captured uncertainty about the number of targets present in the surveillance region as well as their individual states and a Particle Filter(PF) was used to recursively estimate the JMPD.The simulation results demonstrate that the birth and death of target can be estimated accurately as well as its trajectory by the proposed algorithm with smaller detection delays.