Sparse Code Multiple Access

Ziyang Li and Wen Chen

Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai, China

I. S YNONYMS analysis area, researchers compare the capacity of SCMA

Sparse spreading multiple access system with the LDS system [8], and analyse the capacity of
downlink massive MIMO MU-SCMA system [9]. In codebook
II. D EFINITION design area, the method of Shuffling [6] and the method based
Sparse code multiple access (SCMA) is a non-orthogonal on star-QAM [10] are proposed to decrease the Bit Error
multiple access technique based on codebooks of sparse struc- Rate(BER). In low-complexity decoding area, some algorithms
ture. like PM [11], Quasi-ML [12], List Sphere Decoding [13]
are proposed to reduce the decoding complexity. In resource
III. H ISTORICAL BACKGROUND allocation area, the SCMA downlink power allocation to max-
Future 5G wireless networks are expected to support mas- imize weighted sum-rate has been studied in [14] and uplink
sive connectivity, higher throughput, lower latency and lower resource allocation based on a special structure codebook
controlling signal overhead. Compared with the conventional studied in [15]. In blind detection area, contention based blind
orthogonal multiple access techniques such as code division detection is proposed to support Grant-free transmission [16],
multiple access(CDMA) and orthogonal frequency division [17].
multiple access (OFDMA) utilized in current networks, non-
orthogonal multiple access (NOMA), which has the advantage IV. F OUNDATIONS
of higher resource utilization, is more promising in 5G net- SCMA system mainly consists of 2 parts: SCMA encoder
work. and SCMA decoder.
Motivated by the design of CDMA chip sequences, the idea SCMA encoder can be defined as a mapping from log2 (M )
of sparse spreading, called Low Density Signature (LDS), was bits to a K-dimensional codeword of size M selected from a
first proposed by Reza Hoshyar [1]. LDS is a special case of predefined codebook. K dimensions are corresponding to K
multi-carrier CDMA with a few non-zero elements in a long different orthogonal tones, such as OFDMA subcarriers. The
spreading signature. This sparsity characteristic can reduce K-dimensional codeword is a vector with only N < K non-
the complexity of the multi-user detection called Message zero entries. Users can’t transmit data through the subcarriers
Passing Algorithm (MPA) [2], [3]. In an LDS combined with represented by the other N − K zero entries. Theoretically,
Orthogonal Frequency Division Multiplexing (OFDM) system, each user can be allocated to more than one codebook, and
which is called LDS-OFDM [4], the capacity is proved to have each codebook can be utilized by more than one user generally.
a superior performance than OFDMA. Fig. 1 shows an example of an SCMA encoder, with 6
To meet the requirements of 5G, Sparse Code Multiple layered codebooks (variable nodes) and 4 subcarriers (function
Access (SCMA) [5], a new non-orthogonal codebook-based nodes). Each row denotes a dimension, and each column
multiple access method, is proposed by Huawei in 2013. means a 4-dimensional codeword. In each codebook, the
SCMA can be seen as a generalization of LDS. In LDS, in- constellation size is 4, which means there are 4 different
coming bits are mapped to a QAM symbol, and the repetitions codewords can be chosen. The white entries denote the zero
of the QAM symbol are transmitted through the subcarriers ac- elements and the colored entries denote the non-zero elements
cording to the designed signature. But in SCMA, incoming bits in the codebooks. For example, in Codebook 1, the entries in
are directly mapped to multi-dimensional complex codewords the first row is colored and the entries in the third row is
selected from predefined codebook [6]. Instead of simple white, which means the first dimension is non-zero and third
repetition of QAM symbol in LDS, SCMA provides significant dimension is zero. In each codebook, there are 2 non-zero
shaping gain with the multi-dimensional constellation design dimensions with colored lattice. In an AWGN channel, the
[7]. Similar to LDS, only a small number of dimensions signal received in the base station is the superposition of the
are used to transmit data, therefore the codewords of SCMA codewords selected from the codebooks.
are sparse. In the SCMA codebook, the non-zero dimensions Codebook design is the most important part in SCMA en-
are corresponding to the subcarriers in use. Besides, SCMA coder. The target is to design a multi-dimensional lattice con-
can also utilize iterative MPA detection with near optimal stellation with dimensional dependency and power variation of
performance in the MAP sense. the constellation while maintaining large minimum Euclidean
The current researches of SCMA are mainly focused on distance. Generally, there are 3 stages to design SCMA code
capacity analysis, codebook design, low-complexity decoding, [5], [6]: 1) Mapping Matrix determines the number of layers
resource allocation, blind detection and so on. In capacity interfering at each subcarrier, which represents the complexity
 2

VI. C ROSS - REFERENCES

code-division multiple access
frequency-division multiple access
5G Communication Technology
Internet of Things
Technologies Enabling The Internet of Vehicles
Non-Orthogonal Multiple Access (NOMA) for Future Radio
Access
VII. R ECOMMENDED R EADING
R EFERENCES
[1] R. Hoshyar, F. P. Wathan, and R. Tafazolli, “Novel low-density signature
Fig. 1: SCMA encoder. for synchronous cdma systems over awgn channel,” Signal Processing,
IEEE Transactions on, vol. 56, no. 4, pp. 1616–1626, 2008.
[2] R. Razavi, M. A. Imran, and R. Tafazolli, “Exit chart analysis for turbo
of MPA detection. For example, Fig. 1 can be considered lds-ofdm receivers,” in Wireless Communications and Mobile Computing
Conference (IWCMC), 2011 7th International. IEEE, 2011, pp. 354–
as a mapping matrix, which means that each layer will be 358.
interfered by two other layers. 2) Constellation Points and [3] R. Razavi, M. Al-Imari, M. A. Imran, R. Hoshyar, and D. Chen,
Multi-dimensional Mother Constellation design. First, design “On receiver design for uplink low density signature ofdm (lds-ofdm),”
Communications, IEEE Transactions on, vol. 60, no. 11, pp. 3499–3508,
a base constellation with a maximized minimum Euclidean 2012.
distance. Second, a unitary rotation, which might be designed [4] R. Hoshyar, R. Razavi, and M. Al-Imari, “Lds-ofdm an efficient multiple
to maximize the minimum product distance of the constella- access technique,” in Vehicular Technology Conference (VTC 2010-
Spring), 2010 IEEE 71st. IEEE, 2010, pp. 1–5.
tion, can be applied on the base constellation to control the [5] H. Nikopour and H. Baligh, “Sparse code multiple access,” in Personal
dimensional dependency and power variation. Third, build the Indoor and Mobile Radio Communications (PIMRC), 2013 IEEE 24th
complex constellation using the rotated base constellation by International Symposium on. IEEE, 2013, pp. 332–336.
[6] M. Taherzadeh, H. Nikopour, A. Bayesteh, and H. Baligh, “Scma
shuffling. Last, utilize the rotation to minimize the projection codebook design,” in Vehicular Technology Conference (VTC Fall), 2014
points. 3) Constellation Function Operator, which includes IEEE 80th. IEEE, 2014, pp. 1–5.
several operators like complex conjugate, phase rotation and [7] J. Boutros, E. Viterbo, C. Rastello, and J.-C. Belfiore, “Good lattice con-
stellations for both rayleigh fading and gaussian channels,” Information
dimensional permutation, aims to design distinct codebooks Theory, IEEE Transactions on, vol. 42, no. 2, pp. 502–518, 1996.
for the collision layers. [8] M. Cheng, Y. Wu, and Y. Chen, “Capacity analysis for non-orthogonal
Currently, SCMA decoder usually utilizes MPA detection overloading transmissions under constellation constraints,” in Interna-
tional Conference on Wireless Communications & Signal Processing.
or improved version of MPA detection. There are three steps IEEE, 2015, pp. 1–5.
of the conventional MPA detection: 1) Initialize the condi- [9] T. Liu, X. Li, and L. Qiu, “Capacity for downlink massive mimo mu-
tional probability. For the SCMA system in Fig. 1, in each scma system,” in International Conference on Wireless Communications
& Signal Processing. IEEE, 2015, pp. 1–5.
resource node, there are 3 layers collided and each layer has [10] L. Yu, X. Lei, P. Fan, and D. Chen, “An optimized design of scma
4 possibilities of constellation points. Then there are 4 ∗ 4 ∗ 4 codebook based on star-qam signaling constellations,” in International
combinations of transmitted signals which determine the num- Conference on Wireless Communications & Signal Processing. IEEE,
2015, pp. 1–5.
ber of conditional probabilities. 2) Iterative message passing [11] H. Mu, Z. Ma, M. Alhaji, P. Fan, and D. Chen, “A fixed low complex-
through the variable nodes and function nodes. Function node ity message pass algorithm detector for up-link scma system,” IEEE
passes the updated message to its neighbouring variable nodes, Wireless Communications Letters, vol. 4, no. 6, pp. 585–588, 2015.
[12] J. Zou, H. Zhao, and X. Li, “A low-complexity tree search based quasi-
then the variable node passes the updated message to its ml receiver for scma system,” in IEEE International Conference on
neighbouring function node, which constitutes an iteration. Computer and Communications. IEEE, 2016, pp. 319–323.
3) Log-Likelihood-Rate(LLR) calculation. After several iter- [13] F. Wei and W. Chen, “Low complexity iterative receiver design for sparse
code multiple access,” IEEE Transactions on Communications, vol. 65,
ations, LLR output can be calculated by the probability guess no. 2, pp. 621–634, 2017.
of codeword at each layer. [14] H. Nikopour, E. Yi, A. Bayesteh, K. Au, M. Hawryluck, H. Baligh, and
J. Ma, “Scma for downlink multiple access of 5g wireless networks,” in
Global Communications Conference (GLOBECOM), 2014 IEEE. IEEE,
V. K EY A PPLICATIONS 2014, pp. 3940–3945.
[15] Z. Li, W. Chen, F. Wei, F. Wang, X. Xu, and Y. Chen, “Joint codebook
assignment and power allocation for scma based on capacity with gaus-
Sparse code multiple access technique will be utilized in fu- sian input,” in Ieee/cic International Conference on Communications in
ture 5G networks, which will be applied mainly in 3 scenarios: China. IEEE, 2016, pp. 1–6.
[16] A. Bayesteh, E. Yi, H. Nikopour, and H. Baligh, “Blind detection of
1) Enhanced mobile broadband (eMBB), for example, Ultra scma for uplink grant-free multiple-access,” in International Symposium
HD video transmission and High Speed Mobile Broadband on Wireless Communications Systems. IEEE, 2014, pp. 853–857.
Wireless Communication. 2) Ultra-reliable and low latency [17] K. Au, L. Zhang, H. Nikopour, E. Yi, A. Bayesteh, U. Vilaipornsawai,
J. Ma, and P. Zhu, “Uplink contention based scma for 5g radio access,”
communications (uRLLC), for example, unmaned and auto- in Globecom Workshops (GC Wkshps), 2014. IEEE, 2014, pp. 900–905.
matic driving. 3) massive machine type of communication
(mMTC), for example, Internet of things (IoT).