Blind signal separation for underwater lidar applications

Abstract

Abstract. Blind signal separation (BSS) is demonstrated to improve signal-to-interference ratio (SIR) in underwater light detection and ranging (lidar) applications. Lidar systems are used for high-resolution ranging and imaging underwater. A difficult problem in this area is detecting the return from an object of interest in the presence of a strong "clutter" return caused by backscattering in low visibility water environments. The principal component analysis form of BSS is applied to separate the return into multiple subspaces, with the backscatter subspace suppressed to improve detection capability. Simulations are performed using an underwater optical channel model to simulate a challenging harbor-like environment. We show that the processing gain from BSS is increased when the correlation energy between observations is increased, for example by using closely spaced observations. The SIR gain of BSS is explored as a function of frequency, achieving substantial gain independently of carrier frequency. This result has important implications on the design of lidar systems, suggesting that BSS can be combined with low-cost, low-frequency components to achieve performance similar to systems using high-frequency components.