Autoencoder for Position-Assisted Beam Prediction in mmWave ISAC Systems

Integrated sensing and communication and millimeter wave (mmWave) have emerged as pivotal technologies for 6G networks. However, t he narrow nature of mmWave beams requires precise alignments that typically necessitate large training overhead. This overhead can be reduced by incorpor ating the position information with beam adjustments. This letter propos es a lightweight autorencoder (LAE) model that addresses the position-assi sted beam prediction problem while significantly reducing computational co mplexity compared to the conventional baseline method, i.e., deep fully conne cted neural network. The proposed LAE is designed as a three-layer undercomplete network to exploit its dimensionality reduction capabilities and t hereby mitigate the computational requirements of the trained model. Simulati on results show that the proposed model achieves a similar beam prediction a ccuracy to the baseline with an 83% complexity reduction. This work was supported in part by Natural Sciences and Engin eering Research Council of Canada (NSERC), Discovery program RGPIN-2019-04123 and Canada Re search Chair program CRC-2022-00187.