AMHERST — A UMass professor will be giving new life to old tech, repurposing outdated smartphones to study the environment with the help of a $600,000 grant from the U.S. National Science Foundation.
Dr. V.P. Nguyen, an assistant professor at the University of Massachusetts Amherst’s Manning College of Information and Computer Sciences, will use the discarded devices to monitor plant water consumption and nutrient density in soil.
“On average, trees have about a 45% efficiency rate on receiving nutrients from the soil and water,” he said. “That is not ideal.”
Nguyen, initially from a biomedical background, discovered that many of smartphone tools used to monitor human health also could be used to track environmental health, a sort of “FitBit for trees.”
Initially started as a solar powered-drone project with an ecosystem-monitoring group, The PhenoCam Network, Nguyen hopes that using the recycled devices will solve a major problem in ecology: how to study large plots of conservation land without having massive blind spots in the resulting data.
While exploring the drones as a means to cover the distance, Nguyen stumbled upon an unorthodox and straightforward solution.
“Smartphone cameras are more powerful than (traditional monitoring) cameras,” he said.
Moreover, he noted that phones at their end of life are plentiful and cheap by comparison. Old phones can go for as little as $50 to $75, making the technology both scalable and sustainable.
Smartphones are pocket-sized but powerful, too, the professor said. With the average device having the processing power of a small computer, it’s possible to perform fast calculations on the fly.
But what about power sources? With electricity being hard to find or nonexistent in remote areas in the field and batteries being unsustainable, Nguyen and his team had to find a workaround.
“It is an interesting problem,” Nguyen said. “How to execute the maximum amount of goals based on the energy we have.” The solution? Renewable energy mixed with artificial intelligence.
Nguyen’s devices are both solar- and wind-powered, swapping out traditional batteries for capacitor arrays, which increase the speed and frequency of when a device can be charged. The array and the phone case can be biodegradable, too, making it environmentally sustainable in addition to efficient.
Further, the UMass team is employing a computing concept called energy-aware computing, in which AI systems assess the amount of remaining battery time and quickly perform a series of tasks based on importance to the project. This dynamically optimizes the amount — and hierarchy — of work done in the shortest amount of time.
The AI model, built in-house, uses already-existing technology to perform its tasks. Using a method called object recognition, the devices capture photos at regular intervals and compares them to images on record, highlighting changes based on metrics, such as relative greenness and velocity flow rate (or how fast water “goes up the trees”). Object recognition is a both energy-cost efficient, as well as an easy task, for AI to perform.
Nguyen also is exploring cloud networks to offload some of the more energy-intensive tasks.
With partners — The PhenoCam Network and GaugeCam — Nguyen intends to integrate his smartphone system into their existing networks by next summer, which includes about 20 sites nodes. From there, he aims to set up 20 more developing sites.
“The potential is very high,” Nguyen said. “The technology is already out there, and we should make use of it.”