Mobility Management Paradigm Shift: from Reactive to Proactive Handover using AI/ML

Mobility management is one of the most essential functionalities in mobile networks, providing seamless services for users. Mobility performance has been one of the main focuses up to 5G. 3GPP introduced the conditional handover (CHO) in 5G to improve handover (HO) performance. CHO is a well-rounded technique that can solve the trade-off between HO failure (HOF) and ping-pong. However, it can incur a waste of radio resources due to several extra HO preparations. Additionally, achieving an optimal solution that balances the trade-off between ping-pong and user perceived throughput remains unsolved with the current reactive HO mechanism. In light of these challenges, this article proposes a proactive HO mechanism as a paradigm shift in mobility management for 6G networks. It utilizes measurement predictions to decide an optimal time and best target cell for HO. We employ time series forecasting using artificial intelligence and machine learning (AI/ML) for measurement predictions. We discuss and compare the UE-side model and the network-side model for measurement predictions. The proposed mechanism with the UE-side model realizes a proactive HO that improves mobility robustness and throughput gain. Through the simulation results, we demonstrate that our mechanism can achieve nearly zero-failure HO, solving the two above-mentioned trade-offs.