Backed by a new round of funding completed this month, Swiss artificial intelligence start-up Daedalean is pressing ahead with plans to develop what it claims will be the first certifiable DAL-A level autopilot system to support autonomous aircraft operations. Design assurance levels (DALs) are used by aviation authorities to define safety criticality of a software application. The Zurich-based company is aiming for a DAL-C autopilot solution that will support pilots as an interim step to full autonomy.
The latest round of funding for Daedalean is backed by venture capital groups that include Carthona Capital, Redalpine, and Amino Capital. Daedalean, which was founded by a group of engineers who previously worked at companies such as Google and SpaceX, has now raised a total of nearly $12 million.
According to Daedalean founder and CEO Luuk van Dijk, the company aims to release an aircraft-agnostic DAL-C solution in 2021. It plans to partner with avionics manufacturers to get the technology integrated into aircraft cockpits to operate in tandem with existing flight controls. “We will leave the integration process to companies who make cockpit systems,” he told AIN. “Our systems will create visual signals, mainly using cameras, that match existing equipment like ADS-B and TCAS. Visual awareness is both the first and last instrument for safety-critical decisions and we will be able to show that cameras can do this automatically.”
Daedalean’s business model is to lead the introduction of autonomous flight software for applications such as eVTOL aircraft being developed for passenger- and cargo-carrying use and existing general aviation aircraft. The company acknowledges that regulatory structures are not yet in place to support the immediate integration of fully autonomous flight controls and so wants to support aircraft and avionics manufacturers as they take incremental steps towards the eventual goal of full autonomy.
“We have focused on the eVTOL sector because its business model won’t work if it is completely dependent on human pilots—there won’t be enough pilots and IFR restrictions would constrain traffic growth,” van Dijk said. In the meantime, Daedalean also sees more immediate applications on existing aircraft, including helicopters, for which it feels its autopilot system would provide an added layer of protection against risks such as controlled flight into terrain.
Daedalean has been working with U.S. avionics group Avidyne to develop potential artificial intelligence applications in general aviation aircraft. It installed its cameras on Avidyne’s Cessna 180 test aircraft to gather data to create algorithms to teach an artificial intelligence system how to pilot an aircraft.
In May, Daedalean conducted autonomy-enabling flight tests with German eVTOL developer Volocopter. It has also just begun an Innovation Partnership Contract with the European Union Aviation Safety Agency to examine the challenges posed by the application of neural networks in aviation to support machine learning and other artificial intelligence applications on aircraft.
Daedalean currently employs 22 people and expects this number to grow to 30 by year-end of 2019.
Backed by a new round of funding completed in mid-July, Swiss artificial intelligence start-up Daedalean is pressing ahead with plans to develop what it says will be the first certifiable
DAL-A level autopilot system to support autonomous aircraft operations. As a first step, the Zurich-based company is aiming to DAL-C autopilot solution that will support pilots as an interim step to full autonomy. Design Assurance Levels are used by aviation authorities to define safety criticality of a software application.
The latest round of funding for Daedalean is backed by venture capital groups including Carthona Capital, Redalpine and Amino Capital. Daedalean, which was founded by a group of engineers who previously worked at companies such as Google and SpaceX, has now raised a total of just under $12 million.
According to Daedalean founder and CEO Luuk van Dijk, the company aims to release the DAL-C solution in 2021. It plans to partner with avionics manufacturers to get the technology integrated into aircraft cockpits to operate in tandem with existing flight controls. “We will leave the integration process to companies who make cockpit systems,” he told AIN. “Our systems will create visual signals, mainly using cameras, that match existing equipment like ADS-B and TCAS. Visual awareness is both the first and last instrument for safety critical decisions and we will be able to show that cameras can do this automatically.”
Daedalean’s business model is to lead the introduction of autonomous flight software for applications such as eVTOL aircraft being developed for passenger- and cargo-carrying use and existing general aviation aircraft. The company acknowledges that regulatory structures are not yet in place to support the immediate integration of fully autonomous flight controls and so wants to support aircraft and avionics manufacturers as they take incremental steps towards the eventual goal of full autonomy.
“We have focused on the eVTOL sector because its business model won’t work if it is completely dependent on human pilots, because there won’t be enough pilots and because IFR restrictions would constrain traffic growth,” van Dijk said. In the meantime, Daedalean also sees more immediate applications on existing aircraft, including helicopters, for which it feels its autopilot system would provide an added layer of protection against risks such as controlled flight into terrain.
Daedalean has been working with U.S. avionics group Avidyne to develop potential artificial intelligence applications in general aviation aircraft. It installed its cameras on Avidyne’s Cessna 180 test aircraft to gather data to create algorithms to teach an artificial intelligence system how to pilot an aircraft.
In May, Daedalean conducted autonomy-enabling flight tests with German eVTOL developer Volocopter. It has also just begun an Innovation Partnership Contract with the European Union Aviation Safety Agency to examine the challenges posed by the application of neural networks in aviation to support machine learning and other artificial intelligence applications on aircraft.
Daedalean currently employs 22 people, and expects this number to grow to 30 by the end of 2019.