Multiple Hypothesis Tracker (MHT) for Multitarget Tracking

In this work, a real-time feasible Multiple Hypothesis Tracker will be developed for multiple targets tracking. A computationally feasible approximated version of MHT will be compared against the robabilistic Data Association (PDA) based Tracker as well as multi-frame assignment based multitarget tracker. Intensive testing will also be performed in order to determine the performance, complexity, and resource usage of proposed and existing and proposed algorithms. These analyses will help in determining the best possible tracking scheme to be used in different situations and different architectures. The partner organization, ICx Radar Systems, will make use of this work to upgrade their existing tracker to Multiple Hypothesis Tracker. Further, the Doppler information will be incorporated into tracking framework to improve the tracking performance with noisy measurement in high clutter environments. The performance comparisons will be performed with the help of ICxfs automatic Tracker test software (Tester), and also by the development and analysis of simulated target scenarios.

Faculty Supervisor:

Dr. T. Kirubarajan

Student:

Sutharsan Sivagnanam

Partner:

ICx Radar Systems

Discipline:

Engineering

Sector:

Information and communications technologies

University:

Program: