AI-DRIVEN DYNAMIC PILOT PLACEMENT FOR 5G MMWAVE MASSIVE MIMO: A RANDOM FOREST REGRESSION APPROACH

Efficient pilot placement in 5G millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems is critical to enhancing performance, achieving high spectral efficiency (SE), low bit error rate (BER), reduced pilot overhead, and minimized latency. However, this requires pilot symbols transmission, which occupies spectral resources and results in reducing spectral efficiency (SE). This paper proposes a novel dynamic pilot placement (DPP) framework, optimized using a Random Forest Regression (RFR) approach, to enhance system performance. Unlike traditional static and semi-static pilot allocation methods, the DPP approach dynamically adjusts pilot positions based on real-time channel state information (CSI) and system requirements, reducing interference and improving channel estimation accuracy. Through comprehensive simulations, the proposed method demonstrates significant enhancements in SE, BER, and latency, while maintaining pilot overhead at minimal levels. Furthermore, the latency reduction achieved through the optimized pilot allocation supports the stringent requirements of ultra-reliable low-latency communication (URLLC) scenarios in 5G networks. The results highlight the robustness of the RFR model in capturing the complex nonlinear relationships between pilot placement, channel conditions, and system performance metrics.