Abstract :
This paper presents a novel and efficient extrema-mapping algorithm, which we call the roller-coaster algorithm. Two versions of the algorithm, the one-dimensional (1-D) and the two-dimensional (2-D) roller-coaster, are developed. Its applicability to array signal processing is demonstrated. We use it to solve a multiple source direction finding problem using multiple signal classification (MUSIC), beamformer, and minimum variance methods, and for antenna array design. The algorithm is based on heuristic assumptions and its properties are not proved. Yet, its performance was tested in many simulated experiments, yielding favorable results
Keywords :
Newton method; antenna arrays; array signal processing; direction-of-arrival estimation; maximum likelihood estimation; optimisation; signal classification; 1D roller-coaster; 2D roller-coaster; DOA estimation; MAXLE; Newton algorithm; antenna array design; array signal processing; beamformer method; computational complexity; extrema-mapping algorithm; heuristic assumptions; minimum variance methods; multiple signal classification; multiple source direction finding problem; numerical optimization method; roller-coaster algorithm; search and initialise; sparse arrays; Array signal processing; Convergence; Direction of arrival estimation; Maximum likelihood estimation; Multiple signal classification; Optimization methods; Phased arrays; Signal processing algorithms; Spectral analysis; Testing;
