CNN and MLP based Beam Squint and Channel Estimation for mmWave Massive MIMO-OFDM System

Abstract: Non-orthogonal multiple access (NOMA) is the promising technology in 5G and can utilize range proficiently. 5G devices have quicker data rate and are lower latency. We can use NOMA either as code, frequency or time. In NOMA we can serve multiple users simultaneously, hence NOMA has higher number of connectivity and improved user spectrum efficiency. There are mainly two types of NOMA plans, which are power-domain multiplexing and code-domain multiplexing. This followed power domain multiplexing by allocating different power levels to every users in according to the distance of the users. Nearest users provided with lowest power and farthest user provided with highest power. The Successive interference in NOMA is cancelled with the help of Coordinated Multipoint system (CoMP). In NOMA system, it is difficult to estimate Channel State Information (CSI). Here we introduce Least Square (LS), Minimum Mean Square Error (MMSE) and Neural Network methods to estimate CSI and compare these methods too.