Credits: 4 hp

Note: We reserve the the right to fine-tune the exercises up until the lecture before they are to be handed in. However, no major changes are to be expected after the start of the course.

Exercise 1: Basics

The purpose of this exercise is to get acquainted with the basic building blocks used in target tracking, as well as build basic infrastructure to facilitate the subsequent exercises. At the end of the exercise, you will have the infrastructure to simulate measurements from simple tracking scenarios, visualize and benchmark the result.

Exercise 2: Single-Hypothesis Tracker

The purpose of this exercise is to introduce multi-target tracking (MTT), and to get acquainted with common algorithms for single hypothesis trackers (SHT). To do this, first, the simple simulation environment will be extended to generate measurements from a multiple target scenario. After that, a framework for multiple target tracking will be developed. In which in this exercise it will make up the basis for a global nearest neighbor (GNN) tracker and then a joint probabilistic detection association (JPDA) tracker.

Exercise 3: Multi-Hypothesis Tracker

The purpose of this exercise is to implement a basic multi-hypothesis tracker (MHT). The intention is to gain fundamental understanding of what is takes to implement an MHT, not to write a MHT ready for use in production. The developed tracker will be compared to the single-hypothesis trackers (SHT) studied in the previous exercise. To make it easy to compare the results, the same synthetic and mysterious data is used to evaluate the algorithm.