Artificial intelligence (AI) can be paired with advanced sensors to provide a higher level of situational awareness for multiple aircraft types, according to Swiss start-up Daedalean. The company is working to integrate the technology into helicopters operated by Brazil’s Helisul, for an initial use case it believes would pave the way to wider adoption of a system it believes can boost both safety and efficiency.
Daedalean has filed an application with EASA for a Part 27 supplemental type certificate for the Ailumina Vista visual traffic detection system to be installed in as-yet-unspecified rotorcraft. According to the company, both EASA and the FAA have shown a willingness for this process to be a test case for certifying machine-learning-based systems.
Ailumina Vista helps pilots to detect and avoid airborne obstacles such as drones, paragliders, balloons, and birds. The company has trained the algorithms of the AI component to recognize these items, combing this with simulated data developed by its annotation team in the Latvian capital Riga.
Kalin Stoyanov, Daedalean’s head of business development, explained to AIN that to gather the images needed to train the AI algorithms, the company installed cameras and recording equipment in a glider and entered a glider competition. It took the same approach with a balloon competition.
Autonomous Flight Comes Next
In the first instance, Daedalean sees its technology supporting piloted flights with an extra layer of safety. However, its longer-term vision is for this to pave a pathway to autonomous aircraft operations, which Stoyanov said will come “when the industry and technology is ready.”
More than 12 months ago, Daedalean tested a visual positioning and landing system with Leonardo. It anticipates this collaboration could go further once the helicopter manufacturer has determined which platform is best suited to adopting the technology.
The company has also fielded at least one approach from a prospective business aviation partner interested in solutions for detecting runway incursions. “Business jets don’t only land at primary airports, and so there are risks such as animals on the runway that has meant interest in a visual landing system that can be used at smaller airports,” Stoyanov explained.
Daedalean also sees major avionics manufacturers as prospective partners. For now, the only publicly-acknowledged alliance is with Moog Avionics (formerly Genesys) with which it is developing an integrated cockpit.
In 2021, Daedalean worked with Avidyne on an FAA STC covering its PilotEye visual traffic detection system for Part 23 aircraft, and potentially for advanced air mobility vehicles. This was integrated with Avidyne’s IFD flight management system to merge visual information with ADS-B data through Avidyne’s SkyTrax technology.
On top of its safety role, Stoyanov said improved situational awareness also supports greater operational efficiency by increasing the availability of aircraft. The company is now looking to integrate infrared cameras and radar to expand the missions the system can support.