Abdesselam Klouche-Djedid,三浦龍
A High Resolution Algorithm for Detection-estimation of Narrow-band Signals using Sensor Arrays without Eigendecompositions of Data Correlations
Abstract:High resolution algorithms in sensor arrays lead to accurate results but with expensive eigendecompositions making its use in real-time applications such as mobile communications relatively difficult. In this technical report, a trade-off between accuracy and computational load is accomplished through a simplified algorithm which instead of eigendecompositions, uses the robust QR factorization for which many efficient parallel (systolic, wavefront array) implementations exist. First, a simple detection scheme is presented and, through simulations, is shown to work very well for sufficient SNR, even when signals are coherent. Outputs of the detection process include simultaneously estimates of signals Direction Of Arrivals (DOA's) and a simple beamformer vector resulting in an estimate of the desired signal.
Extensive simulations are performed assuming different scenarios of variations in SNR, DOA's leading to discussions on the possibilities and limitations of the proposed solution.