A Comparison of the ML - PDA and the ML - PMHT Algorithms Report

The Maximum Likelihood Probabilistic Data Association (ML-PDA) tracker and the Maximum Likelihood Probabilistic Multi-Hypothesis (ML-PMHT) tracker were applied to five synthetic multistatic active sonar scenarios featuring multiple targets, multiple sources, and multiple receivers. For each of the scenarios, Monte Carlo testing was performed to quantify the performance differences between the two algorithms. Both trackers ended up performing well. For most scenarios, MLPMHT slightly outperformed ML-PDA in terms of in-track percentage. However, in a scenario with closely-spaced targets, ML-PDA exhibited superior performance Conference Name: Proc. 14th Intn'l Conf. on Information Fusion Conference Date: July 01, 2011 The Maximum Likelihood Probabilistic Data Association (ML-PDA) tracker and the Maximum Likelihood Probabilistic Multi-Hypothesis (ML-PMHT) tracker were applied to five synthetic multistatic active sonar scenarios featuring multiple targets, multiple sources, and multiple receivers. For each of the scenarios, Monte Carlo testing was performed to quantify the performance differences between the two algorithms. Both trackers ended up performing well. For most scenarios, MLPMHT slightly outperformed ML-PDA in terms of in-track percentage. However, in a scenario with closely-spaced targets, ML-PDA exhibited superior performance A Comparison of the ML-PDA and the ML-PMHT Algorithms Steven Schoenecker NUWC Division Newport Newport, RI Email: [email protected] Peter Willett Electrical Engineering University of Connecticut Storrs, CT Email: [email protected] Yaakov Bar-Shalom Electrical Engineering University of Connecticut Storrs, CT Email: [email protected] Abstract—The Maximum Likelihood Probabilistic Data Association (ML-PDA) tracker and the Maximum Likelihood Probabilistic Multi-Hypothesis (ML-PMHT) tracker were applied to five synthetic multistatic active sonar scenarios featuring multiple targets, multiple sources, and multiple receivers. For each of the scenarios, Monte Carlo testing was performed to quantify the performance differences between the two algorithms. Both trackers ended up performing well. For most scenarios, MLPMHT slightly outperformed ML-PDA in terms of in-track percentage. However, in a scenario with closely-spaced targets, ML-PDA exhibited superior performance.