An Improved S-D Assignment Algorithm for Underwater Multi-Sensor Multi-Target Tracking

This article addresses the issue in underwater environments where the bearing separation between two co-frequency signal targets is smaller than the main lobe beamwidth of the array, resulting in target indistinguishability and missing bearing measurements. These factors render traditional S-D assignment results unreasonable and degrade localization accuracy. To enhance multi-target localization and tracking performance, an improved S-D assignment method is proposed. Firstly, based on the normalized intensity of beamforming, it is determined whether the bearing measurements of targets under each sensor are likely to be merged. Subsequently, suspected merged bearings are split and recombined, using the total association cost of the S-D assignment as the decision criterion, and ultimately selecting the combination with the minimal total cost. Simulation results demonstrate that the proposed algorithm effectively suppresses the merging of bearing measurements from each sensor, significantly improving the tracking performance of underwater multi-sensor multi-target systems.