Multi-Tier Cluster Based Tracker Approach for Battlefield Acoustic Systems

摘要

The Acoustic Signal Processing Branch of the U.S. Army Research Laboratory (ARL) has ongoing research into Battlefield target tracking. The classic approach is to combine information from multiple line-of-bearing (LOB) sensors that are spatially diverse. The triangulations from candidate intersection points that a tracking algorithm can de-ghost and develop track histories on. For these traditional trackers the vehicles of interest must be resolved by multiple sensors simultaneously to form a valid intersection, which is attainable in sparse vehicle scenarios. In the scenario of an active Battlefield, the sensor field performance can be radically different, the sensors will be captured by their nearest targets and lose the ability to produce valid LOB intersections across the field due to each sensor hearing a different target. This capture effect will force traditional trackers to fail. This paper will develop the concept of a multi-tier tracker, which works at micro level and a macro level. At the micro level the sensor field will focus on producing an accurate estimate of vehicle count and rough estimate of cluster geometry. The cluster estimate produced does not require the simultaneous vehicle resolution by the sensors. This cluster estimate can then be tracked at the level via a traditional tracker. The cluster estimation and tiered tracking will provide robust theater level tracker operation with realistic sensor performance.