HUD dashboard development

About what was inconvenient with Vufine display: 
It is supposed to be located very close to eyeball and has lens designed accordingly. But you would not want to put it so close to the eye in wingsuit flight, as it may hit the eye badly. Attempt to position Vufine at reasonable distance creates "tunnel vision" effect - even a slight shift of eye from the optimal position makes the image to disappear. No practical use by any means.
As I posed the image earlier, assembling of AON2 on helmet front works much better. I already made about 200 jumps with this setup and believe it is a practical solution in terms of where and how display should be positioned. The AON2 can be replaced by any kind of custom implementation, but optimal position is as I mentioned. I'm currently designing my own custom solution with 5"  800 x 480 high-brightness LCD .

On 6/28/2021 at 6:21 PM, the.Legend said:

Another alternative is to use transparent OLED display. Surprisingly, there's a product with everything on-board except GPS:

Won't work: https://hackaday.com/2021/08/19/why-you-cant-make-build-a-wearable-display-with-a-just-a-transparent-oled/

Yes, I'm perfectly are of that problem. I'm trying to overcome it by combining screen big enough so that even blurry text/numbers at close distance are readable, and putting it as far as possible for the best results.

So far sticking OLED to the visor on the outside seem to be good enough. That's why it has to be transparent - big screen obscures more view.

Also it's placed not directly in front of an eye, but at the bottom and to the side - again to get more distance and obscure less view.

LilyGo TTV has arrived about a week ago and actually I'm quite impressed with what they've managed to squeeze in such a small package - it even has WiFi! And plenty of power in the processor, so with time the functionality could be extended.

I've hooked up the GPS and it works, now I'm sorting out few technical challenges
* Extend wiring so that unit could be hidden in one of audible pockets

* Test with different batteries. The stock one is too small, it lasted for just about 1-2 hours working only as a display. For that I had to hook up GPS module to external battery with higher voltage

* There's 5-10 seconds delay in GPS readings - displayed speed lags behind actual speed. Likely some bug in the code. Same GPS module on Arduino had better reaction time

In few  days I'll post an update with link to GitHub for anyone interested

On 8/26/2021 at 12:23 PM, the.Legend said:

So far sticking OLED to the visor on the outside seem to be good enough. That's why it has to be transparent - big screen obscures more view.

Also it's placed not directly in front of an eye, but at the bottom and to the side - again to get more distance and obscure less view.

How's the visibility in sunlight?

Prototype ready and assembled. I tested it today riding a bicycle in cloudy weather and it was readable. In bright sunlight this OLED is too weak on its own, but it could be improved by putting some tinted tape on the back of the frame, or in worst case a piece of duct tape.

Will be testing it this Saturday, hopefully, it will be able to catch up with readings. GPS module is rated up to 5 Hz (similar to FlySight) so if anything, the bottlenecks would be in the code.

I'm wrapping up the documentation & source code, will post it after field tests

44 minutes ago, the.Legend said:

but it could be improved by putting some tinted tape on the back of the frame

you can get tinted film for car windows quite easily, just cut it to size. There are even photochromic ones available.

15 hours ago, olofscience said:

you can get tinted film for car windows quite easily, just cut it to size. There are even photochromic ones available.

You can buy non-adhesive tinted window film as well in case you want to experiment without commitment. A tinted visor may also help.

2 hours ago, BMAC615 said:

A tinted visor may also help.

No, you want the tint between the OLED and the outside, not between the OLED and your eyes. So not a tinted visor.

4 hours ago, olofscience said:

No, you want the tint between the OLED and the outside, not between the OLED and your eyes. So not a tinted visor.

Is this not mounted inside the shell of the helmet?

2 hours ago, BMAC615 said:

Is this not mounted inside the shell of the helmet?

Looks like you're right, it is inside. I thought it was outside because he said earlier it would be hard to focus on something so close.

The unit has been tested through the weekend and I'm really happy that it worked! Unlike previous attempts, when proprietary GPS units couldn't handle freefall, this time GPS data was displayed during all phases of the flight.

Update was to 250ms and unit reaction time was pretty accurate. GPS unit rated up to 5Hz, with time it will be upgraded to 10Hz (same as in FlySight)

I didn't have time to order tinted tape, so I put black one on the back, and on a bright day digits were readable no problem at all.

Despite digits being blurry (the image attached is pretty close to what's in reality) it was easy to read them during the flight. Display is not distracting and could be looked at with quick eye movement. You might need some practice on the ground first just to get used to it, but then during the flight it didn't take an effort to check speed.

As promised, basic source code and hastily written assembly instructions:

https://github.com/RomanTheLegend/WingsuitGps

Contributions are welcome because I won't have enough time to develop this project at a good pace, and there are few quite complex functions to be implemented.

Next on the ToDo list:

* Make few more units for my friends for wider testing and feedback (mid-October because shipping is only from China)

* Make 3D models and wiring cleaner and leaner

* Explore the possibility of attaching some control switches - that would allow to write some kind of menu interface in the future

P.S. I got carried away a bit and added Bluetooth intercom made from old cheap headset while at it. This helmet gets more and more advanced

It seems to me like an LCD might be a better choice than OLED, as it will be illuminated by the ambient light instead of trying to compete with it. Did you consider this?

Yes, I've just ordered few more units for wider testing, one of them is LCD-based. It has ~0.2 inches smaller screen and is a bit thicker, but on the other hand everything is fixed to the board i.e. no flex cables to deal with. And it's colorful.

Primary reasons for OLED were transparency (which backfired on me, but still few tests to be done with tinted tape) and the thickness of the display

Parcel would arrive around mid October, meanwhile I'll be looking at improving casing, wiring and writing new code.

5 hours ago, the.Legend said:

Yes, I've just ordered few more units for wider testing, one of them is LCD-based. It has ~0.2 inches smaller screen and is a bit thicker, but on the other hand everything is fixed to the board i.e. no flex cables to deal with. And it's colorful.

Primary reasons for OLED were transparency (which backfired on me, but still few tests to be done with tinted tape) and the thickness of the display

Parcel would arrive around mid October, meanwhile I'll be looking at improving casing, wiring and writing new code.

Are you having trouble finding a transparent LCD? I wouldn't expect it to be hard to find, as they're inherently clear anyway and they only tend to be opaque because they're packaged with an enclosure and backlight. But, I haven't really looked.

LCDs and TFTs are passive by their nature - they require backlight, unlike OLED where pixels are light emitters themselves.

Without backlight LCDs can't be used in night jumps. Also they would be harder to read while looking at some dark background even in bright daylight (I did some basic tests and something that's not glowing is in general harder to read). Basically it needs a backlight and that backlight adds extra thickness, and in this case every millimeter counts.

I've ordered this one with IPS display, and the whole unit is 8mm thick compared to 1.5mm OLED display (4mm when in 3Dprinted frame):

http://www.lilygo.cn/prod_view.aspx?TypeId=50033&Id=1126&FId=t3:50033:3

Project went silent, but not stalled :) There's been a lot of work behind the scenes and I'm releasing major code update tonight: https://github.com/RomanTheLegend/WingsuitGps

* Introduced support for TTGO device

* Fully redesigned architecture. Everything is modular: modes, devices, features. Each mode is a separate class so it's easy to develop new ones.

* You can switch between operation modes using menu

* Support for Bluetooth - you can stream GPS data to BASEline on your phone (friendly waving to @platypii )

It's still in Alpha state - code needs lots and lots of optimizations and stability improvements. I'm a novice C++ programmer, and many things done not the way they should. This will be fixed with time as I gain more experience. Meanwhile code review by someone experienced is really appreciated.

Next I'll be focusing on 3D models for the body, buttons for TTV and also rewiring stuff in more optimal way (pieces would be plugged to each other via JST) . First model would be tailored for Kiss helmets, other models to come later.

What else on the priority list:

* Flash GPS to higher baud rates (by default it does only 2 readings per second, I want 5 like in FlySight)

* Switch to UBX protocol which is faster than NMEA and more data-rich (FlySight also uses UBX)

* Refactoring

* Fonts optimization (currently font file for digits is too big, sometimes program fails to flash)

* Multi-page menu

Less priority:

* Settings

* Save data between device reboots

It's been a year since the last update, eh? Fear not, the project is still moving forward! Now that there's more free time in winter I hope to catch up on what was delayed.

Actually, there was another reason for delay - I switched to new display type and then spent countless hours banging my had against the wall trying to figure out why suddenly the image was distorted, but I can say with confidence it was worth the effort.

The new display solves the problem of blurry image as well as it lets me display more information at once. Also due to it's size it's nicely positioned in front of the eye and fits under visor (it'll look even cleaner and slimmer once all the boards are hidden)

One more benefit of the new display is adjustable focus, meaning people with vision complications can adjust it to their needs

raw changes: https://github.com/RomanTheLegend/WingsuitGps/tree/composite_display

Two primary tasks for the nearest future are

1) Improve the UI

2) Implement WS competition mode. Now that everything on the screen is clear and readable, I can implement visual gauges to display distance to the gates and deviation from the flight path. But this feature would take some time because reliably detecting the moment when competitor leaves the plane would be tricky. I think of implementing Kalman filter on data stream to detect true-positive downwards acceleration. If anyone has the experience with these please lend me a hand (preferably, in form of github merge requests :) )

P.S. I've been thinking of changing the project name to FlySense.

Why? Well, if FlySight doesn't display a thing, calling it FlySense sounds logical because you won't feel anything either.

P.P.S. The more I work with ESP32 platform the more awesome possibilities I'm discovering. Potentially, FlySense will be able to replace Sena/Cardo intercoms in the distant future https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-reference/network/esp-wifi-mesh.html#mesh-self-organized-behavior . Very distant future because the scope of work is huge

When you perfect this HUD, offer it to motorcycle riders. They will buy thousands and return your investment.

14 hours ago, riggerrob said:

When you perfect this HUD, offer it to motorcycle riders. They will buy thousands and return your investment.

I think this will depend on the size of the blind spot it creates...

Maybe, but that may prove be difficult https://www.youtube.com/watch?v=IiACV0Ly4j4

On the other hand, fellow FF trackers were keen about it's "Heading" functionality because apparently less experienced jumpers often loose their bearings and fly in the wrong direction.

what about this item?

Project Obsidian

https://www.aon2.co.uk/products/project-obsidian/

Obsidian has been in making for few years already (last news dates 2021), and it might still take them some time to get it to the market - pre-order summer 2023 with 100 units

I'd be keen to test it but I can tell you already it will not be a good match for competitors

from their own FAQ:
 

Quote

 

Does it have GPS?

No, Project Obsidian does NOT have a GPS receiver. It will be able to receive GPS data from another device (like a smartphone) via Bluetooth Low Energy.

 

Smartphone GPS is 1Hz which is not enough. I tried this frequency in my early prototypes and speed change feedback is too slow. 5Hz like in FlySight is optimal. What's more, depending on phone's firmware, at high vertical speeds some phones go crazy and stop sending their readings.

I'm wondering if Obsidian would be able to display anything else apart from plain numbers - majority of competitors I've spoken to don't actually care about ground speed or GL as much as they do about hitting the gates as precisely as possible to perfectly time the flare, and staying in the lane to avoid penalties.

BTW after a long fight with the library I've finally switched to UBX protocol and now can calculate all the data exactly as the FlySight does, so what you'll see on display will be exactly what will be in scoring on the ground

12 hours ago, the.Legend said:

Smartphone GPS is 1Hz which is not enough. I tried this frequency in my early prototypes and speed change feedback is too slow. 5Hz like in FlySight is optimal. What's more, depending on phone's firmware, at high vertical speeds some phones go crazy and stop sending their readings.

But there are GPS devices that have bluetooth that are NOT smartphones. I think they were just listing it as an example.

Actually, apps like Skyduck do quite a good job of recording skydives, maybe not for competition but okay for most cases.

The images for Project Obsidian also shows a direction arrow, and the specs list 480x640 display resolution, so it looks like it can display more than just numbers.

What about ActiveLook eyewear platform? It looks like they already did everything necessary. For example Engo 2 glasses are available and have connectivity to Garmin wearables/Android/iOs devices.
Also ActiveLook provides open API to build specific applications, but for me it looks like we even don't need app development - just connect them to Garmin watches and setup what you want to see on the display (needs confirmation from those who have these glasses)

So I found a good Black Friday deal and decided to order Engo 2 glasses. They arrived pretty fast and I did short tests with them. I can say that they fit my Kiss helmet nearly perfectly. Also I've connected the glasses not only with phone, but also with Garmin Fenix 7 watches and they work good (it is somewhat non-user-friendly to setup data fields for garmin watches, but absolutely doable). So glasses can show up to 6 data fields on a single page/screen, which I find good enough for WS flying. It is also possible to add more pages/screens, but switching from one screen to another needs hand gestures in front of the glasses, so it is definitely not usable for WS.
If you have any questions about the glasses, I can check and give you more details about it.