Thesis defence V.S. Rao: internet of things
16 October 2017 12:30 - Location: Aula, TU Delft - By: Webredactie
Many novel IoT applications are fueling an exponential growth in the deployment of embedded devices in order to realize the IoT infrastructure. Wireless sensor networks (WSNs) are one of the key enabling technologies to build an IoT infrastructure.
Although IoT (or sensor) devices are required to last for a long time, batteries limit the lifetime of the devices, therefore that of the network and the applications. Powering all the IoT devices through batteries is neither scalable nor environmentally sustainable. Therefore, we adopt ambient energy-harvesting techniques and by tapping into the harvesting opportunities in the ambiance, the nodes gain autonomy with respect to energy. Unfortunately, merely replacing the batteries with energy harvesters does not provide the necessary alternative as the harvested power varies over space and time.
In this dissertation, we analyze and propose methods for power management for energy-harvested WSNs at various layers in the network stack, as communication takes the most amount of power on the sensor nodes. Firstly, we begin with discovering neighbors to form an energy-harvested WSN, wherein we analyze the effects of several parameters such as harvested energy, size of the storage buffer, and directional antenna on discovering neighbors. Next, after the discovery process, we look to construct well-connected topologies while achieving high energy-efficiency. We propose two localized algorithms, which select neighbors based on energy while keeping the global topology well-connected.
Constructive Interference (CI) has generated huge interest since protocols employing it achieve significantly low latency in data dissemination with supposedly high reliability and high energy efficiency, although there appears to be an inconsistent and contradicting picture of CI. With our extensive experimentation and analyses, we provide comprehensive insights into the CI phenomenon and further employ destructive interference to the performance of constructive interference, which is a major contribution of this thesis. Next, CI is often confused with packet capture, which is another well-known synchronous transmission techniques. Again with our extensive experiments, we explain the working of these two fundamental wireless phenomena in WSNs. Many IoT applications require guarantees on data collection, which is difficult on energy-harvesting WSNs due to the varying energy availability. We propose a distributed, energy-management module called GLEAM to gain the benefits of CI based protocols, particularly low latency, and high reliability, in an energy-harvesting WSN, which is a significant contribution of this thesis.
Context-awareness is important for IoT applications to provide user-centric applications. The last contribution of this work is to exploit the energy harvesters as transducers and to detect a change in the context, and thereby achieve high energy-efficiency as well as provide contextual information.
We intend to use ambient energy-harvesting techniques to completely replace the batteries in order to realize virtually immortal and sustainable IoT applications. We move several steps forward in this direction with our contributions by the means of energy-aware power management across the communication stack in the energy-harvesting WSNs.
For access to theses by the PhD students you can have a look in TU Delft Repository, the digital storage of publications of TU Delft. Theses will be available within a few weeks after the actual thesis defence.