Track-to-track Fusion for Multi-target Tracking Using Asynchronous and Delayed Data
ALEXANDER BERG ANDREAS KÄLL, Master’s thesis, Sweden 2017
1.1 Thesis Objective
The aim of the thesis can be summarized with the following questions:
- What existing real-time fusion strategies can be utilised when fusing data from multiple vehicles in an urban environment?
- How can the effect of a varying time delay in the communication channel be minimized in terms of accuracy?
- How can tracks of the same object from multiple vehicles be associated with each other with high certainty?
- How can the algorithm detect if an object has left the field of view of all the vehicles and decide when to terminate the tracking of that object?
- Which algorithms can be utilized when performing fusion of already filtered data?
- Which motion models are most suitable for modeling a vehicle for tracking purposes?