Federated Learning with LSTM Based Active User Detection for CF-mMIMO Network for 6G Communication System

Effective user detection and resource allocation techniques are required in cell-free massive multiple input multiple output (CF-mMIMO), a sixth generation (6G) communication system to improve network performance by providing grant-free access to the active user. Accurate active user detection (AUD) while reducing needless resource allocation to idle users is one of the main issues in such systems. This work proposes a Federated Learning (FL) with long-short-term memory (LSTM) deep learning-based decentralized AUD and user clustering framework for CF-mMIMO systems. We employ an LSTM model for active user classification and K-Means clustering to identify the user’s activity within the cluster based on signal strength and available bandwidth. The Rician fading technique is incorporated into the simulation to ensure a realistic wireless environment. Experimental results demonstrate that the proposed method significantly reduces communication overhead and provides high detection accuracy.