Ultra Low Power Wireless Sensing

The future of personal well-being is all about continuous health monitoring through on-body devices (OBDs) like wearables, implantables, etc. and the consequent proactive health management that is made possible. The biggest bottleneck in realizing this vision is the challenging conditions (power, form factor, attenuation, inter-operability, etc.) under which these OBDs need to be deployed and operate, without sacrificing on their capabilities. In this thrust, we aim to remove several of the critical barriers in enabling future on-body healthcare applications. In particular, we leverage wireless backscattering to design novel, ultra low power (micro-W regime that enables maintenance-free operation) wireless tags that can significantly enhance (i) deployment: inter-operate and harvest energy from commodity wireless devices in the environment; (ii) communication: ability to send and receive high data rate signals over practical ranges with high sensitivity; and (iii) sensing: ability to house sensors, while also leveraging its own physical properties to record changes in the surrounding environment.

xSHIFT: Battery-less RF Tags Deployable with Commodity Devices

A novel, completely passive (battery-less) RF tag that works with commodity WiFi devices. One of the biggest bottlenecks in deploying battery-free wireless tags/sensors for on-body applications today is the need for dedicated, expensive readers to access and read these tags. In keeping the tags simple and low power, bulk of their processing and communication functionalities are offloaded to their reader, which makes the latter expensive (e.g. RFID readers to access the RFID tags). xSHIFT aims to change this by bringing the benefits of passive RFID tags to consumer and personal health applications at "low cost" by designing novel RF tags that can operate with our smartphones, WiFi access points, and other smart devices (e.g. smart speakers), i.e. without the need for a dedicated RFID-like reader infrastructure. This can open the door to a myriad of physical analytics applications in consumer spaces.

xShift's innovation is to accomplish frequency-shifted backscatter, whereby the backscattered signal is on a different frequency compared to the excitation signal, thereby allowing commodity (half-duplex) WiFi/BLE devices to receive the signal; but do it at a very low energy footprint on the tag, by leveraging the notion of an external local oscillator signal (instead of relying on on-tag oscillators) that can be seamlessly embedded into the excitation device, which is another commodity device. More details are available in the technical paper below.

• "Redefining Passive in Backscattering with Commodity Devices", ACM MobiCom 2020. (Online)

RIO: Turning RFID Tags into Passive Sensors for Tracking User Interaction

In addition to integrating various sensors onto RF tags, we are also interested in transforming the RF tags themselves into sensors. In this direction, RIO develops a novel primitive that leverages the physical properties of the RFID tag, namely the change of its antenna impedance (and hence backscattered signal phase) when interacted by the human body, to develop a smart, battery-free, touch-sensing user interface with commodity RFID tags. With RIO, any surface can be turned into a touch-aware surface by simply attaching RFID tags to them. RIO enables UIs to be constructed using off-the-shelf RFID readers and tags, and provides a unique approach to designing smart IoT surfaces. This can also serve as a communication cue for users and patients that are otherwise unable to communicate/signal verbally. More details are available in the technical paper below.

• "RIO: A Pervasive RFID-based Touch Gesture Interface", ACM MobiCom 2017. (Online)

MiXIQ: Novel Passive Wireless Receivers for On-body Applications

It is anticipated that on-body wireless applications are going to serve as the gateway for hyper-individualization in the future. They will go beyond today's wearables/hearables and cater to more personalized user experiences involving mixed reality, as well as personalized health management. One of the fundamental challenges facing these applications is the resource-constrained (form factor, energy) environment under which the associated wireless devices have to operate in. Particularly, in the context of healthcare, future wireless implants might need to operate without a battery altogether to eliminate the need for maintenance. Wireless (active) radios, which serve the important role of delivering information to/from these devices, unfortunately, form a major contributor to the energy footprint (milli-Watt) of these devices.

In this context, a long-standing challenge in passive radios has been to improve performance of a passive receiver both in terms of sensitivity and spectral efficiency, while retaining its low energy footprint. The most commonly used receiver design in low power applications is the envelope detector. While it consumes low power, it has long been known to have poor performance, but a viable alternative has been lacking. In this work, we take a fresh look towards addressing this challenge through the design of a new receiver architecture, called MIXIQ.

Departing from conventional practice, MIXIQ is built on two innovative principles: (i) a passive mixer that is driven by an "external" local-oscillator (LO) for down-conversion of the RF carrier to just sub-MHz without any energy cost; followed by an (ii) ultra low-power hybrid (analog/digital) baseband pipeline that leverages the down-converted signal's sub-MHz frequency to devise -- a high-gain, micro-power amplification, as well as a full-digital IQ demodulation -- that together contribute to significantly improved sensitivity and decode-ability of complex high-rate signals; all while operating in the micro-Watt power regime. In addition to delivering superior performance over envelope detectors, we have also built a simple Hearable system using it, to showcase its potential in future on-body applications.

• "MIXIQ: Rethinking Ultra-low Power Receiver Design for Next-generation On-body Applications", ACM MobiCom 2021.