Jamming-Resilient Handover Triggering for Programmable 6G Radio Access Networks using Reinforcement Learning

Resilient operation under abrupt Radio Frequency (RF) disruption is critical for next-generation cellular networks. Conventional static Event-trigger handover logic creates a fatal race condition under jamming: the handover command, triggered by slow-moving Reference Signal Received Power (RSRP) metrics, arrives too late to be decoded by a User Equipment (UE) whose Signal-to-Interference-plus-Noise Ratio (SINR) has already collapsed, resulting in radio-link failure. This paper proposes a 3rd Generation Partnership Project (3GPP)-conformant handover controller that instead of the fixed rule at the base-station mobility layer uses a Reinforcement Learning (RL) policy. The agent examines a reduced state vector, discretized SINR, serving-to-neighbour signal difference, and Hybrid Automatic Repeat reQuest (HARQ) error density, and outputs a binary trigger-or-defer action; no Radio Resource Control (RRC) or core-network signalling is changed. Realized in the LENA extension of ns-3 and tested in a multi-cell scenario with on-demand interference, the controller maintains connection continuity without extra signalling, computational overhead or ping-pong behaviour. Since its interface is restricted to vendor-agnostic Key Performance Indicators (KPIs) and a one-bit action, the mechanism can be encapsulated as an Open Radio Access Network (O-RAN) near-real-time xApp and migrated as-is to AI-native mobility functions anticipated in the future radio architectures.