Tabu Search Tracker with Adaptive Neuro-Fuzzy Inference System for Multiple Target Tracking
Ilke Turkmen
and
Kerim Guney
In this paper, a tabusearc h tracker with adaptive neurofuzzy inference system (TST-ANFIS) is presented for multiple target tracking (MTT). First, the data association problem, formulated as an N-dimensional assignment problem, is solved using the tabu search algorithm (TSA), and then the inaccuracies in the estimation are corrected by the adaptive neuro-fuzzy inference system (ANFIS). The performances of the TST-ANFIS, the joint probabilistic data association filter (JPDAF), the tabusearc h tracker (TST), Lagrangian relaxation algorithm (LRA), and cheap joint probabilistic data association with adaptive neuro-fuzzy inference system state filter (CJPDA-ANFISSF) are compared with each other for six different tracking scenarios. It was shown that the tracks estimated by using proposed TST-ANFIS agree better with the true tracks than the tracks predicted by the JPDAF, the TST, the LRA, and the CJPDAANFISSF.