Optimizing multistatic sonobuoy placement

Abstract

Sonobuoy patterns for monostatic sensors were developed during the Cold War for deep, uniform underwater environments, where a simple median detection range defined a fixed inter-buoy spacing (usually along staggered lines). Oceanographic and acoustic conditions in littoral environments are so complex and dynamic that spatial and temporal variability of low-frequency signal and noise fields destroys the basic homogeneous assumption associated with standard tactical search concepts. Genetic Algorithms (GAs) have been applied to this problem to produce near-optimal, non-standard search tracks for monostatic mobile sensors that maximize probability of detection in such inhomogeneous environments. The present work describes a new capability, SCOUT (Sensor Coordination for Optimal Utilization and Tactics), to simulate multistatic distributed-sensor geometries and to optimize the locations of multistatic active sonobuoys in a complex, littoral environment. This presentation reviews the GA approach, discusses the new chromosome structure, and introduces a new target-centric geometry. The results show that (a) standard patterns are not optimal even for a homogeneous environment, (b) small distributed sensor clusters are preferred, and (c) standard patterns are grossly ineffective in inhomogeneous environments where 20% improvements in detection are achieved with SCOUT. [Work supported by NAVAIR.]