Wireless Sensor Networks (WSNs) for Critical Applications: analysis and enhancements

In this thesis we study the suitability of Wireless Sensor Networks (WSNs) for suppporting critical applications, i.e. applications where, in addition to energy effciency, other requirements such as reliability, timeliness, security and scalability must be taken into consideration. We organize the thesis into three main parts. First, we consider WSNs compliant to the IEEE 802.15.4 standard that is currently the most popular technology for commercial WSNs. We show that 802.15.4 WSNs suffer from severe limitations in terms of reliability and scalability that make them unsuitable for critical applications. Since these limitations are mainly due to an improper setting of the CSMA/CA algorithm used to regulate channel access in 802.15.4 networks, we propose JIT-LEAP, an adaptive and learning-based algorithm that adaptively tunes the 802.15.4 CSMA/CA parameters so as to provide the level of reliability required by the application with the minimum energy consumption. JIT-LEAP makes 802.15.4 networks suitable for applications having reliability as the main concern. However, it is not a viable solution for time-critical applications since the increase in reliability comes at the cost of increased packet latency. When low and predictable latencies are required, a Time Division Multiple Access (TDMA) scheme is typically used for channel access. Hence, the second part of this thesis focuses on TDMA-based WSNs. Despite TDMA provides guaranteed bandwidth, high energy efficiency, low and predictable latency, it also has a number of limitations (e.g. it requires a strict synchronization between sensor nodes, an allocation of slots to sensor nodes and, suffers from selective jamming attack). Thus, we provide original solutions to overcome some of them. We first propose LOCALL, a localized algorithm for allocation of transmission slots to sensor nodes. Thanks to its localized approach, LOCALL does not require the exchange of messages to establish a communication schedule and, hence, minimizes energy consumption of sensor nodes. Then, we propose JAMMY, a distributed solution to contrast the selective jamming attack in TDMA-based WSNs. The solutions we propose in the first two parts of the thesis significantly improve the performance/robustness of both 802.15.4 and TDMA based networks. However, the fact that these technologies rely on one single channel for communication can signifcanlty degrade their performance in real-world scenarios. WSNs share the wireless spectrum with other ambient technologies such as WiFi, Bluetooth and, hence, suffer from external interference. Moreover, the performance of WSNs is usually affected by multi-path fading since any object in their surroundings acts as a reflector for RF signals. Using multiple channels for communication and a channel hopping scheme, has been shown to be an effective way to mitigate both external interference and multi-path fading. For this reason, IEEE has recently proposed the IEEE 802.15.4e Time Slotted Channel Hopping (TSCH) mode, a new channel access mechanism that combines time slotted access, with channel hopping capabilities. Given these characteristics TSCH is one of the most promising technologies to support real-world WSN critical applications. Therefore, the last part of the thesis is devoted to analyze it.