Multi-Target Tracking via Mixed Integer Optimization
Abstract
The field of multi-target tracking faces two primary challenges: (i) data association and (ii) trajectory estimation. MTT problems are well researched with many algorithms solving these two problems separately, however few algorithms attempt to solve these simultaneously and even fewer utilize optimization. In this paper we introduce a new mixed integer optimization (MIO) model which solves the data association and trajectory estimation problems simultaneously by minimizing an easily interpretable global objective function. Furthermore, we propose a greedy heuristic which quickly finds good solutions. We extend both the heuristic and the MIO model to scenarios with missed detections and false alarms.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 13, 2016
- Accession Number
- AD1033655
Entities
People
- Dimitris Bertsimas
- Shimrit Shtern
- Zachary Saunders
Organizations
- MIT Lincoln Laboratory