UE-Centric Inter-Cell Interference Mitigation: An Anomaly Detection Approach

Inter-cell interference (ICI) from neighboring cells can substantially degrade performance as spectrum reuse increases. In this context, designing effective interference mitigation techniques is essential for boosting overall system performance. This paper introduces a user equipment (UE)-centric interference mitigation framework that efficiently detects ICI using an anomaly detection method and then suppresses the detected ICI through interference whitening (IW). Simulation results show that the proposed technique, termed Z-refined deep support vector data description (ZRD-SVDD), consistently achieves higher accuracy in interference detection compared to baselines, even with restricted training resources in terms of time or frequency. In addition, the ZRD-SVDD-based IW technique nearly matches an ideal genie-aided interference mitigation technique.