A Directed Information Learning Framework for Event-Driven M2M Traffic Prediction

Abstract

Burst of transmissions stemming from event-driven traffic in machine type communication (MTC) can lead to congestion of random access resources, packet collisions, and long delays. In this paper, a directed information (DI) learning framework is proposed to predict the source traffic in event-driven MTC. By capturing the history of transmissions during past events by a sequence of binary random variables, the DI between different machine type devices (MTDs) is calculated and used for predicting the set of possible MTDs that are likely to report an event. Analytical and simulation results show that the proposed DI learning method can reveal the correlation between transmission from different MTDs that report the same event, and the order in which they transmit their data. The proposed algorithm and the presented results show that the DI can be used to implement effective predictive resource allocation for event-driven MTC.