Efficient Hardware Acceleration of QR Decomposition via In-Memory Computing for Signal Processing Applications

QR decomposition is a fundamental matrix factorization technique widely used in signal processing, wireless communication, and machine learning. Traditional hardware implementations face challenges due to high computational complexity, significant data movement, and energy inefficiency. To address these issues, this paper proposes an in-memory computing (IMC)-based approach for efficient realization of QR decomposition using memristor crossbar arrays. By mapping matrix operations directly onto the crossbar, the proposed scheme minimizes data transfers and leverages the inherent parallelism of memristors to accelerate computations. A comparative analysis with conventional designs highlights the advantages of IMC-based approach in computational efficiency, scalability, and hardware cost, positioning it as a promising solution for high-performance QR decomposition in next-generation computing systems.