Analogue In-Memory Computing for Wireless Communication System

The increasing demand for higher frequencies and faster communication speeds in radio frequency (RF) systems poses challenges for traditional digital processing platforms, including high-speed analogue-to-digital conversion and intensive data movement, which lead to increased energy consumption and latency. We present an RF signal processing system that leverages analogue in-memory computing (AIMC) on a multicore memristive crossbar System on Chip (SoC), enabling high-speed and energy-efficient wireless communications. Our system directly processes RF signals within memory, bypassing analogue-todigital conversions and minimizing data movement. We demonstrate an analogue finite impulse response (FIR) filter and signal modulation type classification for real-time signal analysis coordinating multiple memristive computing cores within the SoC, the AIMC-based approach achieves over 99 % signal modulation classification accuracy, with energy efficiency and speed improvements of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$6.8 \times$</tex> and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$6.2 \times$</tex>, respectively, compared to traditional digital platforms, demonstrating its potential for future wireless communication applications.