Track-Oriented Multiple Hypothesis Tracking Based on Tabu Search and Gibbs Sampling

Abstract

In order to circumvent the curse of dimensionality in multiple hypothesis tracking (MHT) data association, this paper proposes two efficient implementation algorithms using Tabu search and Gibbs sampling. As the first step, we formulate the problem of generating the best global hypothesis in MHT as the problem of finding a maximum weighted independent set of a weighted undirected graph. Then, the metaheuristic Tabu search with two basic movements is designed to find the global optimal solution of the problem formulated. To improve the computational efficiency, this paper also develops a sampling based algorithm based on Gibbs sampling. The problem formulated for the Tabu search based algorithm is reformulated as a max product problem to enable implementation of Gibbs sampling. The detailed algorithm is then designed and the convergence is also theoretically analysed. The performance of the two algorithms proposed are verified through numerical simulations and compared with that of a mainstream MDA implementation algorithm. The simulation results confirm that the proposed algorithms significantly improve the computational efficiency while maintaining or even enhancing the tracking performance.

Trajectory Optimization for Multitarget Tracking Using Joint Probabilistic Data Association Filter

“…For the purpose of analyzing the characteristics of the proposed algorithm, the localization and cardinality errors are also used as performance metrics in the simulations. These two performance metrics are defined as As expected, increasing the weighting factor ω in cost function (14) will reduce the localisation error d c p;loc X; Y because more penalty is enforced on the performance of target state estimation. Similarly, the cardinality error d c p;card X; Y can be reduced by decreasing the weight factor ω in cost function (14).…”

“…It should be pointed out that the OSPA distance reduces to the summation of localization and cardinality estimation errors by choosing p 1. With this in mind, it is expected that the proposed cost function (14) with ω 0.5 would provide the best solution in terms of OSPA performance.…”

“…In MTT, except for the fact that the number of targets varies randomly in time, the received measurements are also subject to a certain degree of uncertainties: unknown source origin, miss detections, and false alarms. There are several elegant solutions available in the literature to address the measurement uncertainty problem in MTT: nearest neighbor (NN) filter [8], probabilistic data association (PDA) filter [9], joint probabilistic data association (JPDA) filter [10,11], and multiple hypothesis tracking (MHT) [12][13][14]. This Note adapts the JPDA filter as the baseline multitarget tracker because of its balance between estimation accuracy and computational cost.…”

A New Multiple Hypothesis Tracker Using Validation Gate with Motion Direction Constraint

“…A cluster is formed by a collection of incompatible tracks, each cluster can generate its global hypotheses and independently calculate its global probability of trajectory. In TOMHT, the commonly used algorithms to search for the global optimal hypothesis include the Lagrangian relaxation implementation algorithm, the multiple dimensional assignment (MDA) algorithm, and the greedy randomized adaptive search procedure (GRASP) method [ 22 , 23 ]. Assume that there are J global hypotheses.…”

Multi-Sensor Multi-Target Tracking Using Domain Knowledge and Clustering

“…This technique discerns target-generated measurements from clutters and finds the mappings between targets and measurements. Multiple hypothesis tracking (MHT) is one of the most well-known association algorithms and a well-established paradigm [5]- [7]. Fundamentally, MHT maintains all incompatible (e.g., unresolved) track hypotheses in track trees during a sliding window.…”