Beam Prediction based on Large Language Models
This work addresses beam prediction for wireless communication systems, representing an incremental improvement by applying LLMs to a domain-specific task.
The paper tackles the problem of millimeter wave beam prediction by formulating it as a time series forecasting task and using large language models (LLMs) with a prompt-as-prefix technique, resulting in outperforming traditional learning-based models in accuracy and robustness.
In this letter, we use large language models (LLMs) to develop a high-performing and robust beam prediction method. We formulate the millimeter wave (mmWave) beam prediction problem as a time series forecasting task, where the historical observations are aggregated through cross-variable attention and then transformed into text-based representations using a trainable tokenizer. By leveraging the prompt-as-prefix (PaP) technique for contextual enrichment, our method harnesses the power of LLMs to predict future optimal beams. Simulation results demonstrate that our LLM-based approach outperforms traditional learning-based models in prediction accuracy as well as robustness, highlighting the significant potential of LLMs in enhancing wireless communication systems.