Coded random access for massive MTC under statistical channel knowledge

Abstract

This paper focuses on random uplink transmissions of a subset of nodes disseminated in a cell. Under the constraints of massive Machine Type Communication (mMTC) in cellular Low Power Wide Area Networks (LPWAN) and Ultra Reliable Low Latency Communications (URLLC), improving the capability of a receiver to detect simultaneously several transmissions with a high probability is important. Considering a very limited coordination between the receiver and the distributed transmitters, the usage of coded Non Orthogonal Multiple Access (NOMA) strategies is seducing. In this framework, we target synchronous direct data transmissions and propose an optimal detector of the active users with channel state information at the receiver limited to statistical knowledge. This algorithm is based on a Maximum Likelihood (ML) detector, under statistical channel knowledge only. We give the formulation of the optimal detector and we evaluate its performance, with different codelengths, code types (random Gaussian and Grassmannian codes) and for various number of antennas at the base station.

Active User Blind Detection Through Deep Learning

“…However, the instantaneous channel states are not known, and no pilots are used in this detection phase. The detector thus operates in non-coherent detection mode [5].…”

“…n. This means that the message is sent over a narrowband channel, typically a single carrier in an OFDM frame, as defined in NB-IoT for MTC. The proposed model is similar to the one used in [5], [7].…”

“…Γ ∈ C M ×ω is composed of the singular values γ on its diagonal. From (2.1), following [5], the likelihood of a codeset is given by:…”

“…The maximum likelihood estimate (MLE) has been used in [5] to estimate the active set. Since we know the prior probability on (U, Φ), related to the Poisson distribution, a maximum a posteriori (MAP) detector can be defined and is optimal w.r.t.…”

“…Avoiding a handshake procedure requires efficient detection algorithms to retrieve active users' identity from a "one shot" access message with limited channel state information at the receiver (CSIR). The optimal AUD as described in [5] suffers from high complexity which does not seem compatible with real time implementation. An iterative version of the optimal detector, having a lower complexity -but also lower performance-, is also introduced therein.…”