Memory-Full Context-Aware Predictive Mobility Management in Dual Connectivity 5G Networks

Network densification with small cell deployment is being considered as one of the dominant themes in the fifth generation (5G) cellular system. Despite the capacity gains, such deployment scenarios raise several challenges from mobility management perspective. The small cell size, which implies a small cell residence time, will increase the handover (HO) rate dramatically. Consequently, the HO latency will become a critical consideration in the 5G era. The latter requires an intelligent, fast, and light-weight HO procedure with minimal signaling overhead. In this direction, we propose a memory-full context-aware HO scheme with mobility prediction to achieve the aforementioned objectives. We consider a dual connectivity radio access network architecture with logical separation between control and data planes because it offers relaxed constraints in implementing the predictive approaches. The proposed scheme predicts future HO events along with the expected HO time by combining radio frequency performance to physical proximity along with the user context in terms of speed, direction, and HO history. To minimize the processing and the storage requirements whilst improving the prediction performance, a user-specific prediction triggering threshold is proposed. The prediction outcome is utilized to perform advance HO signaling whilst suspending the periodic transmission of measurement reports. Analytical and simulation results show that the proposed scheme provides promising gains over the conventional approach.