Flying car autopilot “made in Switzerland” – financed in Australia

Interview with Luuk van Dijk, CEO Daedalean, Speaker at the 5. Silicon Valley meets Switzerland Event, March 28 in Zürich

Tell us about yourself and what Daedalean does please.

Daedalean was created in 2016 on the premise that within a decade we can have ubiquitous electric vertical-takeoff-and-landing aircraft (eVTOLs), so called flying cars. Developments in batteries, structures and motors, driven by revolutions in mobile computing and cars, have made electric flight physically possible and because it can be so much simpler, cheaper and safer, also economically inevitable.

To enable mass application of these so called “flying cars” they need to be fully autonomous, so that to operate them safely over the skies of densely populated cities, a pilot license — or indeed a human pilot — is not required. Reports like the ones released by Uber, Airbus and even UBS, shortly after we incorporated in 2016 subscribe to this view and identify full autonomy as one of the critical enabling technologies. There simply aren’t enough humans with pilot licenses, and humans aren’t that good at flying in densely populated airspaces to begin with.

For this reason Daedalean has set itself the goal to invent an autopilot that can be a full replacement for the human pilot in these vehicles. As our product roadmap we took all the functions the human pilot has to prove during his or her flight exam: for each of these (about 40 in total) we want to build systems that convincingly outperform the human, which we intend to show by having our system actually passing that exam at some point.

The first big chunk of these functions, currently uniquely allocated to the human in control, is _visual_ guidance, navigation and collision avoidance, in what is called Visual Flight Rules (VFR). Flying on instruments (Instrument Flight Rules or IFR), computers can already do better (if it wasn’t for the fact you have to talk to an Air Traffic Controller over a voice channel at the moment) but to integrate seamlessly into the way Civil Aviation works worldwide for now and for the foreseeable future, you have to build a system that can see where it is, where other things are, and that can understand what these things are (other aircraft, mountains, wires, the ground, …) and then takes sensible decisions given the situation.

We have demonstrated these functions on a big UAV: we can take off and fly from A to B reliably without GPS, recognize objects in the sky and identify safe and unsafe landing sites and guide the aircraft to the ground, all on pure visual camera input, like a human is required to be able to do.

After we started it turned out that while we are waiting for the eVTOLs to come to market, there is already a huge opportunity for these instruments in existing aircraft (except that the dynamics of the aerospace market, with its regulatory capture and slow product cycles had obscured it).

When will we realistically see these flying cars around us?

We set our full roadmap for ten years starting 2016, so 2026, but many of our target customers, (over 100 listed on evtol.news) think that by 2021 their first piloted aircraft will reach the market, and that by 2023 more or less usable autonomy will be realized. We certainly don’t plan to be the ones holding them back.

What are the main challenges technically and otherwise?

Seeing and understanding require that we bring recent developments in robotics, computer vision and deep learning to the world of certifiable safety critical avionics. We have to develop “tomorrow’s algorithms”, but because of restriction of what computers and computer programs are considered safe enough to put in aircraft, we have to run on “yesterday’s computers”. That also means we not only have to develop systems that are safe enough, but also the means to prove that they are so. Especially when it comes to the application of modern Neural Networks that means we are at the forefront of bringing academic results to practical application, and that requires working with regulators (FAA, EASA) to define how to measure and achieve certifiable safety.

Why would we need this? Does it solve any current transportation challenges?

The electric flying car promises to solve a number of obvious and ubiquitous problems with earth-bound transportation and could change urban and regional transport dramatically, esp as megacities keep densifying. By themselves they won’t solve global warming, but because electricity allows to move the conversion of energy to be moved to other, cleaner, centralized power plants, and because the batteries currently are still the bottleneck there is a huge drive to consume as little energy as possible in flight (just as a Tesla drives 540 km on the energy equivalent of 10 litres of diesel), there will definitely be large positive effects on the world’s energy consumption. Every diesel engine currently stuck in a traffic jam that we can phase out is cause for celebration.

Without full autonomy however, the flying car is doomed to remain a toy for at most a few hundred rich and bored people. That would be a waste of an opportunity.

How is it to have this company in Zurich?

Zürich is a great place to find smart young people, the ETH and the University have a number of research groups in robotics, computer vision and deep AI that are among the world’s most prominent (as also evidenced by some big tech companies having offices here). Because of the high quality of life compared to, say, the San Francisco Bay Area or London, it is also quite easy to attract great people. So recruiting has been fun.

With respect to raising funds, i find the Swiss VCs fairly low in ambition and vision. The local investor scene is mainly interested in preserving wealth rather than taking risk, and the typical highlight of an investor event is a claim that the next Silicon-Valley-of-X is here when someone raises 100kCHF to expand the next big-data blockchain powered online cat food portal into a neighbouring Kanton. A copy of a US- or even DE- proven business model has the highest chances of finding local capital. That has been a bit of a disappointment.

However it is much easier to get the people here and the money from abroad than the other way around. The lead investor in our series A for example is Carthona Capital, a fund based in Australia, that keeps their eyes open for ambition at the far end of the risk-reward curve and considers the whole world their playground, as do we.

How do you view the discussion that Europe/Switzerland has already lost the tech platform race to the US and Asia?

I’m not sure what ‘platform’ and ‘race’ mean in this context but it sounds like you claim something irreversible has happened and that’s definitely not the case. I see no race going on with a well defined start or finish, instead there is a permanent competition of ideas and execution in all industries, with waves of creative destruction begetting next generations of technology and enterprises, and the balance can shift any time.

In the realm of the eVTOL the Europeans are currently in the lead, with Volocopter, Lilium, 2 Airbus projects, and EASA putting out the first regulatory basis for certification of eVTOL aircraft for private and commercial use.

Merely investing megadollars does not shift the balance (although it helps), and a tech industry mainly occupied with scaling up sharing of photos and clicking on ads does not necessarily produce the next ‘winner’. Instead of pursuing growth by chasing hype cycles, i think we can bring more quality to life and planet with climate of free and stimulating academic thought, rules and regulations that encourage innovation and a supply of creative young people who want to work on the most interesting and urgent problems they can think of.

Picture source:

https://www.theverge.com/2017/11/13/16643342/volvo-geely-terrafugia-flying-car-acquisition

Leave a Reply

Scroll to top