Adaptive IIR Diffusion Networks for IoT Applications

Applications for Internet of Things (IoT) are rapidly growing, and in this work we propose infinite impulse response (IIR) adaptive networks to address such applications. We begin by addressing well known convergence limitations of standalone IIR adaptive filters (AFs), which may pose a major obstacle to their widespread use. For that, a new convex hybrid combination of a finite impulse response (FIR) and an IIR AFs based on Padé approximants mapping is introduced in order to accelerate the IIR AF convergence while preserving its steady-state performance. After addressing the standalone IIR AF applications, we propose a distributed estimation implementation, arriving at an IIR adaptive network that envisions IoT and sensor networks (SNs) applications. Simulations show that the proposed method is able to not only improve convergence for IIR standalone AFs, but also to motivate IIR adaptive networks, and that with much less computations, as compared to traditional FIR adaptive networks.