UCSD - University of California - San Diego

11/05/2024 | Press release | Distributed by Public on 11/05/2024 04:14

UC San Diego Researchers Develop Low-Cost, Scalable Passive Sensors

Published Date

November 05, 2024

Article Content

Data is power.

According to Dinesh Bharadia, an associate professor at UC San Diego in the Department of Electrical and Computer Engineering with an affiliate appointment in the Department of Computer Science and Engineering and the Qualcomm Institute (QI), "data will be the next decade's 'silicon.'"

The rapid growth of the Internet of Thingsmeans that data is more readily available and easily accessible than ever. Sensors, "smart" devices and software connect our world to the cloud, gathering information and enabling new types of data sharing and analysis. However, most of these tools are battery-powered and have difficulty sensing changes in real time.

Now, the tide is turning.

New research presented today to be published in the Proceedings of the ACM Conference on Embedded Networked Sensor Systems from Bharadia and lead author Nagarjun Bhat, demonstrates that not only is passive sensing - or sensing without being connected to a power source - possible, it can be done at little cost without any specialized equipment.

Transforming commodities into tools

A Ph.D. student in electrical engineering, Bhat's research focuses on ways to enable passive sensing using simple, widespread commodities.

His commodity of choice? Radio Frequency Identification tags, also known as RFID tags. Essentially, these small, flexible tags receive and transmit data from a chip to an RFID reader, which processes the information and sends it to a computer program for interpretation. They're commonly embedded in products like clothing or library books for tracking inventory or in contactless transit fare payment cards.

Though they sound highly technical - and thus expensive - RFID tags run between a few cents to a few dollars per chipdepending on the specs. And, with up to 90% of retailers using RFID technology, the chips are widespread and easy to access.

To Bhat and Bharadia, who is also a faculty member of the UC San Diego Center for Wireless Communications, these chips appeared prime candidates for further experimentation.

"We wondered whether we could repurpose RFID tags to do battery-free sensing and tracking," said Bhat. He explained that most current approaches to passive sensing rely on analog-digital converters, which measure stimuli, record them in raw data and convert them to digital values that are readable by computers.

But, these types of sensor interfaces are power-hungry; without additional batteries, they can last a matter of hours. Battery-based systems are also bulky, expensive and hard to scale sustainably.

"We were trying to see if we could use the chips to directly sense stimuli without needing converters," he added. "We wanted to know if our environment could be automated in a way that was battery-free, able to sense parameters like temperature and humidity, and could connect to the Internet of Things to send raw data to a reader that could make sense of it all."

Real-time data through RFID tags

Bhat and Bharadia aren't the first ones to attempt making passive, wireless interfaces. Other researchers have pursued ultra low-power digital sensing that couples a sensor, converter and microprocessor into a single package. While an efficient design, these types of devices are expensive, bulky, and lack the ability to sense and report stimuli in real time. They only send data to a reader when it's requested and need complex electronic circuitry for their interface.

One of Bhat and Bharadia's sensors feeding data to a computer

"If I wanted to use digital sensing for a biomedical application like monitoring a patient's heart rate, I might not be able to access that data for 10 minutes," Bhat said. "That's a problem."

Analog sensing - the category in which Bhat and Bharadia's sensors fall - directly perceive environmental stimuli. Unlike digital interfaces, analog ones convert the change in voltage/current produced by sensors into parameters of a wireless signal.

Bhat noted that although "there has been good work done so far" on passive analog systems, most of the research has relied on customized sensors that are purpose-built and only suited for a particular application. These systems are difficult to generalize, he explained, adding that "you'd have to redesign all sensors on the market to make them commercially available."

That's why he chose RFID tags as the workhorse of his passive sensors: they're commercialized, cheap and require little custom hardware to be deployed or read.

"We took the concept of analog sensing and made it real-time," Bhat said. "You don't need any fancy interfaces, specialized readers or batteries to access the data - all you need are some commercially available RFID tags, antennas and readers."

The future of data collection

Bhat's battery-free RFID sensors enable new use cases like improved agricultural management, real-time athletic performance metrics and occupancy detection.

Currently, automatic irrigation systems generally rely on a smaller quantity of bigger sensors that cover large areas. This can be cost-effective, although it comes at the expense of data specificity. RFID-based passive trackers can do both. By deploying soil moisture sensors at scale around a field, it's possible to use a few RFID readers to remotely measure moisture content at a much more granular level and adjust how water is distributed based on current conditions.

This type of immediate data can also be valuable for athletes. For instance, many UC San Diego athletes engage in force plate testing as part of their training, where they jump on force plates that measure their strength, power and posture. These tests must be done at a special facility and can be expensive. Bhat's paper describes how RFID sensor tags could be used to bring these tests "in-house" by embedding them in shoe soles to measure an athlete's jumping force.

Or, RFID tags can be placed in parking garages to measure occupancy and map where and how many spaces are being used. A chip could be added to the floor of every space; when a car pulls into the spot and covers the light-sensitive sensor, the tag recognizes that the spot is occupied and can send that information to a central location.

At the end of the day, however, Bharadia and Bhat see bigger uses for their work.

"AI is everywhere now," Bharadia said, adding that AI is powered by data enabled by sensors. "We're at the cusp of a revolution where new sensors will be collecting the data that will power the next generation of AI. Using batteryless sensors lets us collect a lot of information that's otherwise challenging to access - they can empower data collection, and this innovation marks a really important direction for the future."

Bhat and Bharadia presented their research on November 5 at the 22nd ACM Conference on Embedded Networked Sensor Systems (SenSys 2024) in Hangzhou, China.

Learn more about research and education at UC San Diego in: Artificial Intelligence