Bounding end-to-end delays in energy harvesting wireless sensor networks

Wireless Sensor Networks (WSNs) are comprised of autonomous, self-organized sensors that collaborate to monitor surrounding conditions, such as temperature, motion, sound and vibration. The data collected by sensors are then forwarded to a base station or sink via multiple hops. A key challenge in WSNs is that sensor nodes are constrained by their physical size and battery capacity. As a result, they have limited energy, storage, processing speed, radio frequency (RF) bandwidth and transmission range. Consequently, the finite energy problem has attracted the attention of a number of researchers given that it directly determines the lifetime, and hence, usability of a WSN. In recent years, researchers have proposed to use energy harvesting technology to address the energy capacity constraint of sensor nodes. More specifically, sensor nodes are equipped with devices that are able to harvest energy from the ambient environment or human motion, and thereby, have the potential to operate perpetually. However, the energy-harvesting rate is time varying, and may not be sufficient to guarantee the required operating duty cycle. Hence, in order to ensure continuous operation, sensor nodes are required to maintain a low duty cycle. In a WSN, each node operates in accordance to a given duty cycle and may wake and sleep synchronously or asynchronously. In this thesis, we assume nodes operate using an asynchronous duty cycle due to its simplicity and low signaling overheads as nodes are not required to be synchronized globally. However,