An AI-Driven Solution for Efficient 6G MIMO Networks Performance

The proposed study describes the architecture of an artificial intelligence solution known as GBT-AINet for sixth-generation networks. By combining Deep Neural Networks (DNN) and Gradient Boosting Decision Trees (GBDT), the model achieves a better balance between detection accuracy and computational efficiency. BGBT-AINet’s reduced-layer DNN architecture optimizes feature processing, while GBDT enhances feature extraction, enabling robust performance across various MIMO configurations (2×2, 4×4, and 8×8). The evaluations demonstrate GBT-AINet’s significant improvements over conventional methods such as Maximum Likelihood Estimation (MLE) and Minimum Mean Square Error (MMSE), achieving up to a 60% reduction in Symbol Error Rate (SER) under various SNR conditions. Furthermore, the model’s computational efficiency and scalability meet the requirements of high-dimensional 6G networks, guaranteeing flexibility in noisy and nonlinear signal settings. This work also demonstrates the reliability of GBT-AINet in real-time signal identification and anticipates a future for improvement, which is introduced in this paper. A future study could improve the training process, include recurrent patterns for temporal modeling in larger Multiple Input Multiple Output (MIMO) systems, and use advanced training techniques to enhance our detection of the transmitted signal.