Sum Data Rate and Energy Efficiency optimization of MIMO NOMA using CDNN

The Non-Orthogonal Multiple Access (NOMA) is considered as a technique that satisfies increasing demands of spectrum efficiency and system capacity in future 5G networks since the technique serves numerous number of users in one resource block with multiple power levels. The application or execution of multiple input multiple output (MIMO) to NOMA system helps in building up the performance gains of NOMA. Here an effectual communication deep neural network (CDNN) that addresses problem of power allocation in MIMO NOMA system is designed and used which helps to integrate the deep learning technique with the MIMO NOMA system in order to provide an improved sum data rate and energy efficiency of the system. Improved sum data rate and energy efficiency is realized further using extensive simulations and it is then compared to realise the performance of the neural network. The results then shows that the CDNN integrated MIMO NOMA system shows the best performance and also the algorithm is efficient for the systems with both mobile and stationary users.