Secure Transmission for Uplink SCMA Systems Over $\kappa$-$\mu$ Fading Channels

Sparse code multiple access (SCMA) is an interesting proposal for supporting massive machine-type communications (mMTC) in future mobile networks. In this use case, mobile networks connect to Internet of Things (IoT) devices, which may have low computational processing capacities, and thus, they are exposed to different security threats. In this regard, physical-layer security (PLS) arises as an alternative technique to protect information. This work proposes a new scheme based on the design of unknown codebooks to an external eavesdropper aiming at guaranteeing PLS in the uplink of SCMA systems over generalized
$\kappa$
-
$\mu$
fading channels. Considering that the SCMA codebook design can be realized from the rotation of base quadrature-amplitude modulation (QAM) constellations using the signal-space-diversity (SSD) technique, a novel upper-bound expression to calculate the average bit error probability (ABEP) of multidimensional SSD systems is derived. Then, a metric to select the best rotation matrices is determined. With this, a set of optimal and suboptimal rotation matrices can be acquired. A strategy to obtain secret keys in the SCMA system is presented by considering imperfect channel estimation. The secret keys are used to select matrices for rotating the base QAM constellations used in the codebooks generation. Finally, an uncoded and turbo-encoded SCMA system is evaluated using Monte Carlo simulations. The results show that our proposal ensures a near-optimal SCMA performance while the eavesdropper cannot decode the information.