Tensor-based Random Access for Ambient IoT Contention Resolution

Abstract

Ambient Internet of Things (A-IoT) deployments impose stringent hardware constraints, including low-order modulations and simple transceiver architectures. We propose a tensor-based grant-free random access scheme for the initial access message (Msg1) tailored to these limitations. Each device transmits replicas of its short payload structured as a rank-1 tensor over multiple access occasions. The A-IoT receiver performs joint activity detection, channel estimation, and decoding via tensor decomposition and successive interference cancellation. Results show significant throughput gains over conventional schemes while preserving low-complexity transmitter designs.