Design and performance of irregular sonobuoy patterns in complicated environments

Abstract

Patterns for optimal monostatic sonobuoy fields were developed during the Cold War for use in deep, uniform undersea environments, where a simple median detection range can be used to define a useful fixed spacing between sonobuoys. However, oceanographie and acoustic conditions in the littoral environments where current operations are often conducted are so complex and dynamic that spatial and temporal variability destroys the homogeneous assumption associated with traditional tactical search concepts. Several research efforts have been undertaken to design better placements of passive and monostatic-active sonobuoys, but most of these are evaluation algorithms, as opposed to true planning algorithms. A different algorithmic approach, which begins with a random set of sensor locations and then uses genetic algorithms to find a near-optimal solution, was successfully developed and initially applied to monostatic mobile sensors. The genetic algorithm solutions were non-standard search paths that adapted to complex oceanography, to variable bottom properties, and to assumed target tactics [D.P. Kierstead and D.R. DelBalzo, Military Operations Research Journal (March/April 2003)]. A new capability was then developed to optimize the locations (latitude, longitude, and depth) and ping times of multistatic active sonobuoys in a complex, littoral environment. These algorithms are called SCOUT (Sensor Coordination for Optimal Utilization and Tactics). SCOUT contains two major modifications to the mobile-sensor genetic algorithm approach in order to account for bistatic and multistatic sonobuoy fields, where every receiver is capable of observing data from every source. The first is in structure, where a new chromosome was introduced to describe the tactical plan. It has one gene for each sonobuoy, consisting of a location, an ordered deployment sequence, and a set of ping times. Positions and times in the new chromosome mutate independently and are characterized by an irregular pattern and a non-sequential ping sequence. The second modification is in detection modeling, where a new model for bistatic detection was introduced. It allows for a combination of coherent and incoherent processing. For this work, we postulated that all sonobuoys could be monitored simultaneously. The SCOUT algorithms are an extension of our previous genetic algorithm work and, to the best of our knowledge, they represent the only solution that designs multi-static active sonobuoy placements in complicated environments from scratch, as opposed to recommending general effort allocations or simply evaluating standard patterns with different parameters. This paper discusses the new chromosome structure and simulation results in a realistic environment. The results show the following: a) SCOUT can effectively adapt multistatic sensor fields to the environmental complexity found in littoral areas; b) standard patterns are not optimal even for a homogeneous environment; c) standard patterns are grossly ineffective in inhomogeneous environments where 36–60% improvements in detection performance are achieved with SCOUT; d) 8–16 sonobuoys with SCOUT placement can perform as well as 32 regularly spaced sonobuoys; and e) extra flight time spent on laying irregular patterns is more than compensated by the additional tactical performance achieved.