A Massive MIMO Framework for Power-Constrained Devices Without Error-Correcting Memory

Error correcting codes (ECCs) are frequently avoided by power-limited devices due to hardware complexity and financial limitations. During transmission, memory errors produced by these devices may spread over the wireless channel, making decoding more difficult at the receiver. It can be a crucial problem in a large MIMO system. We address this challenge in this article by proposing a two-stage solution. The first stage employs an equalizer-based detection method to lessen interference and communication channel impairments, while the second stage uses gradient boosted decision tree (GBDT) and neural network (NN) algorithms to forecast and fix memory errors. Using the GBDT and NN model through the Scikit-Learn library, we used Python simulations to show the effectiveness of our method. Following approximate inversion-based equalization and NN-based error detection at the massive MIMO base station receiver, the results show a detection accuracy of over 99%. Finally, we present a motivating example demonstrating how our proposed model can be integrated into the base station.