Improvements to end-to-end performance of low-power wireless networks

Over the last decades, wireless technologies have become an important part of our daily lives. A plentitude of new types of networks based on wireless technologies have emerged, often replacing wired solutions. In this development, not only the number and the types of devices equipped with wireless transceivers have significantly increased, also the variety of wireless technologies has grown considerably. Moreover, Internet access for wireless devices has paved the way for a large variety of new private, business, and research applications. Great efforts have been made by the research community and the industry to develop standards, specifications, and communication protocols for networks of constrained devices, we refer to as Wireless Sensor Networks (WSNs). The Institute of Electrical and Electronics Engineers (IEEE) defined the 802.15.4 standard for Personal Area Networks (PANs). With the introduction of an adaptation layer which makes IEEE 802.15.4 networks IPv6-capable, interconnecting billions of constrained devices has become possible and is expected to become a reality in the near future. The vision that embraces the idea of interweaving Internet technology with any type of smart objects, such as wearable devices or sensors of a WSN, is called the Internet of Things (IoT). The main goal of this thesis is the improvement of the performance of low-power wireless networks. Given the wide scope of application scenarios and networking solutions proposed for such networks, the development and optimization of communication protocols for wireless low-power devices is a challenging task: The hardware restrictions of constrained devices, specific application scenarios that may vary from one network to another, and the integration of WSNs into the IoT require new approaches to the design and evaluation of communication protocols. To face these challenges and to find solutions for them, research needs to be carried out. Mechanisms and parameter settings of communication protocol stacks for WSNs that are crucial to the network performance need to be identified, optimized, and complemented by adding new ones. The first contribution of this thesis is the improvement of end-to-end performance for IEEE 802.15.4-based PANs, where default parameter settings of common communication protocols are analyzed and evaluated with regard to their impact on the network performance. Physical evaluations are carried out in a large testbed, addressing the important question of whether the default and allowed range settings defined for common communication protocols are efficient or whether alternative settings may yield a better performance. The second contribution of this thesis is the improvement of end-to-end performance for ZigBee wireless HA networks. ZigBee is an important standard for low-power wireless networks and the investigations carried out address the crucial lack of investigation the ZigBee HA performance evaluations through physical experiments and potential ways to improve the network performance based on these experiments. Eventually, this thesis focuses on the improvement of the congestion control (CC) mechanism applied by the Constrained Application Protocol (CoAP) used in IoT communications. For the handling of the possible congestion in the IoT produced by the plethora of the devices and/or link errors innate to low-power radio communications, the default CC mechanism it lacks an advanced CC algorithm. Given CoAP's high relevance for IoT communications, an advanced CC algorithm should be capable of adapting to these particularities of IoT communications. This thesis contributes to this topic with the design and optimization of the CoAP Advanced Congestion Control/Simple (CoCoA) protocol, an advanced CC mechanism for CoAP.The investigations of advanced CC mechanisms for CoAP involve extensive performance evaluations in simulated networks and physical experiments in real testbeds using different communication technologies.