Massive Access over an OFDM Platform

Project: Research

View graph of relations


Machine type communication (MTC) is an important service for the 5th and 6th generation (5G and 6G) cellular systems. MTC is characterized by a large number of mobile terminals (MTs), sporadic traffic, multiple concurrently active MTs and very short data messages. MTC brings about new challenges. Consider the following basic receiver functions at a base state (BS):(i) identify active MTs who have information to transmit,(ii) estimate the channels from active MTs, and(iii) decoding information bits.A random-access system, such as ALOHA or CSMA, completes these functions in a single phase. However, collisions of multiple concurrently active MTs seriously affect the throughput in ALOHA or CSMA. The existing cellular system, on the other hand, separates these functions into multiple phases, among which four phases are for function (i) (such as RACH in 4G for connection setup), and one additional phase for functions (ii) and (iii) (using pilot and data slots). In MTC, the amount of data per session is typically small. Then the overhead due to connection setup becomes a serious burden. Recently, a two-phase massive-access scheme attracted wide research interests for MTC, in which functions (i) and (ii) are carried out jointly in the first phase, and function (iii) in the second phase. This two-phase structure is sub-optimal. We can see such optimality by noting that the outputs of function (iii) can be used to enhance the accuracy of functions (i) and (ii). For example, a decoding failure indicates a high probability of error in activity detection. If this information is utilized, we may refine the outputs of the first phase. In this project, we will work on this idea and develop a scheme that can complete the functions (i) – (iii) in a single phase. Our focus is on the orthogonal frequency division multiplexing (OFDM) platform. OFDM is the underlying technology for 5G, and possibly for 6G as well. The sub-carriers in an OFDM system are typically strongly correlated. The existing AMP based schemes cannot perform well in such correlated environments. Our recent work on the orthogonal approximate message passing (OAMP) algorithm revealed an efficient treatment for the correlation problem. In this project, we will develop a joint OAMP and FEC decoding scheme to complete functions (i) – (iii) in a single phase. It will provide a reliable and low-cost implementation technique for MTC. The outputs of this project may find useful applications in future cellular systems.


Project number9043131
Grant typeGRF
Effective start/end date1/01/22 → …