Optimization of Statistical Processing Algorithms for Wireless Communications in Dynamic Environments

This study investigates the performance of various channel estimation and signal detection techniques, including Kalman Filtering, Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs), with a focus on their application in 5G/6G networks. We evaluate these methods based on key metrics, including Bit Error Rate (BER), Mean Squared Error (MSE), and computational complexity, under different Signal-to-Noise Ratio conditions. Our results demonstrate that Deep Learning models (CNNs and RNN) significantly outperform traditional methods in terms of accuracy, achieving lower BER and MSE values. However, these improvements come at the cost of increased computational complexity, making them less feasible for real-time applications in resource-constrained environments. Reinforcement learning models also show promise, offering real-time adaptability for dynamic spectrum management and beam tracking but they also face challenges regarding computational efficiency. Despite some limitations, Kalman Filtering remains valuable for applications where low latency and computational efficiency are critical. Our findings highlight the importance of optimizing these models to balance accuracy and computational load for large-scale 5G/6G networks.