Core Concepts
DU-JAD algorithm optimizes grant-free transmission in cell-free systems using deep unfolding for joint activity and data detection.
Abstract
Massive grant-free transmission and cell-free wireless communication systems are crucial for mMTC.
DU-JAD algorithm optimizes joint activity and data detection using deep unfolding.
FBS-based JAD algorithm is enhanced with deep unfolding, momentum strategy, PME, and soft-output AUD module.
Simulation results confirm the effectiveness of DU-JAD for AUD and DD.
Comparison with baselines shows significant improvement in AUD and DD performance.
Training procedures for FBS modules and AUD module are detailed.
System setup, performance metrics, and baselines for evaluation are provided.
Results show superior performance of DU-JAD in AUD and DD compared to baselines.
Stats
P = 400 UEs, P = 20-100 APs, M = 4 antennas, Pa = 0.2.
Active UEs transmit RP = 50 pilots and RD = 200 QPSK signals.
UDER of DU-JAD is 0.135 of Baseline 3 and 4 at P = 60.
ASER of DU-JAD is 0.129 of Baseline 3 and 4 at P = 60.
Quotes
"DU-JAD algorithm significantly outperforms existing baseline methods in active user detection and data detection performance."
"Simulation results confirm the efficacy of DU-JAD for AUD and DD."