Robust Array Processing for DOA Estimation With Widely Spaced Sensors

This paper deals with DOA estimation techniques used for tracking a mixed layer float. Cross correlation followed by additional filtering gives us satisfactory delay estimates. These delay estimates are then suitably weighted to obtain a DOA estimate whose tracking is enhanced by the use of Kalman filtering. To improve the resolution of this preliminary DOA estimation and to provide versatility, high resolution subspace-based algorithms are used. The high resolution techniques for this problem require time-shifting to overcome the problem of widely spaced sensors.