Multiuser Superposition Transmission

Description

Multiuser Superposition Transmission (MUST) is a downlink physical layer transmission scheme standardized in 3GPP for LTE and studied for NR. It is a form of Non-Orthogonal Multiple Access (NOMA) where the base station (eNB or gNB) transmits a composite signal by superimposing the modulated symbols intended for multiple users onto the same physical resource block (PRB) in the time and frequency domain. The superposition is achieved by allocating different power levels to each user's signal. Typically, a cell-center user (with good channel conditions) is assigned lower power, while a cell-edge user (with poor channel conditions) is assigned higher power. The composite signal is broadcast, and each user employs Successive Interference Cancellation (SIC) at the receiver to decode its intended information.

The architecture of MUST involves enhancements to the scheduler and the physical downlink shared channel (PDSCH) processing chain. The scheduler pairs users with significantly different channel gains (e.g., different distances from the base station) for superposition on the same resources. It determines the power allocation ratio between the paired users. The transmitter then performs constellation superposition, where the constellation points for the far user and the near user are combined into a new, denser constellation for transmission. Key components include the MUST user pairing algorithm, the power allocation controller, and the generation of the superimposed constellation mapping, which must be known or signaled to the receivers for successful SIC.

At the receiver, the near user (with the better channel) first decodes the far user's signal by treating its own signal as noise, thanks to the higher power allocation of the far user's signal. After successfully decoding and reconstructing the far user's signal, the near user subtracts (cancels) it from the received composite signal. It then proceeds to decode its own signal from the cleaner residual signal. The far user, with its poorer channel, simply decodes its own signal directly, treating the near user's lower-power signal as additional noise, which has a minimal impact due to the power disparity. This process allows both users to share the same radio resources, thereby increasing the overall spectral efficiency and system capacity compared to traditional Orthogonal Multiple Access (OMA) techniques like OFDMA.

Purpose & Motivation

MUST was developed to address the ever-increasing demand for higher spectral efficiency and user capacity in cellular networks, especially as data traffic grows exponentially. Traditional orthogonal multiple access schemes, like OFDMA used in LTE, allocate exclusive time-frequency resources to each user to avoid interference. While this simplifies receiver design, it limits the number of simultaneously served users, particularly in dense scenarios. MUST was motivated by the need to break this orthogonality barrier and serve more users within the same bandwidth, a concept central to NOMA.

The historical context for MUST's introduction in 3GPP Release 14 was the industry's exploration of NOMA techniques as a key candidate for 5G. Early research showed significant potential gains in throughput and connectivity. MUST specifically aimed to enhance LTE networks in a backward-compatible manner, providing a smooth evolution path. It targeted scenarios with diverse user channel conditions, such as a mix of cell-center and cell-edge users, where the power-domain superposition is most effective. By allowing these users to share resources, MUST improves fairness and cell-edge throughput, which are critical metrics for user experience.

Furthermore, MUST addresses the limitation of purely orthogonal scheduling in handling massive connectivity, a requirement for the Internet of Things (IoT). While its primary study was in LTE, the principles informed NOMA discussions for NR. The technology solves the problem of resource scarcity by exploiting the power domain as an additional multiplexing dimension. Its creation was driven by the need for more efficient use of licensed spectrum, ultimately aiming to deliver higher data rates and support more concurrent users without requiring additional bandwidth, which is a scarce and expensive resource for operators.