Design and Testing of Mixed-Range Low-Power Wireless Networks

Abstract

Nowadays, more and more objects are being connected to the Internet as the resulting Internet of Things (IoT) offers a wide range of opportunities. Low-power wireless communication technology is frequently used to connect objects to the Internet because it has the advantage that no infrastructure is required at the object location. Until recently, low-power wireless technology was limited to short distances. However, in the last few years, advances in transceiver technology have enabled low-power communication over long distances which enables connecting objects wirelessly at a city scale. A fundamental difficulty of the new technology is that transmissions suitable for different distances have significantly different characteristics. Low-power long-distance transmission schemes typically offer only low data rates, resulting in relatively long transmission times and significantly higher energy consumption compared to short-distance transmissions. This makes exchanging or extending short-range with long-range communication technology difficult. In this thesis, we focus on the design of reliable and energy-efficient schemes that provide high coverage in scenarios with a mix of long- and short-range wireless connections. First, we present infrastructure and tools for the investigation of such schemes. Then, we extend existing and design novel schemes for low-power mixed-range application scenarios. More concretely, we make the following contributions in this thesis: * We present a platform design for wireless IoT devices that support multiple modulation schemes and therefore offer low-power short and long-range communication capabilities. The use of clear abstractions and a well-defined interface enables modular integration, which allows the platform to be used for prototyping as well as for real-world deployments. A detailed analysis of relevant timing overheads form the basis for the implementation of timing-critical communication schemes. * We introduce a testbed architecture for testing and validation of mixed-range communication protocol implementations with large link distances between testbed nodes. In addition, the testbed architecture allows simultaneous high-quality tracing of different aspects including timing, power consumption, and functional correctness. Moreover, it provides support for remote accessing the on-chip debug and trace sub-system that is built-in in modern microcontrollers. * Based on the introduced testbed architecture, we propose a methodology for non-intrusive and instrumentation-free tracing of the program execution on all testbed nodes simultaneously. In contrast to commonly used approaches, our methodology allows tracing with high temporal resolution and without requiring interactions with the experimenter. Furthermore, our proposed method includes a time synchronization scheme allowing to align the traces obtained from the distributed testbed nodes on a common time axis. * We propose two communication schemes based on time-division multiple access (TDMA) channel access that are compatible with the widely used Long Range Wide Area Network (LoRaWAN) communication protocol. With an analysis that takes into account legally enforced duty-cycle limits, we show that our proposed schemes increase reliability and energy-efficiency for a wide range of application scenarios when compared to the default random channel access scheme of LoRaWAN. * We present a protocol that makes use of synchronous transmission floods with multiple modulation settings to enable reliable and high coverage multi-hop communication in networks with inhomogeneous link characteristics. We propose a scheme based on non-destructive probing within floods that allows to collect connectivity information of the network in parallel to data transmissions without adding significant overhead. The obtained connectivity information enables a wide range of optimizations. In an extensive evaluation, we show how the use of multiple modulations and applying the proposed optimizations can reduce the power consumption significantly for certain application scenarios compared to a state-of-the-art protocol using a single modulation and no optimizations.