Capacity Analysis of Partially Overlapping NOMA System with Two Users

Abstract

In recent years the Internet of Things and other massive machine-type communication applications demand increased spectral efficiency for future radio multiple access schemes. One of the key enabling technologies is power domain non-orthogonal multiple access (NOMA) with successive interference cancellation (SIC), which works by superposing multiple users in the same frequency band, and at the receiver side successive decoding the signals. The key advantage here compared to a traditional orthogonal multiple access (OMA) scheme like frequency division multiple access (FDMA) is that each user can utilize the full available bandwidth. The SIC method however has limitations. It requires precise estimation of the channel conditions of each user by the receiver. It works best when there is a large enough difference between the receiving signal strengths of the different superposed user's signals. At the marginal case, when all the signals has the same signal to noise ratio (SNR) there is no advantage compared to FDMA. In this paper we propose a method that lies between the full spectral overlap of the power domain NOMA and the no overlap of the FDMA schemes, allowing the user's signal to partially overlap, by stretching the frequency band occupied by one user to utilize wider bandwidth, allowing the neighboring signal to partially overlap. The overlapping part of the neighboring signal causes interference, which is treated as part of the noise, reducing the capacity, but the wider bandwidth will increase the capacity. In this contribution we investigate the balance between this gain and loss of this proposed scheme for the two user case where both users have similar SNR conditions, and we show that depending on the actual conditions in many cases significant capacity increase can be reached.