As weather predictability grows increasingly unstable in the face of climate change, a sudden, unpredicted episode can have costly consequences to military action. One Colorado State University professor is attempting to curtail this threat by employing artificial intelligence technology to refine weather monitoring systems.
Sudeep Pasricha, a professor of electrical and computer engineering, received a $1.8 million grant from the United States Department of Air Force to develop enhanced weather monitoring systems through harnessing the potential of AI technology and the Internet of Things.
“The main goal of the project is to explore the use of nontraditional sensors such as terrestrial internet connected camera networks and resulting data to strengthen the overall weather intelligence capability at the DAF, making it more robust, resilient and adaptable to changing operational needs,” Pasricha wrote in an email-based interview.
As the current director of the Embedded, High Performance and Intelligence Computer Lab, Pasricha’s research focuses on the application and design of innovative software algorithms, including AI and machine-learning programs. Pasricha was named the first recipient of the Aram and Helga Budak ECE Professorship in 2024, a testament to both his research and teaching capabilities.
“Professor Pasricha is a globally recognized leader in technologies for a wide range of applications,” said Edwin Chong, head of the department of electrical and computer engineering. “His research on AI for weather monitoring has the potential to impact how we anticipate and adapt to threatening weather conditions.”
Current weather monitoring systems rely on measurements taken by polar orbiting satellites and geostationary satellites that are a part of the Defense Meteorological Satellite Program. Geostationary satellites orbit the Earth at a height of 22,200 miles and are capable of capturing images of the entire planet simultaneously. But their orbital distance can also cause an image of lower-resolution.
A lack of accurate weather monitoring can increase preventable risk faced by the DAF — a risk the Air Force described in a publication entitled “Weather Operations” in 2020 as, “Gaps in weather sensor coverage, limitations on the accuracy of weather observing systems and prediction models and the complexity of atmospheric processes can all reduce accuracy.”
Pasricha further explained the current weather-based changes DAF faces in daily operations.
“(DAF) faces significant challenges in managing diverse airfield, airspace and transportation scheduling operations due to the impact of unpredictable weather events, which can hurt mission success, aircraft performance and terrestrial transportation safety,” Pasricha said.
AI technologies directly counter this risk by greatly increasing DAF’s ability to measure real-time weather systems, both mitigating risks and potential limitations.
“AI algorithms have the ability to detect nascent patterns in data and make robust predictions based on these detected patterns,” Pasricha said. “Such algorithms will be crucial to predict weather changes in real time, by distilling useful information from gigabytes of daily data obtained from thousands of terrestrial sensing devices.”
The algorithm will be trained on data collected from devices associated with the Internet of Things: an interconnected ecosystem of devices, vehicles, appliances and other objects capable of collecting and exchanging stored data autonomously.
“By prototyping and deploying these algorithms in the cloud as well as on small (IoT) devices, we will bolster mission planning, situational awareness and operational effectiveness, ensuring that DAF commanders have the best possible environmental intelligence to support their decisions,” Pasricha said.
Pasricha is collaborating with FrostyFlake LLC, a company specializing in “hyperlocal, real-time alerts using cloud infrastructure and AI,” as stated on their website. Their inclusion will allow Pasricha to draw from their network of sensors deployed across over 20,000 locations.
Implementation of an AI-based software offers several advantages to DAF operations, including decreasing the gap in time between detection and reaction in the event of weather episodes.
Pasricha said the algorithm is projected to be designed and prototyped within 18 months, or at the end of 2026.
While development begins, the weight of the application’s growth is not lost on Pasricha. The hyper-local nature of the algorithm’s monitoring will also allow for advancements across departments while improving safety for active duty servicemen.
“This approach could potentially minimize the effects of weather on Air Force operations, saving lives, time and money while enhancing domain awareness for Joint Force Commanders across multiple geographic locations,” Pasricha said.
Reach Katie Fisher at science@collegian.com or on social media @RMCollegian.
