Association Methods
The Ufil framework provides several association methods for optimal measurement-to-track assignments. Next to well known algorithms such as the Hungarian and Jonker–Volgenant algorithms, Ufil also includes more experimental approaches such as a linear programming (LP) solver-based method and the greedy associator.
The inclusion in this frame does not necessarily indicate that these methods are recommended for general use, but they are available for users who want to experiment with different approaches or have specific requirements that may be better served by these methods. Please refer to the runtime analysis in runtime analysis for performance comparisons between these methods.
All association methods in Ufil are designed to work with the cost matrices generated by the hypothesizer functions presetned in hypothesizer functions. These cost matrices represent the “cost” of associating a particular measurement with a track, and the association methods use these costs to find optimal or near-optimal assignments.