This repository contains MATLAB assignments completed for the course "Multi-Object Tracking for Automotive Systems" offered by EDX Chalmers University of Technology. Each assignment focuses on implementing various tracking algorithms for automotive applications.
- Nearest Neighbors Filter (NN)
- Probabilistic Data Association Filter (PDA)
- Gaussian Sum Filter (GSF)
- Global Nearest Neighbors Filter (GNN)
- Joint Probabilistic Data Association Filter (JPDA)
- Track-oriented Multiple Hypothesis Tracker (TO-MHT)
- Probability Hypothesis Density Filter (PHD)
- Gaussian Mixture Probability Hypothesis Density Filter (GM-PHD)
- Multi-Bernoulli Mixture filter (MBM)
- Poisson Multi-Bernoulli Mixture filter (PMBM)
To run the code in this repository, ensure the following:
- MATLAB (R2020b or later recommended)
- Statistics and Machine Learning Toolbox (if applicable)
This section provides a summary of the results obtained from the implemented algorithms, captured in the following files:
-
Cardinality
- This file contains the cardinality estimates for each time step, comparing the predicted and ground truth values.
- Screenshot:
-
Metrics
- Includes performance metrics such as OSPA (Optimal Subpattern Assignment) and GOSPA (Generalized OSPA) scores, highlighting the accuracy of the tracking algorithms.
- Screenshot:
-
Nonlinear Prediction Ground Truth
- This file visualizes the ground truth trajectories versus the predicted trajectories for nonlinear models.
- Screenshot:
