Owl Platform comes with two different types of sensors, one for light and one for water, as well as a receiver that connects to computers via USB or Ethernet, said Martin, an associate professor in the Department of Computer Science.
“This is a low-power wireless research project worked on across various generations of students,” he said. “Every generation, the boards get better.”
Moore, a graduate student in the Department of Computer Science, tests the sensors in a 24×24 grid that broadcasts a signal about five times per second.
Instead of having a sensor passively wait for signals, the team designed the sensors to broadcast their own signal, he said. Any object passing between the sensor and receiver weakens the radio signal, providing data that could be used to locate the object.
“As I walk through the grid, the salt water of my body interacts with the radio signals and starts to block them,” Moore said. “The weakening of these radio signals are picked up by the receiver and you can find my location from this interference.”
Saving power over a wireless network requires carefully controlled information trafficking, Martin said. Wireless communication protocols such as Wi-Fi or Bluetooth are generalized, resulting in large data traffic that can be simplified for low-power needs.
The team developed their own protocol that uses radio waves, Moore said, named “roll call.”
“‘Roll call’ is a reason we have such good lifetime in our chips,” he said. “Right now, these are the lowest energy wireless sensors in the world.”
It was designed originally for inventory management, using a small amount of energy to track many devices, he said.
Bernhard Firner, a Ph.D. student in Electrical and Computers Engineering, worked on much of the math behind “roll call” and Owl Platform’s data prediction.
Data prediction is necessary because large networks of sensors could result in data loss from communication collisions to the receivers, he said. He created accurate models to prevent such data loss.
“If you told me what you were tracking and in what space, I could tell you how many receivers you would have to put and where,” he said.
Moore says his responsibilities involve the software end of the project — he develops the user interface and activity of the sensors.
Owl Platform offers a notification system that can be tailored to anyone’s needs, he said. The notification system includes emails, texts and tweets.
Undergraduate students also contributed, he said. One student developed a system that would turn off a heater when someone leaves a specific chair absent for more than 50 minutes.
“I made a sort of Marauder’s Map-style application to track objects, too,” he said. “During the summer we get a lot of people, so we also used sensors to send alerts whenever someone tried to take a chair.”
The data solvers and USB drivers of Owl Platform is open source, offering freedom for tech-savvy people as well, he said. They hope to support both the home users and the application developers.
Howard is working on upgrades to Owl Platform, including soil moisture detection, humidity detection and improvements to the water sensor.
He said the benefit of Owl Platform is that it does not operate from a rigidly defined set of applications — it is a general-purpose platform fit for any user.
“We had two students who made the data solver software for the water sensor in a day,” Howard said. “It’s a pretty straightforward platform to use.”
The Nelson Biological Laboratories on Busch campus also use the sensors for their benefit, Moore said, to track temperature and light levels in the animal rooms.
Notifications are sent to the researchers in Nelson Biological Laboratories if the temperature is out of range for too long or if the day cycles of the lighting are incorrect.
“The goal is you buy the thing, set it up, and you can forget about it for 10 years,” Firner said. “If someone had a vacation home they wanted to monitor and make sure their house was OK, that can be done quickly and easily.”
They want to be able to make a system that is easily accessible to anybody and does not require professional or troublesome installment and maintenance, he said.
“This can help many types of people — some people can check if their houses are beginning to flood,” Martin said. “One woman we gave the kit to put a sensor in her liquor cabinet to see if her kids were opening it.”
According to the Owl Platform website, the open source information is available on GitHub. Owl Platform will also soon to have its own Kickstarter campaign to finalize development and commercialize the product.