Complex-Weight Sparse Linear Array Synthesis by Bayesian Compressive Sampling

• Published in: IEEE transactions on antennas and propagation Vol. 60; no. 5; pp. 2309 - 2326
• Main Authors: Oliveri, G., Carlin, M., Massa, A.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.2012; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Antennas; Applied sciences; Array synthesis; Arrays; Assessments; Bayesian analysis; Bayesian compressive sampling; Bayesian methods; complex-weight pattern; Electromagnetism; Engineering Sciences; Exact sciences and technology; Layout; Linear arrays; Matching; Mathematical models; Pattern matching; Probabilistic logic; Radiocommunications; Sampling; shaped-beam pattern; Signal and Image processing; sparse arrays; Sparse matrices; Synthesis; Telecommunications; Telecommunications and information theory; Vectors; Performance evaluation; Sparse array; State of the art; Probabilistic approach; complex-weight pattern; Hermite interpolation; Bayesian compressive sampling; linear arrays; Pattern recognition; Vector method; Flexibility; Implementation; Constrained optimization; Array synthesis; Multiple task; Relevance criterion; shaped-beam pattern, sparse arrays; Linear antenna arrays; Reliability; Pattern matching; Beam shaping; Compressed sensing; Antenna synthesis
• ISSN: 0018-926X; 1558-2221
• DOI: 10.1109/TAP.2012.2189742

An innovative method for the synthesis of maximally sparse linear arrays matching arbitrary reference patterns is proposed. In the framework of sparseness constrained optimization, the approach exploits the multi-task ( MT ) Bayesian compressive sensing ( BCS ) theory to enable the design of complex non-Hermitian layouts with arbitrary radiation and geometrical constraints. By casting the pattern matching problem into a probabilistic formulation, a Relevance-Vector-Machine ( RVM ) technique is used as solution tool. The numerical assessment points out the advances of the proposed implementation over the extension to complex patterns of and it gives some indications about the reliability, flexibility, and numerical efficiency of the MT - BCS approach also in comparison with state-of-the-art sparse-arrays synthesis methods.