A Semi-Supervised Location-Aware Anomaly Detection Method for Ultra-Dense Indoor Scenarios.

Abstract

Over the past few years, indoor cellular deployments have been on the rise. These scenarios are characterized by their user density and fast-changing conditions, thus, being prone to failures. Moreover, the steady development of indoor and outdoor positioning techniques is expected to provide a reliable source of information. Thus, the availability of user location is being considered to be a key enabler to improve the resilience and performance of automatic failure management and optimization techniques. Taking this into consideration, the present work proposes a semi-supervised location-aware anomaly detection method for the management of failures such as cell outages and interference problems.