Some of you may have already seen our rugged professional mission control units also presented here on DIYdrones. Most of those are produced to at least industrial IP specifications, but developers and DIYers are constantly bombarding us with requests for lighter and especially cheaper integrated solutions that would be at least on the edge of the affordable in applications where fully rugged builds and tightly waterproof features are not an absolute must.
Well, this is our first shot at this, so we're trying to get as much of your valuable input as possible to create a practical hand-held controller that may as well help a lot of people. We've already built the first prototype based on our current RHH line, some of our own ideas and customer input, so it's probably a good start for a conversation on the technical details for a more generic purpose unit with extensive DIY options.
Unfortunately, even if we drop most of the IP67 parts and build features of the RHH design, this particular product will still cost around 1200 Euros in very small series, but at least it's not the fully rugged €4k-€7k range. The final price is going to be determined mainly by production numbers, but at the moment we're just aiming for a batch of a few dozen units, just to be on the safe side.
The current technical content of the prototype can serve both direct FPV operation and/or payload control, both in the very same neat little package:
- 7" high contrast and high brightness (700 nit) 1280x800 screen, non-reflective matte or low reflectance glossy surface
- non-bluescreening LCD controller with AV/HDMI/VGA input and full menu control
- USB programmable er9x/OpenTX compatible open source RC TX logic with full menu control (eepe compatible)
- high quality quad ball bearing Hall effect sensor RC style joysticks with full access to slide/ratchet/spring settings on both sides, thumber and pincher operation supported by design
- fully managed Li-ion battery system with simple micro USB charge port and high current connector alternative
- industrial grade ultra miniature toggle and pushbutton switches, optional pot and rotary switch
- high impact industrial ABS enclosure, 210x280mm compact footprint
- charger, USB programmer, PPM and HMD AV connections as standard
- optional built-in RC TX RF module (FrSky XJT, TBS Crossfire, EzUHF, etc.)
- optional built-in AV RX (5.8GHz, 1.2GHz, etc.)
- optional external RF box setup with single field cable
- optional external video connection (CVBS/HDMI/VGA)
- optional wireless ground video relay (Miracast)
- optional IT hw for IP video streaming, telemetry, etc.
As you can see from the above, this is not a telemetry station with any kind of embedded IT hardware, just a direct FPV and payload controller. We currently manufacture a fully rugged 10" all-in-one telemetry GCS with a powerful Android hw, which we would also like to make a light version of, but we're still working out the kinks there, and we'll let you know.
The main questions we would like to ask You are mainly about the acceptable or necessary build and kit readiness levels, the most generic but still practical control layout and labels, built-in or optional RF modules, etc. We also welcome any other input that you may find important here. This is just a short list of issues that we're contemplating at the moment:
- Toggles, pots, other controls... shall we adhere to a generic RC control layout (especially 9x type)?
- What is your opinion of the current custom layout that caters for both payload and FPV operation with remapping?
- Pots in particular, do you need them at all in a professional controller?
- 6 position rotary flight mode switch, do you need them while you barely use 2-3 flight modes in general?
- Joystick trims, do you need them at all in a professional controller? After all, they are a major source of mishaps in non-RC trained fiddly hands...
- Single PCB construction or PCB mounted RC switches with connectorised/solder pad-based wiring for DIY purposes?
- Moderate waterproof features, we can seal the electronics of the open joysticks, so the water can go simply through the body with port holes in the bottom service covers (or built-in silica pad) ... is this something you need to survive a passing shower?
If it's not clear from all of the above, this is NOT a crowdsourced project, at least not financially, so you don't need to pledge any money to take part in this little endeavour. We're not a one-trick pony start-up venture as we've been in this industry for a decade now, but we have to keep learning from our customers and peers.
You should look into Saiy or any other similar voice command application, because the above is already possible with Android tablet level hw. Because Saiy is now open source it may as well be easier to customise the code to this very purpose. As we're also working on a light version of our 10" tablet RHH we'll soon have a more economical all-in-one telemetry station able to perform these functions at the hw level, as well.
I think most of what you need is feasible if the RC override function is used, so the RC commands go through the IT hw and the telemetry modem using a game controller module, instead of a fully fledged RC TX logic like er9x/OpenTX. RC override has long been supported by Mission Planner on a PC, but QGC on Android also has it, so there's nothing against implementing it with voice command.
Operating the controls blindly i.e. without having to look at the panel can be made easier by careful arrangement of the switches and by using finger guides. The guard tabs between the switches can also help. We still have to simplify and rearrange the top row here, however.
To tell you the truth, speech recognition will be such a breakthrough that I am afraid to even dream about it for fear of disappointment. Often, taking eyes off the model even for a second can be extremely inconvenient. Very recently, I spent panicky seconds trying to find a black, small and fast wing after having very briefly glanced at the remote to see battery level.
Speech recognition opens up a world of possibilities, least of which is simplification of the hardware, you can replace a zillion buttons with one mic. External noise filtering and mic waterproofing can be even more simplified by using a wireless throat mic.
This would make field operations a lot easier, I currently either have a second person getting bored out of their minds sitting at the laptop so I can fire an occasional question at them or I have to make sure the vehicle does not need much baby sitting so I can have a look and do whatever needs doing.
Parsing and avoiding accidental input can be simplified by following the paradigm of "OK Google..." that primes the software for a command input. In terms of a remote it could be something like commands and queries, e.g.
Command examples (COMMAND: argument list):
- Pilot "MODE: stab"
- GCS "beep" indicating command parsed and executed.
- GCS "Mode changed to stabilize", once MAVLINK confirmation has been received from vehicle
- Pilot: "SET: throttle, 30"
- GCS "beep" indicating command parsed and executed.
- GCS "Throttle set to 30 percent", once MAVLINK confirmation has been received from vehicle
Query example (QUERY: item):
- Pilot "SAY: airspeed"
- GCS "Airspeed 14 metres per second"
I don't think it's a dumb question, because voice command can be a viable alternative in certain UAV applications. Since it needs at least some embedded IT hw it may as well be an option in the 10" telemetry version, which is based on a fairly powerful Android tablet. There are already a number of 3rd party voice command applications available that can be used as a programmable interface, and open source ones are out there as well. Coupled with the use of the RC override function of the GCS sw it may even be more straightforward for some direct RC functions.
I like radio buttons in general, and it would be a great alternative for flight modes, but unfortunately we were unable to find any decent ones in this sub- and ultra-miniature size. If you have any idea where to find some suitably small and reasonably well protected types I'd be glad.
This is probably a dumb question but what is the current state of speech recognition? If a small subset of commands (the flight modes really, they only a handful) can be recognised by a relatively low powered processor, that would be a dream come true. Any means of changing flight modes currently available is horrible.
Failing that, a physical possibility for mode switching is mappable radio buttons (or three buttons, up/down and select, like you already have but for modes only, somewhere where a thumb can get to them and with audio feedback). There is space left and right of the screen. I don't know if they would make the cost explode.
As many people have pointed out both rotary pots and sliders as important features, we have been looking into this in detail. It is possible to install two sliders on the top edge of the frame at the corners to be operated by your index finger.
The main reason why we haven't even thought about installing them in the first place is the rugged lineage of the original RHH design, since sliders are notoriously hard to protect from the elements. By using the vertical side wall of the lower half of the clamshell enclosure the sliders would stick out horizontally, thus limiting the amount of dust or water that can fall into them. Although the camera operation is already taken care of through joystick remapping during any of the automatic flight modes, sliders or pots are still an option.
One of the things that stand out right away is the lack of slider switches. Slider switches for me are key as I like to use pan and tilt on my cameras.
I have no worries about the ergonomy of the original design since we've been manufacturing it in small series for some time now. See here: RHH
The "square" shape is a given with a monitor if we don't want to enlarge the footprint with unnecessary protrusions. The rounded corners and the neatly chamfered edges make it a very comfortable hand-held package, especially with a 30mm main enclosure thickness.
If you need more flight modes a rotary flight mode switch is still an option somewhere below the current logo. However, we may then have to mirror the current layout since most multirotor pilots prefer keeping their fingers on the throttle while changing modes.
Pots are also possible, but as this layout can be remapped during any of the automatic flight modes to payload control with a flick of a switch, your camera is already taken care of within the same set. The current preliminary switch layout is an aerial video application combo that comes from several existing client specifications, but it's far from fixed or final in any way. The switch guard tabs help you navigate even the single row on top without looking, though. It is also possible to mount the same type of switches around the top chamfer and edge of the set, but since these are protruding controls we also have to protect them. Sliders on the other hand are out of the question as they take up a lot of space, and it's extremely hard to protect them from the elements.
The reason I use 2-3 flight modes on the Tx is that it is easy to accomodate on most Txs. This is also why I never head to the field without a laptop, Tx alone just doesn't cut it.
Personally would absolutely love to have at the very least MANUAL, FBWA, FBWB, LOITER, AUTO and RTL. Also, a visual interface for GUIDED is a strong must, which is why I generally prefer laptops over tablets.
Regarding pots, how would you move gimbals if you don't have pots (by pot I take it you also mean slider)?
Trims are useful but don't really end up using them that much once the vehicle is finished. If you don't want to include them in the hardware, perhaps somehow enable them through the software?
Waterproof features would be nice but not crucial, especially if they raise the price a lot.
at a first glance that looks nice. Directly to your questions:
- adhering to a generic RC control would let most used pilots to feel at home, but hey, there's no reason to stick at that if reasons are good
- can't understand what you mean, without a concrete example
- pots are nice if you have direct, in flight control e.g. over gains
- yes, rotary knob mandatory (on serious flying things I usually use 5-6 flight modes). An alternative that I personally use right now and like so much are three positions switches on top of the sticks, so you can have at least 5 flight conditions
- trims also mandatory for fixed wings (is that controller meant for those too, isn't it?). Of some use on some multirotors too. You could just disable them as default and those who really need them could select the functions and assign a channel (Nice to have them crossed, so with left hand one trims the right stick etc)
- I like some DIY possibilities on controllers, so I can suit my taste positioning extra switches, connections etc.
- some sort of splash proof would be nice too
About the case itself, one should have it in hand and eventually try to pilot something to see how it feels. My doubts from pictures are about stick advanced position and especially the missing of any pad at bottom. Some sort of ergonomy is finally appearing in almost every controller on the market now. Your's seems a bit ... squared.
About switches position I have big doubts about the usability of the row of switches: how can one recognize (and remember!) which one is the right one? A differentiated positioning and switch type, long and short knob etc., could help a lot.
Switches and sliders on the sides and at bottom are great too, even for a pincher.
What about a panic switch, big and handy? I always program such a function on my controllers.
This resumes how I pilot, so strictly personal (summarizing, performance gliders on slope, even equipped with UDB autopilot, powered planes, autonomous or not, DIY autonomous multirotors, task and racers,...)