Application of Polynomial and Radial Basis Function Maps to Signal Masking

摘要

The objective of this research was to develop and demonstrate a technique for encrypting information by using a masking signal that closely approximates local ambient noise. Signal masking techniques developed to date have used nonlinear differential equations, spread spectrum, and various modulation schemes to encode information. While these techniques can effectively hide a signal, the resulting masks may not appear as ambient noise to an observer. The advantage of the proposed technique over commonly used masking methods is that the transmitted signal will appear as normal background noise, thus greatly reducing the probability of detection and exploitation. A promising near-term application of this technology presents itself in the area of clandestine minefield reconnaissance in shallow water areas. Shallow water mine-counter-mine (SWMCM) activity is essential for minefield avoidance, efficient minefield clearance, and effective selection of transit lanes within minefields. A key technology area for SWMCM is the development of special sonar waveforms with low probability of exploitation/intercept (LPE/LPI) attributes. In addition to LPE/LPI sonar, this technology has the potential to enable significant improvements in underwater acoustic communications. For SWMCM, the chaotic waveform research provides a mechanism for encrypted communications between a submarine (SSN) and an unmanned underwater vehicle (UUV) via an acoustic channel. Acoustic SSN/UUV communications would eliminate the need for a fiberoptic link between the two vessels, thus increasing the robustness of SWMCM. Similar applications may exist in the areas of radar masking and secure communications. The original approach called for the use of polynomial maps to generate a masking signal. Because polynomial maps were found to have highly restrictive stability criteria, the approach was modified to use radial basis function (RBF) maps. they have shown that stable RBF maps that closely approximate an ambient sea state can be derived using nonlinear systems theory. In doing so, they have shown that the measured ambient state has a deterministic structure that implies eight-order dynamics. The RBF maps were used to successfully encrypt a continuous wave signal across a high-fidelity, low-noise transmission path. Attempts to duplicate this result across a low-fidelity, high-noise path were not successful.