08-04-2025
Predicting The Unpredictable: AI's New Role In Aviation Safety
Just last week, a small plane crashed in Tuscarawas County, Ohio, and another small plane flying from Iowa to Minnesota crashed into a home, killing all on board. According to the USA's NTSB (National Transportation Safety Board), between 2008 and 2023, there were 20,684 total accidents, of which 3,764 were fatal, resulting in 6,254 fatalities and 3,745 serious injuries.
As the predictive maintenance market is witnessing significant growth, could there finally be a solution for these fatal incidents?
Israeli video-sensor solutions for critical systems developer OdysightAI (Nasdaq: ODYS) claims to have successfully implemented Predictive Maintenance and System Health Monitoring across the Aviation & Aerospace, Energy, Transportation, and Industry 4.0 markets. 'Our technology predicts hazards early, reduces maintenance costs and labor, and enhances safety,' says CEO Col. (res.) Yehu Ofer, a former IAF Hercules squadron commander. 'With advanced micro-cameras and specialized AI models, we provide predictive maintenance and monitoring in the most challenging environments.'
A small propeller plane crashed into a field in St. Augustin near Bonn and caught fire. Photo: ... More Benjamin Westhoff/dpa (Photo by Benjamin Westhoff/picture alliance via Getty Images)
While the company's solution was initially created for the healthcare realm, they discovered, almost by accident, that it could be a game-changer for Aerospace & Aviation.
'In 2017, an IDF Apache helicopter crashed during a training mission, resulting in the death of a pilot and critically injuring another,' Ofer explains. 'According to the final report, the crash was caused by the improper installation of the aircraft's push-pull mechanical rod, which is part of the flight control system. The helicopter was equipped with a fully functional Health and Usage Monitoring System that monitors dynamic components such as gears, bearings, and shafts but was not designed to monitor other critical components, including the one that caused the accident.'
After the investigative committee determined the root cause of the failure, the IAF and IMOD approached OdysightAI for a proof-of-concept system to prevent similar future mishaps. The system integrated vision-based technology for health monitoring of critical flight control components like the one that failed. 'This accident led us to develop a new layer of preventative maintenance which we believe will save future lives and valuable equipment,' Ofer says.
The company's first priority became improving flight safety and preventing incidents that aren't currently addressed by existing PHM technologies.
Today, the first IDF Apache is already flying with OdysightAI's cameras installed at critical fail points. These cameras record sensitive parts that are not readily accessible even to mechanics and use artificial intelligence to analyze images for cracks starting to propagate or parts that are loosening.
'In the Apache, the part that failed was checked only once every five years during the Level D check at the Air Force Maintenance Unit. But now, the camera prototype constantly monitors this part, an increasingly important matter given that some of the Air Force's Apache fleet is 30 years old. The entire military aviation world is dealing with the challenge of keeping them safely flying.'
"'We are rapidly increasing the adoption of our system in the aerospace industry while expanding ... More into new verticals with significant value-add.' CEO Col. (res.) Yehu Ofer.
Their solution also won a Ministry of Defense contract to install its system in the new Seahawk helicopters, second-hand U.S. Navy aircraft that have been taken out of service, renovated, and are now operated by the Air Force for the Navy's missile boats.
'There are now systems that are supposed to identify malfunctions before they happen by analyzing vibrations. Our system is based on a camera and is more accurate,' Ofer says. 'For example, the Seahawk will have 20 cameras at critical points, enabling the early identification of cracks that could cause the tail of the aircraft to disintegrate.'
The company recently marked another milestone collaborating with NASA. Odysight's PdM and Condition-Based Monitoring (CBM) systems are now being integrated into several aeronautical flight trials, investigating their operations at high speed and extreme environments.
NASA aims to embed cameras within aircraft to continuously photograph key components and detect any damage during takeoff, damage which often occurs unnoticed during launch and is only discovered in space or upon return.
As recent research indicates that predictive maintenance solutions today result in cost reductions of up to 40% compared to reactive maintenance and between 8% and 12% when compared to preventive maintenance, this solution seems to be aligned: 'The information collected can dramatically reduce preflight check time by 40% compared to current procedures, saving mechanics' work and shortening turnaround time between sorties, a critical factor during wartime, and beyond.'
These days, the company is already targeting the civil aviation industry as its next sector. 'Civil aviation makes decisions more slowly than the military, but we intend to become its strong maintenance standard.'
It seems like the future of transportation will combine intelligent infrastructure with advanced AI technologies. As global aviation regulators begin mandating AI-based predictive maintenance, how drastically could the safety landscape change?
Ofer believes that adding a vision-based AI layer is a natural progression of safety improvements in the industry and he doesn't plan on stopping there: 'We are rapidly increasing the adoption of our system in the aerospace industry while expanding into new verticals with significant value-add.'
If AI-powered predictive maintenance can help us catch what we've been blind to, perhaps fewer stories will begin with news of another preventable crash.
