부산대학교 도서관

Rate Splitting Multiple Access (RSMA) has emerged as a powerful physical (PHY) layer multiple access technique for multi-antenna networks that achieves better outcomes – in terms of spectral efficiency, energy efficiency, etc. – than the conventional Space Division Multiple Access (SDMA) used in 3GPP 5G NR and IEEE 802.11.In RSMA, to simultaneously serve K receivers (RXs), a transmitter (TX) splits the message, $W_{k}$, intended for the $k$-th RX into common and private portions denoted by $W_{c, k}$ and $W_{p, k}$, respectively. All the common portions - i.e, $\left\{W_{c, k}: k=1, \cdots, K\right\}$ - are combined and encoded to form a common data stream, $s_{c}$. Likewise, each private portion $W_{p, k}$ is individually encoded to form a private data stream, $s_{k}(k=1, \cdots, K)$. Thus, the K messages are used to form $K+1$ data streams (one common and K private streams). Each stream is then individually precoded and transmitted over the air. At the $k$-th RX, the common stream $\left(s_{c}\right)$ is first decoded by treating the other interfering private streams as noise to retrieve $\left\{W_{c, j}: j=1, \cdots, K\right\}-$ i.e., not just (a part of) the desired message ($W_{c, k}$), but also (a part of) the multi-user interference $\left(\left\{W_{c, j}: j \neq k\right\}\right)$. Then, the corresponding private stream $\left(s_{k}\right)$ is decoded (e.g., through successive interference cancellation (SIC)) to retrieve $W_{p, k}$. From $W_{c, k}$ and $W_{p, k}$, the $k$-th ${\bf R X}$ can reconstruct its intended message, $W_{k}$ [1]. The salient features of RSMA are:•Enhanced interference suppression: Each RX partially decodes the multi-user interference as described above, which increases the spectral efficiency.•Versatility: The extra precoded stream over SDMA (i.e., $K +1$ instead of $K)$ enables over 40 new/emerging applications in future wireless networks, e.g., Integrated Sensing and Communications (ISAC) [1, Section V].•Generality: Crucially, SDMA is a special case of RSMA (i.e., no message splitting and hence, no common stream).The above features make RSMA promising for standardization in future evolutions of 3GPP 5G NR and IEEE 802.11.To raise RSMA’s technology readiness level (TRL) for standardization, two independent RSMA prototypes have been built using software-defined radios by teams at Imperial College London [2] and Viavi Solutions UK Ltd [3, Section III]. For the two RX case, the RSMA sum throughput was measured using the Imperial prototype over line-of-sight (LoS) channels. RSMA achieved a gain of upto 57% and more fairness than SDMA over highly interfering channels. Additionally, with quantized channel state feedback, RSMA experienced a smaller throughput loss relative to the unquantized feedback case – 37%, on average, as opposed to 44% for SDMA. For a detailed discussion, see [2].These impressive early results notwithstanding, RSMA’s benefits need to be further demonstrated for different use cases and operating environments with minimal increase in complexity, especially at the RX. Our approach towards addressing these challenges is summarized below:•RX Complexity: RSMA is often criticized for the increased RX complexity due to SIC. However, SIC-free RXs can also achieve near-SIC performance [4]. Realizing these RX architectures in our prototype upgrades is a work in progress.•Extended enhanced Mobile Broadbanc (eMBB): Along with unicast, raising RSMA’s TRL for other eMBB use cases, like multicast and joint unicast/multicast, is essential to demonstrate its suitability to handle the traffic demands of future networks. Currently, these measurements are underway using the Imperial prototype.•Emerging Use Case – ISAC: RSMA is ideally suited for ISAC, as the extra precoded stream can be used for both communications and sensing. Realizing this potential would no doubt enhance RSMA’s standardization appeal.•Challenging Operating Environments: Non-LoS and high mobility are benchmark environments for evaluating RSMA’s performance.Finally, in terms of standardization efforts, BBC [5] and Viavi [6] submitted technical contributions on RSMA to the 3GPP Release 19 workshop in June 2023, as a precursor to promoting it as a study item for Release 20.