Abstract
Grant-free random access is a key enabler in massive machine-type communications (mMTC) to reduce signalling overhead and latency thereby improving the energy efficiency. One of its main challenges lies in joint user activity identification and channel estimation (JUICE). Due to the sporadic mMTC traffic, JUICE can be solved as a compressive sensing (CS) problem. We address CS-based JUICE in uplink with single-antenna transmitters and a multiantenna base station under spatially correlated fading channels. We formulate a novel CS problem that utilizes prior information on the second order statistics of the channel of each user to improve the performance. We propose a method based on alternating direction method of multipliers to solve the JUICE efficiently. The simulation results show that the proposed method significantly improves the user identification accuracy and channel estimation performance with lower signalling overhead as compared to the baseline schemes.
Highlights
M ASSIVE machine-type communications aim to provide wireless connectivity to billions of low-cost energy-constrained Internet of Things (IoT) devices [1]
We address the joint user activity identification and channel estimation (JUICE) problem in M ASSIVE machine-type communications (mMTC) under spatially correlated fading channels
Numerical results are provided to show the performance in terms of user activity detection accuracy, channel estimation quality, and convergence rate for the proposed method with comparison to different measurement vector (MMV) reconstruction algorithms
Summary
M ASSIVE machine-type communications (mMTC) aim to provide wireless connectivity to billions of low-cost energy-constrained Internet of Things (IoT) devices [1]. The main advantage of grant-free access compared to conventional random access is the reduced signalling overhead and the improved energy-efficiency of the IoT devices. The vast majority of JUICE related works assume that the communication channels are spatially uncorrelated. This assumption leads to analytically tractable solutions, it is not always the case in practice [8]. As the channel spatial correlation varies in a slower time-scale compared to the channel realizations, the channel covariance matrices for all IoT devices can be estimated with high accuracy in practice [10]. The proposed approach is empirically shown to significantly enhance the user activity detection accuracy and, the channel estimation performance. The proposed approach achieves the same performance as baseline MMV JUICE, yet with a smaller signalling overhead
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
