“Providing a real-time, in-depth picture of the aircraft and its surroundings — both immediate and further away — would be a game-changer, massively enhancing situational awareness…” We couldn’t agree more. Except to say the impact of #AI will be so massive that a new term is warranted: #SituationalIntelligence — the ability to understand and make sense of the current environment and situation while anticipating and reacting to a future threat. Pilot Fahad ibne Masood, MRAeS, MBA-Av.Mgmt. details the many benefits of AI in the #cockpit in an article in FlightGlobal. #MachineLearning #ML #AviationSafety
Daedalean AI
Herstellung von Bauteilen für die Luft- und Raumfahrt
Autonomous flight control for the aircraft of today and tomorrow
Info
Daedalean creates safety-critical AI systems for situational awareness and flight control, with the eventual goal of developing certified autonomy in the air.
- Website
-
https://daedalean.ai/
Externer Link zu Daedalean AI
- Branche
- Herstellung von Bauteilen für die Luft- und Raumfahrt
- Größe
- 51–200 Beschäftigte
- Hauptsitz
- Zurich
- Art
- Privatunternehmen
- Gegründet
- 2016
- Spezialgebiete
- aviation, safety-critical software, computer vision, machine learning, robotics, neural networks, avionics, ai, artificial intelligence, autonomous flight und pilot aid systems
Orte
-
Primär
Zurich, CH
-
Phoenix, AZ, US
-
Riga, LV
Beschäftigte von Daedalean AI
Updates
-
Wire strikes are a leading cause of fatalities for low-flying aircraft. We believe that a vision-based detection system, capable of identifying both powered and unpowered wires and masts, will significantly enhance flight safety. Daedalean is currently developing such a system (see product factsheet here – https://lnkd.in/dBUiA7Zh). In collaboration with EPFL and supported by a grant from Innosuisse, we have developed a deep learning approach that allows a single #NeuralNetwork to perform both detection tasks. This month, EPFL submitted the final scientific report, marking the successful completion of this project that exemplifies the fruitful collaboration between academia and industry. The researchers combined modern convolutional architecture with transfer learning and an adapted loss function for curvilinear structure delineation, demonstrating its performance on two benchmarking datasets. The neural network was then implemented on Daedalean's onboard system, which includes cameras and a computing unit installed on our test helicopter. During the test flights, the system's performance exceeded our expectations in terms of detection distance (several hundred meters) and accuracy. See this short video clip (https://lnkd.in/dithMNB7) filmed on one of the flights. The preprint of the paper by our joint teams (https://lnkd.in/dy5ZBFFQ) provides detailed information on the model’s architecture, implementation, and performance. #MachineLearning #cockpit #avionics
-
-
Something for the (obligatory) vacation time: if your kids are 3-7 year old, they and you are welcome to join our Musica-Matematica-Mare (the sea) camp in Sardinia in October 2024. For more details and reservations visit https://mumama.net/
MuMaMa
mumama.net
-
On Monday at EAA AirVenture in #Oshkosh, Daedalean’s President of the Americas, Dr. Yemaya Bordain, will outline how #AI is transforming flight and how disruption of the aviation industry is coming from General Aviation. If you are attending, please come by Forum Stage 2—Aero Hose Shop at 11:00 AM on Monday, July 22. #generalaviation #avionics #osh24
-
-
A few months ago, we announced a strategic collaboration with Xwing (Acquired by Joby Aviation) to share data as we both work toward certifying #AI systems. Since then, Joby Aviation acquired Xwing’s autonomy division but the collaboration continues to move ahead. In the July issue of Avionics News, David Hughes does a deep dive in an article (https://lnkd.in/dUUdXN3m) on each company’s assurance approaches to certifying #NeuralNetworks as software. #MachineLearning #ML #avionics
Avionics News July 2024 Page 20
digital.avionicsnews.net
-
Test #flying an experimental aircraft requires a unique skill set: a test #pilot is not only highly skilled at specific flight test maneuvers but also specially trained to evaluate new technologies with deep knowledge of aeronautical engineering and analytical skills. That’s why we were especially pleased to offer an in-depth look at Daedalean systems to James E. Brown III when he recently toured the Zurich offices with our CEO Luuk van Dijk. Who would be better suited to assess what our technology will bring to the #cockpit and how #pilots could benefit than the legendary (https://lnkd.in/d5Z9fJfp) president of the US National Test Pilot School (https://www.ntps.edu/) with 30 years of test flying experience? We are happy to say that Brown summed up the day as ”‘quite informative and inspirational.” #AviationSately #Aerospace
-
-
This winter, we proudly shared a video (https://lnkd.in/d9X4_KQx) of our joint flight tests with Leonardo Helicopters, who evaluated our #AI-based situational awareness functions. Recently, Paul E Eden of Aerospace Testing International featured this story in an article (https://lnkd.in/dZEMiJYU) covering advancements towards #AutonomousFlight by all the major #aerospace players. We are offering a detailed account in our new blog about how it happened, where the flights occurred, what Leonardo’s engineers tested, and the results of the testing. #avionics #cockpit #helicopter #helicopters #aviation
OmniX: A deep dive into Leonardo's flight tests of Daedalean's visual awareness system
daedalean.ai
-
Congratulations to the consortium on the successful conclusion of the Machine Learning Application Approval research project. Another step towards making certifiable, safety-critical ML-based systems a reality. #ArtificialIntelligence #MachineLearning #Aviation
The final report of the Machine Learning Application Approval (MLEAP) research project is out! MLEAP plays an important role in #EASA ’s #AI Roadmap. Three main technical axis were investigated during the project: - #Data representativeness - Generalisation guarantee - Model and learning #algorithm robustness and guarantee Get the report 👉 https://lnkd.in/eExJSVfw The MLEAP project lasted 2 years and was implemented by a consortium consisting of: Airbus Protect, Numalis and LNE MLEAP will hold its final event during the EASA AI Days, on July 3rd. You can still register 👉 https://lnkd.in/eJ6vWBqY #HorizonEurope #aviation #ArtificialIntelligence
-
This was a stunning polylogue. Thank you, Adam Twidell, for assembling such a panel of experts and for the honor of being a part of it! What a brilliant idea to give a stereo view showing the concerns and approaches to the upcoming autonomy and to single-pilot operation along the way—from the POVs of an operator, a novel aircraft OEM, and an avionics developer.
My thanks to Ivor van Dartel from VÆRIDION, Kalin Stoyanov from Daedalean AI and Colin Russell from Flexjet for joining me at #EBACE for our panel debate on Empowering Aviation Safety through Autonomy. We started by discussing what autonomy means to aircraft operators, third party suppliers and aircraft manufacturers. The discussion moved onto the expected speed of rollout of new innovative products and what was slowing this down. Most controversial was the concern that the huge investment into Advanced Air Mobility could itself be a safety concern, as OEMs begin to feel huge commercial pressure from their investors to get a minimum viable product into the market, even if that involves stripping out more complex safety related systems for later generation models. "What could possibly go wrong?" when flight test timelines need to be hit. That debate continues...
-