Information theoretic multivariate change detection for multisensory information processing in Internet of Things

Internet of Things (IoT) is one of the main technological trends in the recent years. It allows machine-to-machine communication over the internet. Almost each device may transmit information from its sensors over the web to enable centralized insights derivation in an appropriate cloud architecture. In this paper we review analytical aspects of the sensory information processing. We emphasize the importance of multisensory approach, in which the joint distribution of all sensors values of a device is used to derive insights out of the stream of sensory data. We introduce a novel information theoretic multivariate change detection method based on k-nearest neighbor (kNN) estimation. The algorithm is designed and implemented to satisfy the requirements of IoT for fast online parallel multisensory information processing. We provide a numerical evidence of the validity of the proposed method on simulated and real world data.