Five different methods suitable for tracking single targets in clutter are compared: the Nearest Neighbor algorithm, the Probabilistic Multi-Hypothesis Tracking filter, the Probabilistic Data Association Filter, the Mixture Reduction algorithm, and the Mean-Field Event-Averaged Maximum Likelihood Estimator. Across a range of clutter densities, comparison results were generated for a common, fixed set of Monte Carlo target, target measurement, and clutter measurement realizations. The relative performances, as measured by track lifetime, RMS tracking error, and computational complexity are compared.
展开▼
