22 - 26 April 2019
Westin Waterfront Hotel
Boston, Massachusetts USA

Tutorial: Radar Tracking State Estimation and Association

22 April Monday Morning Session 11:00 AM – 3:00 PM

Dr. Chaw-Bing Chang
MIT Lincoln Laboratory, Lexington, MA

Dr. Keh-Ping Dunn
MIT Lincoln Laboratory, Lexington, MA

State estimation is to determine target’s motion dynamics, association is to identify measurements of the same target in a multiple target environment for estimation.  Put them together, it is known as target tracking. The purpose of this tutorial is to introduce the audience with techniques of state estimation and association for tracking.  Theory and application of different estimation and association algorithms will be discussed. Applicability of recursive and batch processing algorithms will be compared.  Specific attentions will be given to practical filter designs as well as their estimation performance prediction and evaluation. A variety of adaptive techniques for tracking maneuvering targets including multiple model estimation algorithms will be presented. Methods for mitigating radar biases will be discussed. Integrating multiple radar collections in a network can improve tracking accuracy, enhance surveillance coverage, and achieve a variety of additional advantages.  Architectures and algorithms for integrating multiple radars will be presented with the pros and cons of different approaches compared. A real world challenge in multiple target tracking is the presence of dense target environment, hence the problem of association. A variety of association techniques will be presented including nearest neighbor, global nearest neighbor, probabilistic data association, multidimensional assignment, multiple hypothesis tracking, etc. Several approaches for jointly solving estimation and association problems in a multiple radar system with measurement biases will be compared using examples. Numerical examples representing tracking of air and ballistic targets will be included throughout the lecture to help the audience to understand practical aspects of applying state estimation and association to target tracking.  Open issues and further areas of research together with more recent approaches such as “particle filters” will be briefly discussed at the end.