Hi there! This is the DBX Drones team. Some time ago we presented you our design concept for a hybrid cargo drone. and since then we have been working on the idea and two months ago we decided to start working full time in this project and found a start-up: DBX Drones! During these last two months we have been creating the team, moving to a start up incubator, Lanzadera in Valencia, and starting to test our controller.
Right now we divide our time into two main objectives, the business team, trying to learn what do our future clients really want to have, and the engineering team, that is developing the flight control system and testing the platform so our client desires are satisfied.
After the last survey we made, we realised that people prefer to avoid touching low level code and have plug and play solutions, but at the same time have access to the inner workings of the system. For the time being, we have decided to go with Pixhawk autopilot due to the big community behind it, because it is open source, and because we would like to learn enough to give some knowledge back to the community. We have had to redesign the vehicle concept several times, due to unwanted vibrations that were resonating along the structure. You can se above, the evolution and different design proposals. We have decided to add an additional carbon fibre tube to increase rigidity during the multicopter mode flight tests.
Software wise, one of the principal efforts has been put into developing a good model of our platform, in order to test the control schemes as much as we can before performing the risky, time consuming (and deeply satisfying when they are successful) tests. Since our main idea is to be able to fly with different cargo weights, we have chosen to implement a model following neuro-adaptive non linear control, this way we are aiming to minimize the time we spend reconfiguring and retuning anytime we change something in the drone.
Here you can see an attitude test from the last month. We were thinking about what video to post from this last month, we chose this one because it was the first time we were testing our nonlinear controller. We were expecting it to turn wild, the obtained performance wasn't expected given that we hadn't tuned the controller at all. To give you a rough idea the controller thought the vehicle weighted 3kg more than it really did. The vehicle was capable of maintaining the attitude despite the lateral wind, the torque the straps produced and the severely damaged yaw mechanism we had at that time. We are now actively working in obtaining a good measurement of the ground speed so we can implement a speed loop control scheme. The next step will be testing the platform as a fixed wing and finally process to the cumbersome yet exciting transition tests!
We would love to have any feedback you can give to us about how we can improve our vehicle, development process or business. We are happy to learn everyday from the huge diydrones community and to be here doing what we like the most.
Please take our SURVEY so we can make something you like!
Thank you and see you soon.
Comments
Hi Andreas
Parcel delivery is still an option, but is simply not one I personally endorse. ;-)
In our use case it's for SAR operations, where access is confined and we need to operate out of limited glideslope ad hoc areas as close as possible to the required area requiring coverage. Aerial coverage is not so much dictated by overall range as it is by speed, especially when fighting changing weather and a setting sun. Long endurance is not the same as long range and large coverage (hot balloon vs jet fighter). There's only so much photography that can be done in daylight hours, and being able to deploy regardless of the site is invaluable.
The platform we're developing right now for the OBC is capable of electric 60km flight at over 100kmh cruise with 2kg VTOL. That's beyond what is possible with electric helicopters and borderline what can be done with a deep stall type aircraft like the Cumulus, which can't do vertical takeoff at all (They've been taken over recently?). Reliably returning the aircraft, even under pilot control in harsh conditions because it can VTOL both up and down at low speeds, means it can fly another mission after physically returning images to the ground (even if RF comms fails etc). In comparison I doubt that the Cumulus can keep up with the rate of coverage.
On the subject of launching; we used to use a catapult for our X8 frames, but was way to clumsy to cart around and deploy. A simple bungee worked better, but neither give true vertical performance, let alone VTOL. A small sub 1m backpack-able sub 2kg quadplane without any is much easier to handle.
One of the main considerations for safe landings of conventional aircraft is having a low approach speed which requires a low stall speed. This dictates the design of the airfoil to be large, which means for cruise it's at best a compromise. (Flaps etc can mitigate the effect) A quadplane need not have any such compromise as the quad can maintain lift until wing lift is achieved. Accordingly, in time i'm expecting better than conventional plane range and speed performance from a quadplane as a result, at the price of a significantly reduced flight envelope.
Note that NASA and co are developing using hybrid designs (including gas/electric hybrids) for the same reasons, like the Greased Lightning etc.
Regards
@Jonny Glad you like it! We are building it for people like you!
Your are right, you could point rear prop downwards and then use it for propulsion.
The thing is, we haven't seen we need it and want to avoid the mechanical complexity of doing so.
But, if you guys says so in the survey, we will try to add that feature!
Love this!
Is it not better to face the tail motor downwards and make it a pusher? That way during transition you still have vertical force, and it just slowly translates to horizontal.
With your upward-facing tail motor I would think that starting a transition would instantly decrease vertical force, and you would sink a whole bunch.
Just some thoughts!
Fair enough, so, if delivery is not the use case, what is (genuinely can't think of anything)?
If the vehicle is light, you can hand launch it. If it is heavy enough to necessitate carrying it around in the boot of a car, you can also carry a catapult. If you have a short length of tarmac, you can use it as a runway to takeoff.
For landing, any field without too many trees can serve as a landing area for out of the box landings (APM scripted via MP) and for the demanding, there is Cumulus One (0:34).
So, unless you can think of a take-off or landing scenario that is specifically not addressed by the these, the only use case remaining is mid-flight stop and hover. This is pretty much covered by traditional helicopters, which also offer better efficiency than multirotors.
Of course, once something is out there, people might always find really cool uses for it, it's just that I 'm struggling right now to come up with something "serious" off the top of my head. Of course, "serious" is not necessary, once you get the control algorithms rock solid, you could make an absolutely awesome Imperial Shuttle.
Hey Andreas
Although I can mostly agree with what you are saying about alternative delivery methods, overall I'm fundamentally against drone deliveries of any sort, apart from urgent medical types. For delivery purposes there are much much better solutions available that don't even need to leave the ground. In fact I'd say most deliveries will become redundant in the next 10-20 years as digitally distributed manufacturing take up the bulk of mass production.
Technically you are right with your analysis that "hybrids" are a solution looking for a problem. I'd however add the problem is only in the the mind of men, that cannot conceive better solutions! ;-)
But on the subject of quadplanes/hybrids an co. If you take my previous comments to the extreme end of the scale there's not much "aircraft" left to fly the payload. It could be a aerdymnamically shaped payload container with propulsion and nearly no wings at all. (In this case the payload could be a camera instead of a delivery etc to make it worth while to fly!)
Unlike a normal plane or quad, a quadplane can be designed to operate at just two velocities: Stop and go ;-)
The point that it needs to "overcome" with the quad is the general landing and takeoff, which is achieved in hover to gain altitude above the ground, to get out of the way of obstacles in horizontal flight. At altitude the aircraft can then continue to use the quad for lift until what ever little bit of wing takes over (or flying body) and is sufficient to maintain lift for flight at cruise .
The result is a "hybrid" that can fly fast and far and can land anywhere. Neither a plane nor a quad can do this by themselves.
Until we create readily available morphing wings this is the closest system we have to biological type flight capabilities.
Regards
@Andreas that is a good point and some of the questions all of us must make. That is why we are making a survey: to understand what people value and need most on drones and to Validate if a VTOL is something interesting.
See here the survey, where all of us can give our opinion about this.
If we collect enough samples, we will share the answers with everybody and then we @All could take a decision about VTOLs :)
@Bart Thank you guy. There are great plans but it is a hard path to follow. that is why we need your help. Help from people like you and others here that have great experience :)
@JB Absolutely yes. We did design and optimizing everything for forward flight, while ensuring enough controllability and efficiency for VTOL. About wing surface area: the more you have the longer you fly only if you power plant can propel you.
The question @JB and @All is: What is the advantages and disadvantages of Tilting-props versus quadplanes? You guys have more experience on this
I 'm generally wondering what the point of hybrids is, beyond the coolness factor of course which is epic.
If you want to make a delivery and don't want to use a multirotor, an airdrop is a perfectly acceptable method of delivery that can, in principle, be reasonably accurate (by flying low and slow) and delicate by using a parachute and/or suitable packaging.
If the problem you are trying to address is not the delivery, is take off and landing really that important? If you have a designated base, you can hand, catapult or runway launch. Landings are also pretty accurate, even using generic scripted commands, all you need is a suitable spot. My feeling about hybrids is that they look a lot like a (multirotor inspired) solution looking for a problem that ain't there.
Not if you halve your wings/body area in the process... ;-)
The drag quadruples as well. Good job you can simulate all these things in software for free these days!
-
1
-
2
of 2 Next