Group Target Tracking 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.

UAV Swarm Centroid Tracking for Edge Computing Applications Using GRU-Assisted Multi-Model Filtering

When an unmanned aerial vehicles (UAV) swarm is used for edge computing, and high-speed data transmission is required, accurate tracking of the UAV swarm’s centroid is of great significance for the acquisition and synchronization of signal demodulation. Accurate centroid tracking can also be applied to accurate communication beamforming and angle tracking, bringing about a reception gain. Group target tracking (GTT) offers a suitable framework for tracking the centroids of UAV swarms. GTT typically involves accurate modeling of target maneuvering behavior and effective state filtering. However, conventional coordinate-uncoupled maneuver models and multi-model filtering methods encounter difficulties in accurately tracking highly maneuverable UAVs. To address this, an innovative approach known as 3DCDM-based GRU-MM is introduced for tracking the maneuvering centroid of a UAV swarm. This method employs a multi-model filtering technique assisted by a gated recurrent unit (GRU) network based on a suitable 3D coordinate-coupled dynamic model. The proposed dynamic model represents the centroid’s tangential load, normal load, and roll angle as random processes, from which a nine-dimensional unscented Kalman filter is derived. A GRU is utilized to update the model weights of the multi-model filtering. Additionally, a smoothing-differencing module is presented to extract the maneuvering features from position observations affected by measurement noise. The resulting GRU-MM method achieved a classification accuracy of 99.73%, surpassing that of the traditional IMM algorithm based on the same model. Furthermore, our proposed 3DCDM-based GRU-MM method outperformed the Singer-KF and 3DCDM-based IMM-EKF in terms of the RMSE for position estimation, which provides a basis for further edge computing.

Belief-Propagation-Based Resolvable Group Target Tracking With Unknown Detection Probability

Belief-Propagation-Based Resolvable Group Target Tracking With Unknown Detection Probability

MMF-GSTIW-PMBM Adaptive Filter for Multiple Group Target Tracking With Heavy-Tailed Noise

MMF-GSTIW-PMBM Adaptive Filter for Multiple Group Target Tracking With Heavy-Tailed Noise

Group Target Tracking for Highly Maneuverable Unmanned Aerial Vehicles Swarms: A Perspective.

Group target tracking (GTT) is a promising approach for countering unmanned aerial vehicles (UAVs). However, the complex distribution and high mobility of UAV swarms may limit GTTs performance. To enhance GTT performance for UAV swarms, this paper proposes potential solutions. An automatic measurement partitioning method based on ordering points to identify the clustering structure (OPTICS) is suggested to handle non-uniform measurements with arbitrary contour distribution. Maneuver modeling of UAV swarms using deep learning methods is proposed to improve centroid tracking precision. Furthermore, the group's three-dimensional (3D) shape can be estimated more accurately by applying key point extraction and preset geometric models. Finally, optimized criteria are proposed to improve the spawning or combination of tracking groups. In the future, the proposed solutions will undergo rigorous derivations and be evaluated under harsh simulation conditions to assess their effectiveness.

Group target tracking via jointly optimizing group partition and association

This paper considers the problem of indistinguishable group target tracking, where the individual targets within a group are closely spaced and resolvable, move in a coordinated manner, and the group splitting and merging may occur. In such a case, the performance of the commonly used multi-target tracking methods may be degraded. We focus on three aspects of the problem: deal with the group association, group splitting and merging, and derive the state estimates of the individual targets. First, in order to deal with the group partition and the group association globally, we propose a novel group target tracking method within the multiple hypothesis tracking (MHT) framework by extending the target partition and the measurement partition as parts of global hypotheses. An efficient algorithm is given to find the M-best group partition and association simultaneously. Second, we use the undirected graph to represent the interaction between groups and then obtain an identification method of the group splitting and merging. Third, in response to the state estimate of the group target, a group-individual Bayesian filter is designed to estimate the states of the virtual leader and the individual targets. Finally, numerical experiments are presented to illustrate the effectiveness of the proposed method.

Multi-Target State and Extent Estimation for High Resolution Automotive Sensor Detections

This paper discusses the perception and tracking of individual as well as group targets as applied to multi-lane public traffic. Target tracking problem is formulated as a two hierarchical layer problem-on the first layer, a multi-target tracking problem based on multiple detections is distinguished in the measurement space, and on the second (top) layer, group target tracking with birth and death as well as merging and splitting of group target tracks as they evolve in a dynamic scene is represented. This configuration enhances the multi-target tracking performance in situations including but not limited to target initialization(birth), target occlusion, missed detections, unresolved measurement, target maneuver, etc. In addition, group tracking exposes complex individual target interactions to help in situation assessment which is challenging to capture otherwise.

Resolvable Group Target Tracking via Multi-Bernoulli Filter and Its Application to Sensor Control Scenario

In this paper, we address the problem of resolvable group target tracking (RGTT) and the application of RGTT in the sensor control scenarios. The aim of our approach is to further improve the tracking performance of multi-Bernoulli (MB) filter by exploiting the group structure information. Firstly, a single-target state transition function (SSTF) is derived from the direct integral solution of the single-target state stochastic differential equation (SDE), which is used to revise the prediction process of the standard MB filter. Then, combing the Gaussian mixture (GM) implementation, the prediction process of the proposed SSTF-based MB (SSTF-MB) filter is detailed. Next, considering the benefits and importance of group structure in the sensor control scenarios, the proposed SSTF-MB filter is applied into a mobile sensor platform whose control actions (e.g., motion direction) are uncertain. Specifically, the proposed SSTF-MB filter is merged into a sensor control strategy where the Cauchy-Schwarz (CS) divergence is selected as the objective function to compute the discrepancy between two first-order moment approximations of MBs. Finally, the efficiency and performance of the proposed SSTF-MB filter and its application to sensor control scenario is well demonstrated in the simulation experiments.

Multi object tracking using gradient-based learning model in video-surveillance

On accomplishing an efficacious object tracking, the activity of an object concerned becomes notified in a forthright manner. An accurate form of object tracking task necessitates a robust object tracking procedures irrespective of hardware assistance. Such approaches inferred a vast computational complexity to track an object with high accuracy in a stipulated amount of processing time. On the other hand, the tracking gets affected owing to the existence of varied quality diminishing factors such as occlusion, illumination changes, shadows etc., In order to rectify all these inadequacies in tracking an object, a novel background normalization procedure articulated on the basis of a textural pattern is proposed in this paper. After preprocessing an acquired image, employment of an Environmental Succession Prediction algorithm for discriminating disparate background environment by background clustering approach have been accomplished. Afterward, abstract textural characterizations through utilization of a Probability based Gradient Pattern (PGP) approach for recognizing the similarity between patterns obtained so far. Comparison between standardized frame obtained in prior and those processed patterns detects the motion exposed by an object and the object concerned gets identified within a blob. Hence, the system is resistant towards illumination variations. These illumination variation was interpreted in object tracking residing within a dynamic background. Devised approach certainly outperforms other object tracking methodologies like Group Target Tracking (GTT), ViPER-GT, grabcut, snakes in terms of accuracy and average time. Proposed PGP-based pattern texture analysis is compared with Gamifying Video Object (GVO) approach and hence, it evidently outperforms in terms of precision, recall and F1 measure.

Group Target Tracking Based on MS-MeMBer Filters

This paper presents a new group target tracking method based on the standard multi-sensor multi-target multi-Bernoulli (MS-MeMBer) filter. In the prediction step, the group structure is used to constrain the movement of the constituent members within the respective groups. Specifically, the group of members is considered as an undirected random graph. Combined with the virtual leader-follower model, the motion equation of the members within groups is formulated. In the update step, the partitioning problem of multiple sensors is transformed into a multi-dimensional assignment (MDA) problem. Compared with the original two-step greedy partitioning mechanism, the MDA algorithm achieves better measurement partitions in group target tracking scenarios. To evaluate the performance of the proposed method, a simulation scenario including group splitting and merging is established. Results show that, compared with the standard MS-MeMBer filter, our method can effectively estimate the cardinality of members and groups at the cost of increasing computational load. The filtering accuracy of the proposed method outperforms that of the MS-MeMBer filter.

A Detection and Tracking Algorithm for Resolvable Group with Structural and Formation Changes Using the Gibbs-GLMB Filter.

In the field of resolvable group target tracking, further study on the structure and formation of group targets is helpful to reduce the tracking error of group bluetargets. In this paper, we propose an algorithm to detect whether the structure or formation state of group targets changes. In this paper, a Gibbs Generalized Labeled Multi-Bernoulli (GLMB) filter is used to obtain the estimation of discernible group target bluestates. After obtaining the state estimation of the group target, we extract relevant information based on the estimation data to judge whether the structure or formation state has changed. Finally, several experiments are carried out to verify the algorithm.

Seamless group target tracking using random finite sets

In many surveillance applications, target tracking algorithms have to deal with dense group targets in contrast to independently moving and well-separated targets as assumed in most scenarios. An effective strategy to handle such a group of targets is to first track the overall group and then attempt to extract the states of individual targets. Based on the Random Finite Set theory, the τ-Cardinalized Probability Hypothesis Density (τ-CPHD) filter is proposed in this paper as an effective method for group target tracking. This filter accurately extracts target states and represents the distribution of the legacy PHD. Association and track extraction are done in the post-processing step, without affecting the filtering process or overloading the computing resources. Furthermore, the group motion is modeled in conjunction with individual target motion models, with model proportion being updated by the proposed filter. Initialization process is carried out adaptively using the group motion trend through group state fitting. All these characteristics make the new filter flexible and seamlessly applicable to real-world tracking problems with group targets. Simulations demonstrate the superior performance of the proposed filter with individual target and group motion models being considered in combination or separately.

Recursive update filtering-based variational Bayesian approach for extended target or group target tracking with nonlinear measurements

For extended target and group target tracking, most existing random matrix approaches assume that the measurements are linear in the kinematic state and in the noise with its covariance being a random matrix to represent the extension of an extended target or target group. However, in practice, the measurements (e.g., range and bearing) are nonlinear in the kinematic state and extension noise. Only the matched linearization (ML) approach has been proposed to linearize the nonlinear measurement function based on minimum mean square error criterion. However, if the state prediction error is not small, it will incur large linearization error and then the variational Bayesian measurement update approach produces inaccurate estimation results. To this point, using the ML technology, we present a recursive update filtering-based variational Bayesian update (RUF-VBU) approach for better estimation performance. In addition, the square-root implementation of RUF-VBU is proposed for the recursive interruption problem caused by the nonpositive definite covariance matrix. The simulation results demonstrate the effectiveness of the proposed approach.

Resolvable Group State Estimation with Maneuver Based on Labeled RFS and Graph Theory.

In this paper, multiple resolvable group target tracking was considered in the frame of random finite sets. In particular, a group target model was introduced by combining graph theory with the labeled random finite sets (RFS). This accounted for dependence between group members. Simulations were presented to verify the proposed algorithm.

Tracking Maneuvering Group Target with Extension Predicted and Best Model Augmentation Method Adapted

The random matrix (RM) method is widely applied for group target tracking. The assumption that the group extension keeps invariant in conventional RM method is not yet valid, as the orientation of the group varies rapidly while it is maneuvering; thus, a new approach with group extension predicted is derived here. To match the group maneuvering, a best model augmentation (BMA) method is introduced. The existing BMA method uses a fixed basic model set, which may lead to a poor performance when it could not ensure basic coverage of true motion modes. Here, a maneuvering group target tracking algorithm is proposed, where the group extension prediction and the BMA adaption are exploited. The performance of the proposed algorithm will be illustrated by simulation.

Tracking of extended object or target group using random matrix: new model and approach

The approach of using a random matrix for extended object and group target tracking (EOT and GTT) is appealing when the scattering centers of the object or the group targets themselves are (partially) unresolvable. Designing and effectively applying this approach relies on effective modeling of the extended object and the target group. To describe complex dynamical variation and practical observation distortion of the extension in size, shape, and orientation, two random matrix-based models are proposed. True measurement noise can also be incorporated into our proposed measurement model easily. Facilitated by special properties of the models, an approximate Bayesian approach is proposed to estimate the kinematic state and the extension jointly. For maneuvering EOT and GTT, a multiple-model approach is derived by moment matching. To evaluate what is proposed, a scenario for maneuvering EOT is simulated. The results illustrate the effectiveness of the proposed models and approach.

Sensor selection for tracking multiple groups of targets

Group target tracking is a challenge for sensor networks. It occurs where large numbers of closely spaced targets move together in different groups. In these applications, the sensor selection scheme plays a vital role in extending network lifetime while providing high tracking accuracy. Existing schemes cause an extreme imbalance between energy usages and tracking accuracy. They are capable of tracking only individual groups and without using prior knowledge about the groups. These problems make them impractical for group target tracking. With the aim of balancing the trade-off between lifetime and accuracy, we present a novel Multi-Sensor Group Tracking (MSGT) scheme. MSGT comprises the following steps to accomplish concurrent tracking of multiple groups: (1) Clustering to capture changes in the behavioural properties of groups, such as formation, merging, and splitting; (2) Sensor selection to activate the contributory sensors for the estimated group regions; and (3) Group tracking using the activated sensors. We develop a probabilistic decision-making strategy that triggers the clustering step adaptively with any detected change in group behavioural patterns. The sensor selection step coordinates periodic selection of leader and tracking sensor nodes in a distributed manner. We introduce cost metrics that include sensor′s energy parameters in the selection of active sensors that fully cover the group regions. The tracking step is a Bayesian modelling of the target groups which uses particle filtering algorithm to estimate the group locations. Simulation results show that our scheme achieves substantial improvements over existing approaches in terms of network lifetime and tracking accuracy.

Tracking of Maneuvering Non-Ellipsoidal Extended Object or Target Group Using Random Matrix

For non-ellipsoidal extended object and group target tracking (NEOT and NGTT), using a random matrix to simplify the extension as an ellipsoid, although efficient, may not be accurate enough because of the loss of useful information in shape and orientation. In view of this, we model a non-ellipsoidal extended object or target group as a combination of multiple ellipsoidal sub-objects, each represented by a random matrix. Different sub-objects are distinguished by different dynamic models. Based on such models, a Bayesian approach is proposed to estimate the kinematic states and the extensions of the sub-objects. For maneuvering NEOT and NGTT, a multiple-model approach is proposed. The overall maneuver model is combined with the sub-object models to describe the common and the individual motions of the sub-objects. Three simplifications of the approach are presented. Two scenarios for maneuvering and nonmaneuvering NEOT and NGTT are simulated. The results demonstrate the effectiveness of the proposed modeling and estimation approach.

Group Target Tracking in WSN Based on Convex Hulls Merging

When a mass of individual targets move closely, it is unpractical or unnecessary to localize and track every specific target in wireless sensor networks (WSN). However, they can be tracked as a whole by view of group target. In order to decrease the amount of energy spent on active sensing and communications, a flexible boundary detecting model for group target tracking in WSN is proposed, in which, the number of sensors involved in target tracking is adjustable. Unlike traditional one or multiple individual targets, the group target usually occupies a large area. To obtain global estimated position of group target, a divide-merge algorithm using convex hull is designed. In this algorithm, group target’s boundary is divided into several small pieces, and each one is enclosed by a convex hull which is constructed by a cluster of boundary sensors. Then, the information of these small convex hulls is sent back to a sink. Finally, big convex hull merged from these small ones is considered as the group target’s contour. According to our metric of precision evaluation, the simulation experiments confirm the efficiency and accuracy of this algorithm.

Track labeling and PHD filter for multitarget tracking

Multiple target tracking requires data association that operates in conjunction with filtering. When multiple targets are closely spaced, the conventional approaches (as, e.g., MHT/assignment) may not give satisfactory results. This is mainly because of the difficulty in deciding what the number of targets is. Recently, the probability hypothesis density (PHD) filter has been proposed and particle filtering techniques have been developed to implement the PHD filter. In the particle PHD filter, the track labeling problem is not considered, i.e., the PHD is obtained only for a frame at a time, and it is very difficult to perform the multipeak extraction, particularly in high clutter environments. A track labeling method combined with the PHD approach, as well as considering the finite resolution, is proposed here for multitarget tracking, i.e., we keep a separate tracker for each target, use the PHD in the resolution cell to get the estimated number and locations of the targets at each time step, and then perform the track labeling ("peak-to-track" association), whose results can provide information for PHD peak extraction at the next time step. Besides, by keeping a separate tracker for each target, our approach provides more information than the standard particle PHD filter. For example, in group target tracking, if we are interested in the motion of a specific target, we can track this target, which is not possible for the standard particle PHD filter, since the standard particle PHD filter does not keep track labels. Using our approach, multitarget tracking can be performed with automatic track initiation, maintenance, spawning, merging, and termination