On Simultaneous Multipacket Channel Estimation and Reception in Random Access for MTC Under Frequency-Selective Fading

Abstract

In random access for machine-type communications, it is required not only to detect active devices, but also to estimate their channel state information when they transmit preambles. In this paper, we consider an approach for simultaneous multipacket channel estimation and reception (SMuCER) with a multicarrier system under frequency-selective fading, where the channel estimation becomes challenging due to frequency-selective fading and different round-trip delay between the devices. In the proposed approach, circular-shifted preambles are used to effectively increase the number of preambles (that can lower the probability of preamble collision), and compressive sensing (CS)-based multiuser detection (MUD) is employed. In order to have a good performance of CS-based MUD for SMuCER, we consider Alltop sequences. Furthermore, we formulate an optimization problem to maximize the number of successfully recovered active devices, where a large single system is to be divided into multiple subsystems. It is shown that multiple subsystems have not only a higher throughput, but also a lower computational complexity than a large single system.