Frequency Agile Wireless Sensor Networks: Design and Implementation

In this paper we propose an integration of dynamic channel allocation capabilities, typical of cognitive radio networks, in wireless sensor networks based on IPv6 protocol. We analyze the fundamental aspects of the proposed frequency-agile wireless sensor networks, including spectrum sensing, channel selection, network architecture, extensions of the IPv6 stack, and system operation. We then present a complete implementation of a prototype which illustrates feasibility and effectiveness of the proposed solutions. Results show that dynamic frequency allocation provides significant improvement of performance and reliability in wireless sensor networks operating in highly interfered environments.

I. Introduction

Networks of low-power short-range wireless devices, such as wireless sensor networks (WSN) [1], are used in a variety of applications, including environmental and industrial monitoring, surveillance, traffic control, healthcare, domotics, and military applications. In the future, it is envisaged that such devices will be communicating through the Internet protocol (IP), in particular IPv6 [2], participating in a wide-scale network of interconnected objects, called the “ Internet of things” (IoT) [3]. The increasing number and dense deployment of low-power wireless devices involves potential coexistence issues. Specifically, WSN devices are subject to interference for the following reasons. •

Most of current WSNs, e.g., based on IEEE 802.15.4 [4] and Zigbee [5], do not implement advanced frequency selection features. Therefore, if interference occurs on the operating channel, communication can be seriously degraded.

•

WSNs communicate in unlicensed bands, typically the 2.4 GHz ISM band, which is shared by other technologies, including IEEE 802.11 Wi-Fi [6] and Bluetooth [7].

•

WSN devices use relatively low transmit power, hence they are potentially vulnerable to several sources of interference. Wi-Fi, due to its high transmission power, represents today the main source of interference for WSNs.