Memory Optimization for FPGA Implementation of Correlation-Based Beamforming

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2024

Autores

Avelar, H (Autor) Outra; João Canas Ferreira (Autor) FEUP

Ata de Conferência Internacional

Título: 2024 IEEE 22ND MEDITERRANEAN ELECTROTECHNICAL CONFERENCE, MELECON 2024

Páginas: 196-201

IEEE 22nd Mediterranean Electrotechnical Conference (MELECON)

Porto, PORTUGAL, JUN 25-27, 2024

Indexação

ISI Web of Knowledge - 0 Citações

Scopus - 0 Citações

Outras Informações

ID Authenticus: P-016-W3E

DOI: 10.1109/melecon56669.2024.10608567

Abstract (EN): This paper proposes a method to avoid using a CORDIC or external memory to process the steering vectors to calculate the pseudospectrum of correlation-based beamforming algorithms. We show that if we decompose the steering vector equation, the size of the matrix to be saved in memory becomes independent of the antenna array size. Besides, the amount of data needed is small enough to be saved in the internal block RAMs of the FPGA SoC. Besides, this method greatly reduces the number of memory accesses, by offloading some processing to hardware, while keeping the frequency at 300MHz with a precision of 0.25 degrees. Finally, we show that this approach is scalable since the complexity grows logarithmically for bigger arrays, and the symmetry in the matrices obtained allows even more compact data.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 6