An Adaptive Density-Based Fuzzy Clustering Track Association for Distributed Tracking System

The problem of duplicate track determination called "Track-to-Track association'' occurs when a target is reported by different sensors, and it is regarded as one of the most important challenges in distributed multi-sensor tracking systems. The present study aimed to propose a density-based fuzzy clustering method for solving the track-to-track association problem in distributed multi-sensor tracking systems. Unlike the previously published solutions, the proposed method does not need any information about the number of targets, due to the use of the density-based clustering approach. Proposed method has low computational overhead and can be used in real-time tracking systems. In addition, the proposed method uses the maximum entropy approach to determine the membership degree of single target related tracks and combines them. This paper presents three scenarios including sensors with complete and incomplete overlapping by considering the bias and a different number of sensors and targets for evaluating the proposed method based on the Monte Carlo simulation. The results indicate the improvement of the efficiency in comparison with the FTF approach. The efficiency of proposed method's results is close to the results of Bayesian minimum mean square error criterion that gives best possible results.