Joint Optimization of Decoding Order, Power Allocation and Reflection Coefficients in an IRS Assisted Untrusted Secure NOMA

This letter studies a non-orthogonal multiple access (NOMA) based downlink secure communication system in the presence of an intelligent reflecting surface (IRS). More specifically, we look at the joint optimization of decoding order, power and IRS reflection coefficients. After obtaining optimal decoding order, the power and reflection coefficients are obtained by splitting the problem into individual problems. Power allocation coefficient is derived for a known reflection coefficients, and reflection coefficients are obtained for a known power allocation. Both these sub-problems are solved analytically with closed form solutions. By switching between these independent optimizations, the joint optimization is accomplished. The numerical findings highlight design lessons for securing IRS-assisted untrusted NOMA systems, and demonstrate that the jointly optimal solution gives better performance compared to the benchmarks.