Evaluation of the endfire response of a superdirective line array in simulated ambient noise environments

Superdirective line arrays can provide high gains whenever the inter-element spacing is much less than half a wavelength. In practice these arrays often fall short for two reasons: First, performance degrades appreciably in the presence of uncorrelated system noise. Second, the optimum array weights are specific to the form of the noise field and the choice of these weights is sensitive to deviations from that noise field. Furthermore, since the process involves computing signal differences between hydrophones, SNR degrades as the steering angle approaches broadside. In spite of these drawbacks, a superdirective array can provide substantial improvement over a conventional array in specific instances. For example, the present application requires an endfire array of limited extent (<1 m in length) and wide bandwidth (1-6 kHz) making a conventional array impractical. Moreover, improvements in array hardware have substantially reduced uncorrelated system noise, and increased computer speed makes it reasonable to process data using several weighting schemes. Given these conditions, a superdirective array provides a useful alternative. In this paper, the superdirective array will be described briefly and its performance will be evaluated using an ambient noise simulation model. The simulation allows one to estimate array performance in a variety of ambient noise environments using realistic values for system noise and number of averages, rather than relying on the results obtained in the theoretical limit. For example, assuming a 3-D isotropic noise field, and system noise of -30 dB relative to the ambient noise background, the simulation indicates that a 6-element array can achieve an array gain of 13 dB at 2.5 kHz and 9 dB at 0.8 kHz