Ultra Low-power Wireless Sensor Node Design for ECG Sensing Applications

Ubiquitous computing, such as smart homes, smart cars, and smart grid, connects our world closely so that we can easily access to the world through such virtual infrastructural systems. The ultimate vision of this is Internet of Things (IoT) through which intelligent monitoring and management is feasible via networked sensors and actuators. In this system, devices transmit sensed information, and execute instructions distributed via sensor networks. A wireless sensor network (WSN) is such a network where many sensor nodes are interconnected such that a sensor node can transmit information via its adjacent sensor nodes when physical phenomenon is detected. Accordingly, the information can be delivered to the final destination through this process. The concept of WSN is also applicable to biomedical applications, especially ECG sensing applications, in a form of a sensor network, so-called body sensor network (BSN) where affixed or implanted bio-signal sensors gather bio-signals and transmit them to medical providers. The main challenge of BSN is energy constraint since implanted sensor nodes cannot be replaced easily, so they should prolong with a limited amount of battery energy or by energy harvesting. Thus, we will discuss several power saving techniques in this thesis.

The main low power techniques are low voltage operation and duty-cycling. The design challenges for the former are reliability and signal timing since performance is deteriorated as the supply voltage is scaled down. The challenge of the latter is that it is effective only to the blocks which are power hungry and seldom used. Thus, combination of these two techniques are presented to maximize the power saving for the proposed system on chip (SoC) where low-voltage operation was applied to always-on blocks, while a power hungry RF transceiver was duty-cycled.