Performance Analysis of Channel Estimation Schemes for Phase Shift Optimization: An Analysis of Bit Error Rate and Armijo Step Size Behavior

In the recent past, there has been a growing need for ultra-low latency and high-data-rate communication. In Non-Line-of-Sight (NLoS) communication, the channel capacity and accuracy of transmission are significantly affected by interferences, lowering the Quality of Service (QoS). An intelligent Reflecting Surface (IRS) has risen as a potential solution to challenges associated with NLOS communication including low data rate, multipath fading, and high BER. However, to leverage the performance gains of the IRS, effective and highly accurate channel estimation is crucial as it facilitates optimal phase shift optimization. This work investigated the performance of four main channel estimation algorithms in an IRS-aided system; LS, DD, DFT, and MMSE in terms of their BERs and effects on the convergence behavior of the Stochastic Convex Approximation (SCA) algorithm following the Armijo rule. Results indicate that in cases without statistical knowledge of the channel, the DD method provides the best performance. This work shows that the communication needs, complexity, and accuracy should be carefully considered when selecting the channel estimation method for IRS-aided communication systems.