Empowering Communication Via Optimizing Non-RIS and RIS aided Models with Gradient Descent: A Comparative Analysis

Beyond 5G and upcoming 6G communication technologies uses high frequency signals which has less penetration through obstacles. The use of Reconfigurable Intelligent Surfaces (RIS) is to act as an intermediate between the transmitter and the receiver end, by reflecting the signal from the transmitter to the destination with appropriate phase shifts. Optimization techniques are implemented in phase shift matrix as well as at the receiver end. Gradient Descent (GD) is an optimization algorithm which is frequently used in Deep Learning (DL) neural networks to decide the weightage of the nodes. This paper utilizes the Gradient Descent algorithm for the receiver side optimization in both non-RIS and RIS aided communication system and analyzed the effectiveness of using GD in these scenarios.