Wireless Sensor Networks (WSN) development leverages recent progress in electronic devices power consumption and in energy harvesting technologies in order to create smart sensing structures useful for improvements in various topics such as health monitoring or farming. Thanks to wireless communication circuits lower power consumption, it becomes possible to create networks of sensing systems capable of extracting information from the environment and of transmitting data through the network to the global intelligence. Because of hard and costly maintenance requirements, limited lifespans batteries are a brake on such networks development. Thanks to environmental energy harvesting on solar, thermal or mechanical sources, a system containing sensors and a wireless communication circuit can be powered. Global energy autonomy is thus improved and the node's life is enhanced. Works done during this PhD aim to study energy management within a sensing wireless communicating node. Thanks to the use of advanced multiple power paths architecture leveraging direct power path between the sources and the power loads, the power management system can optimize its energy efficiency when energy is harvested in the environment. Nevertheless, a precise digital control is mandatory to continuously determine the best power path between the energy harvesters, the energy storing capacitors and batteries, and the power loads. An integrated asynchronous controller implements an event-driven management of the power paths and gives the system robustness to environmental energy variations. After modeling and analyzing the power efficiency gain granted by the advanced architecture, an event-driven controller is proposed to ease implementation of wireless sensing applications. The controller is implemented in asynchronous quasi delay insensitive (QDI) logic and presents high intrinsic robustness to environemental variations while maintaining ultra low power consumption. A power management circuit suited for wireless sensing systems is thus fabricated using 180nm CMOS process and includes both architecture and digital management innovations. Its global power consumption close to 1µW allows considering the creation of wireless sensing nodes running for applications in the range of microwatts, consequently enabling development of ultra low power wireless sensor networks.
