Idaho researchers use AI to give drones new roles in industryPosted to Idaho National Laboratory in the Digital Utility Group
image credit: © Thekaikoro | Dreamstime.com
- May 12, 2021 4:35 pm GMTMay 11, 2021 9:06 pm GMT
- 527 views
This item is part of the Special Issue - 2021-04 - Innovation in the Power Industry, click here for more
Every day across the U.S., companies spend precious time and money sending employees to perform routine tasks such as taking inventory in a warehouse or monitoring equipment in a power plant.
In the past decade, drone technology has advanced to the point that an off-the-shelf, camera-equipped quad copter could prove a good option for performing some of these tasks. But before companies can realize the full potential of drones, one big challenge remains: How can a drone navigate through tight spaces in industrial settings quickly and efficiently without crashing?
Now, researchers with the U.S. Department of Energy Light Water Reactor Sustainability program have developed a solution. Route Operable Unmanned Navigation of Drones (ROUNDS) uses artificial intelligence (AI), advanced image analysis and modern control techniques to enable an off-the-shelf, camera-equipped drone to accurately find its way, at high speed, in an indoor environment where GPS is not available.
What ROUNDS can do
ROUNDS opens the door for inexpensive drones to help a wide range of industries save time, save money and increase efficiency while reducing human exposure to heights, chemicals, radiation and other hazards. For example, with ROUNDS, drones could perform security patrols in place of or in conjunction with human security guards (at loading docks, warehouses, buildings, etc.).
ROUNDS-equipped drones can collect readings from nonnetworked gauges and instruments that otherwise would have to be captured manually. They can conduct visual inspections to identify abnormal and dangerous situations, such as fires. Drones can also check inventory in a warehouse, which is currently an expensive and time-consuming process.
“ROUNDS is especially useful when the tasks are frequent, tedious or hazardous,” said Ahmad Al Rashdan, a senior research and development scientist at Idaho National Laboratory. “Instead of requiring a human to perform the same task over and over again, the operator can use ROUNDS to program the drone to perform the same task at preset times without human assistance.”
How ROUNDS works
ROUNDS reimagines the way off-the-shelf drones navigate. First the drone operator places visual cues such as QR codes along a route for the drone to follow. ROUNDS then exploits the imaging and communications capabilities of the drone to follow those visual cues, like following breadcrumbs through a forest.
ROUNDS uses several novel techniques, including AI, to accurately locate itself within its environment by analyzing the relative shape of the QR codes it is following. Like the human eye, the drone’s camera then perceives changes in the shape of an object as it is viewed from different angles.
ROUNDS uses artificial intelligence to recognize QR codes from multiple angles and different backgrounds. ROUNDS measures the angles and dimensions of the QR code and extrapolates its relative location, determining its location in the environment within inches of its true position. The QR code is processed, and the location is determined in less than a hundred milliseconds using a normal computer. Every time the drone calculates its relative location, high performance controllers, typically used for rockets and satellite applications, update the drone’s trajectory.
When the next QR code becomes visible, ROUNDS identifies it and again establishes the drone’s location, feeding that information into the controller to continuously adjust trajectory. This process repeats until the drone has completed its mission.
Each QR code contains identifying information about more than its location. Commands within the code also guide the drone to perform tasks at a particular location, such as taking a picture of a gauge.
“ROUNDS makes it easy for the drone operator to change these locations and tasks by simply printing out a new QR code. The software does not need to be reconfigured, making it very user-friendly,” Al Rashdan said.
Why ROUNDS is better
ROUNDS holds any number of advantages over competing systems, especially its cost and user-friendly operation. In the end, however, the most important advantage is that ROUNDS can reliably perform high-speed, accurate indoor navigation at very low cost.
The primary reason that off-the-shelf drone systems cannot reliably navigate indoors is because they lack the ability to accurately locate their own position. Compared with ROUNDS, which uses AI and QR codes to find its location to within a few inches, the location accuracy of low-cost consumer drones is typically several feet.
“Using ROUNDS, the image analysis and control requires roughly tens of milliseconds,” Al Rashdan said. “This enables a drone equipped with ROUNDS to navigate quickly, covering much more distance and reaching far more locations before it returns to the charging pad.”
ROUNDS holds another big benefit over competing technologies: It’s easy to modify off-the-shelf drones to use the technology. Retrofitting a standard commercial drone to use mapping, lidar or Wi-Fi triangulation typically requires expensive equipment and expertise to implement. ROUNDS does not require any modifications to the drone itself, nor does it require software configuration.
An expanded role for drones
ROUNDS achieves the speed, precision and accuracy necessary for inexpensive drones to play an expanded role in industrial settings.
“The technology overcomes the limitations of modern commercial drone navigation systems and makes drone technology useful in dozens of new venues – such as plants, security and warehouses – where GPS is not available,” Al Rashdan said. “ROUNDS is user-friendly and performs these tasks without the expensive hardware and software currently used by other, more expensive drone navigation systems.”