A Self-Adaptive Spectrum Management Middleware for Wireless Sensor Networks

Abstract

The vision of the Internet of Things, wherein everyday objects are embedded with smart wireless sensor devices, is making these sensor devices increasingly pervasive. As the density of their deployment in overlapping or adjacent areas increases, the contention for the unlicensed 2.4GHz ISM band will also increase. To deal with the crowded spectrum, nodes must use the channels more judiciously and be able to adapt by detecting and switching to the most available channel. The SAS middleware that we have developed, is a self-adaptive spectrum management middleware for wireless sensor networks that enhances single-frequency MAC protocols with multi-frequency capability, without any change in hardware. It allows a single-frequency MAC protocol, like B-MAC, to automatically adapt to the least congested physical channel at runtime. SAS supports a combination of receiver-initiated and sender-initiated schemes to decide when to switch the channel and which channel to switch to. We have implemented the B-MAC protocol integrated with SAS in TinyOS 2.1 on TelosB sensor devices and evaluated its performance on the conditions of varied data flows and the interference produced by a jammer. The results demonstrate that the integrated B-MAC protocol outperforms B-MAC in terms of packet reception ratio, system throughput, average packet delay, and energy consumption.