Reliability Enhancement in 5G/6G Millimeter Wave and Terahertz Cellular Systems with a LSTM Model

Enhancing reliability in 5G/6G millimeter wave (mmWave) and Terahertz (THz) Cellular Systems is a pressing concern due to these bands' vulnerability to path loss, blockage, and interference. As these technologies are slated to support critical applications like autonomous vehicles and telemedicine, achieving ultra-reliable and low-latency communication becomes imperative. Any shortfall in reliability could lead to severe repercussions, emphasizing the need for robust solutions to fully realize the capabilities of next-generation networks. To address this, this research introduces the Long Short-Term Memory based Reliability Enhancement (LSTM-RE) model for mmWave and THz networks. Unlike traditional methods that falter in dynamic conditions, the LSTM-RE model utilizes LSTM networks to capture long-term dependencies, providing a more adaptive and resilient approach to reliability enhancement. Further, it employs the Multi-Head Attention Mechanism (MHAM) to emphasize the important time steps in the input sequence. The model's performance metrics, framed as functions of latency, throughput, and packet loss, validate its efficacy. In larger 6G network configurations, the model achieves a low latency of 1.1 ms and a high throughput of 100 Mbps, positioning the LSTM-RE as a promising solution to enhance the performance of next-generation networks.