Dissertation Defense

Energy-Efficient and Robust Wireless Connectivity and Sensing Solutions for the Internet of Things

Milad Moosavifar

In recent years, as smart sensor nodes are being ubiquitously adopted in different environments and applications, there has been an exponential growth in the number of Internet of Things (IoT) sensor nodes which is expected to reach a 1-trillion-node milestone sometime in our lifetime. Wireless connectivity and sensing are the centerpieces to the promised massive IoT networks, which commonly experience strict energy restraints. It has been demonstrated that wireless communication and sensing are some of the major barriers in Ultra-Low-Power (ULP) Wireless Sensor Node (WSN) design due to their high power consumption. The rollout of modern wireless solutions has not been able to fully meet the requirements of the modern wireless era, leading to highly scalable wireless communication and sensing network infrastructure while maintaining the low-power and ULP regime. The small form factor and low power consumption requirement pose severe limitations on the performance of wireless systems. One of the main challenges is achieving high levels of interference tolerance in densely populated wireless networks, in which ULP receivers experience significant degradations. Second, non-integrated millimeter-wave (mm-wave) wireless systems encounter excessive losses due to their distributed nature and fail to preserve the miniature form factor.

The objectives of this dissertation are to analyze and address such challenges by proposing new system design techniques as well as circuit architectures to offer end-to-end energy-efficient wireless solutions for connectivity and sensing. Three prototypes of the proposed systems were implemented for evaluation in this thesis. The first prototype is a ULP interference-tolerant 433MHz receiver utilizing a novel Dual-Chirp On-Off-Keying (DC-OOK) modulation scheme that simultaneously enables low-power operation, in-band/out-of-band blocker rejection, and long communication range. The second work is a 900MHz low-power blocker-tolerant receiver with chirped OOK modulation that showcases a highly-selective receiver architecture leveraging a novel chirped N-path filter. This work showcases the feasibility of operating low-power radios in extremely congested spectrums with millions of connected WSNs, by leveraging chirped N-path filters. Finally, the last prototype is a 50mW PLL-less fully integrated 60GHz FMCW radar transceiver with on-chip antennas that leverages an open-loop VCO in conjunction with an energy-efficient chirp linearization scheme. Unlike conventional FMCW transceivers which employ power-hungry PLLs to generate linear chirps and detect target distance, this work benefits from a low-power and low-frequency ADC-assisted chirp linearization method, which is facilitated by TX/RX co-design, to compensate for chirp non-linearities with minimum power overhead.

Chair: Professor David Wentzloff