Direction of Arrival and Channel Estimation using Machine Learning for Multiple Input Multiple Output System

The massive multiple input multiple output (MIMO) system is a prodigious method that could improvise a communication system's capacity with the aid of an increased number of antennas. The fundamental challenges of massive MIMO system are complicated channel modeling, high dimensional channel state information (CSI), and narrow radio frequency chains (RF chains), etc. For solving this problem we introduce deep neural network (DNN) framework for direction of arrival estimation (DOA estimation). Using the information obtained about DOA and gain that is complex, the channel is estimated and mean square error (MSE) performance is evaluated for both DOA and channel estimation.