VISION-ASSISTED PROACTIVE HANDOVER AND RESOURCE ALLOCATION IN HYBRID LIFI/RF 6G NETWORKS USING EDGE AI

Ali Salah Mahdi Khafaga
Department of Computer Networks and Software, Al-Furat Al-Awsat Technical University, Babil, Iraq ali.khafaja@atu.edu.iq

DOI: 10.36724/2664-066X-2026-12-4-2-18

SYNCHROINFO JOURNAL. Volume 12, Number 4 (2026). P. 2-18.

Abstract

Light fidelity (LiFi) offers large unlicensed bandwidth for indoor 6G access, but its line-of-sight links are interrupted whenever a person walks between a luminaire and a device, or when the device tilts out of the receiver field of view. Handover schemes in hybrid LiFi/RF networks react only after the link is lost and therefore leave users without service for the detection and execution time of the handover. In this paper, we use the cameras that already monitor many smart buildings to anticipate these outages. An edge-AI pipeline tracks every person, converts the scene geometry into link-level features, fuses them with the measured signal-to-noise ratio, and predicts the probability that each LiFi link becomes unusable within the next second. A deep reinforcement learning agent, pre-trained from demonstrations and fine-tuned online, uses this forecast to select the access point of every user, while proportional-fair time sharing allocates resources inside each cell. In a room with six users and up to fifteen walking pedestrians, the learned vision-radio predictor raises the area under the ROC curve for 200-ms outage prediction from 0.79 (Kalman/Monte-Carlo forecast) to 0.89, and the proposed controller reduces link-interruption time by 42-47% with respect to a standard event-A3 handover, with no loss in mean throughput. Without retraining, the controller also reduces interruption by 53% when the pedestrians follow real trajectories from the ETH/UCY datasets.

Keywords:  LiFi, hybrid LiFi/WiFi networks, handover, blockage prediction, computer vision, edge AI, deep reinforcement learning, 6G

References

[1] H. Haas, L. Yin, Y. Wang, and C. Chen, “What is LiFi?” J. Lightw. Technol., vol. 34, no. 6, pp. 1533–1544, 2016, doi: 10.1109/JLT.2015.2510021.

[2] “IEEE standard for information technology–Telecommunications and information exchange between systems–Local and metropolitan area networks–Specific requirements–Part 11: Wireless LAN medium access control (MAC) and physical layer (PHY) specifications–Amendment 6: Light communications,” IEEE Std 802.11bb-2023, 2023.

[3] X. Wu, M. D. Soltani, L. Zhou, M. Safari, and H. Haas, “Hybrid LiFi and WiFi networks: A survey,” IEEE Commun. Surveys Tuts., vol. 23, no. 2, pp. 1398–1420, 2021, doi: 10.1109/COMST.2021.3058296.

[4] M. D. Soltani, A. A. Purwita, Z. Zeng, H. Haas, and M. Safari, “Modeling the random orientation of mobile devices: Measurement, analysis and LiFi use case,” IEEE Trans. Commun., vol. 67, no. 3, pp. 2157–2172, Mar. 2019.

[5] Y. Wang and H. Haas, “Dynamic load balancing with handover in hybrid Li-Fi and Wi-Fi networks,” J. Lightw. Technol., vol. 33, no. 22, pp. 4671–4682, 2015, doi: 10.1109/JLT.2015.2480969.

[6] D. Frómeta Fonseca, B. Genovés Guzmán, G. L. Martena, R. Bian, H. Haas, and D. Giustiniano, “Prediction-model-assisted reinforcement learning algorithm for handover decision-making in hybrid LiFi and WiFi networks,” J. Opt. Commun. Netw., vol. 16, no. 2, pp. 159–170, 2024, doi: 10.1364/JOCN.495234.

[7] S. Karimkhan and H. Shokrzadeh, “Reinforcement learning-based dueling DQN for access point selection and handover management in mobile hybrid LiFi/WiFi networks,” Wireless Pers. Commun., 2026, doi: 10.1007/s11277-026-12215-w.

[8] P. Hou and N. Cen, “Deep reinforcement learning-based user association in hybrid LiFi/WiFi indoor networks,” arXiv preprint arXiv:2503.01803, 2025.

[9] G. Charan, M. Alrabeiah, and A. Alkhateeb, “Vision-aided 6G wireless communications: Blockage prediction and proactive handoff,” IEEE Trans. Veh. Technol., vol. 70, no. 10, pp. 10193–10208, 2021, doi: 10.1109/TVT.2021.3104219.

[10] A. Alkhateeb, G. Charan, T. Osman, A. Hredzak, J. Morais, U. Demirhan, and N. Srinivas, “DeepSense 6G: A large-scale real-world multi-modal sensing and communication dataset,” IEEE Commun. Mag., vol. 61, no. 9, pp. 122-128, 2023, doi: 10.1109/MCOM.006.2200730.

[11] S. Karunasena, E. Khordad, T. Drummond, and R. Senanayake, “Deep learning-based computer vision for beam selection and proactive blockage prediction,” arXiv preprint arXiv:2605.04514, 2026.

[12] Y. Ahn, J. Kim, S. Kim, and B. Shim, “Computer vision-aided proactive mobility management for 6G terahertz communications,” in Proc. IEEE Global Commun. Conf. (GLOBECOM), 2023, pp. 1513-1518, doi: 10.1109/GLOBECOM54140.2023.10437835.

[13] S. Naser, O. Alhussein, and S. Muhaidat, “Towards multi-modal and cross-modal integration in LiFi-based sensing,” Sci. Rep., vol. 16, Art. no. 6038, 2026, doi: 10.1038/s41598-026-36891-7.

[14] S. Pellegrini, A. Ess, K. Schindler, and L. van Gool, “You’ll never walk alone: Modeling social behavior for multi-target tracking,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), 2009, pp. 261-268.

[15] A. Lerner, Y. Chrysanthou, and D. Lischinski, “Crowds by example,” Comput. Graph. Forum, vol. 26, no. 3, pp. 655-664, 2007.

[16] S. Chatterjee, D. Sabui, and B. Roy, “Mobility aware blockage management of a multi-user multi-cell hybrid Li-Fi Wi-Fi system with freeform diversity receiver,” Opt. Commun., vol. 560, Art. no. 130487, 2024, doi: 10.1016/j.optcom.2024.130487.

[17] J. M. Kahn and J. R. Barry, “Wireless infrared communications,” Proc. IEEE, vol. 85, no. 2, pp. 265-298, 1997.

[18] S. Dimitrov and H. Haas, “Information rate of OFDM-based optical wireless communication systems with nonlinear distortion,” J. Lightw. Technol., vol. 31, no. 6, pp. 918–929, 2013.

[19] V. Erceg et al., “TGn channel models,” IEEE 802.11 document 03/940r4, 2004.

[20] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 779-788.

[21] Y. Zhang et al., “ByteTrack: Multi-object tracking by associating every detection box,” in Proc. Eur. Conf. Comput. Vis. (ECCV), 2022, pp. 1-21.

[22] V. Mnih et al., “Human-level control through deep reinforcement learning,” Nature, vol. 518, pp. 529-533, 2015.

[23] H. van Hasselt, A. Guez, and D. Silver, “Deep reinforcement learning with double Q-learning,” in Proc. AAAI Conf. Artif. Intell., vol. 30, 2016, pp. 2094-2100.

[24] T. Hester et al., “Deep Q-learning from demonstrations,” in Proc. AAAI Conf. Artif. Intell., vol. 32, 2018, pp. 3223-3230.

[25] T. Salzmann, B. Ivanovic, P. Chakravarty, and M. Pavone, “Trajectron++: Dynamically-feasible trajectory forecasting with heterogeneous data,” in Proc. Eur. Conf. Comput. Vis. (ECCV), 2020, pp. 683-700.