Multi-Object-Tracking

Repository for the course "Multi-Object Tracking for Automotive Systems" at EDX Chalmers University of Technology

Note: I have completed all assignments in the course and obfuscated key function by using encrypted p files, since it is not allowed to release solution publicly on github, according to EDX policy. Any interest in implementation details will be discussed privately.

Home-Assignment 01 (SA2) - Single-Object Tracking in Clutter

Implementation of the following algorithms:

• Nearest Neighbors Filter (NN)
• Probabilistic Data Association Filter (PDA)
• Gaussian Sum Filter (GSF)

The main class is located at SA2/singleobjectracker.m

Simulations can be done using SA2/simulation1.mlx

Home-Assignment 02 (SA3) - Tracking n Objects in Clutter

Implementation of the following algorithms:

• Global Nearest Neighbors Filter (GNN)
• Joint Probabilistic Data Association Filter (JPDA)
• Track-oriented Multiple Hypothesis Tracker (TO-MHT)

The main class is located at SA3/n_objectracker.m

Simulations can be done using SA3/simulation2.mlx

Home-Assignment 03 (SA4) - Random Finite Sets

Implementation of the following algorithms:

• Probability Hypothesis Density Filter (PHD)

The main class is located at SA4/PHDfilter.m

Simulations can be done using SA4/simulation3.mlx

Filter output

Cardinality Estimation	GOSPA

Home-Assignment 04 (SA5) - MOT Using Conjugate Priors

Implementation of the following algorithms:

• Poisson Multi-Bernoulli Mixture filter (PMBM)

The main class is located at SA5/PMBMfilter.m

Simulations can be done using SA5/simulation4.mlx

Filter output

Cardinality Estimation	GOSPA