Thats why it is different than a drone flight controller. Imagine a Sport car in automated gearbox and comfort mode . Things are smooth traction control acceleration … then you switch to sport mode, manual gearbox and full control of the car. In hydrofoil I think controller will just smooth out user inputs or external impacts to avoid crashing. Imagine a nee user curving with a low speed, control system will just boost throttle to minimum possible speed to be able to do that curve.
Another scenario, you are going straight, but there is a sudden side wave, with a little roll control and little throttle increase flight control will help avoiding crashing. Assisted foil would Better word for it
1 Like
I evaluated betta flight code and hw. Hw hs everything we need. Although there is lots of overlaps in Firmware I think control concept is different, plus I found firmware it so convoluted that if I just write my own or just use modules from it would be easier and faster.
You missing the point completely.
In a car you have manual input controls for all those automated variables (steering wheel, accelerator, gears, brakes etc). The automation works because they can assist all those manual input commands. In foiling there is no manual input command except for throttle. You can’t compare shifting your weight to a steering wheel, they are entirely different mechanisms of control. The second you lean in any direction your board will be fighting back.
How is the controller going to distinguish between a sudden wave and a sudden turn that has the same force and angle? I’ll tell you…IT WON’T, it will react the same way in both situations and ultimately force a crash. There’s a reason there’s no electronic stabilisation on a skateboard 
It won’t and it will damp user inputs as well to avoid crashing to some degree. And if user wants full agility and control just needs to get out of assisted mode. Any inputs from user or outside that is going to lead to crash will be damped to make it smooth ride rather than sporty ride. I also see your point but we shouldn’t be so biased to new ideas before testing and fully evaluating them. Isn’t it why we all here because someone believed that they can make foil electric?
3 Likes
I understand the concept of what you are trying to do, its just that the logic is flawed.
eg. Honda has assisted balancing on one model of their motorbikes, but the steering is still controlled by a steering wheel. When you turn the wheel it lets the control algorithms know that the user wants to turn and hence adjusts the algorithms accordingly.
In a standard efoil that’s not possible if you have balance assist, unless you add a joystick (steering wheel) to let the balance assist know you turning and not to fight back. But then you may as well add a rudder…
Simplifying the balance issues on an efoil are quite easily done by changing to more stable wings. Why complicate something simple with more underwater electronics, actuators or levers. That will just add to the cost and complexity of an already expensive product.
What you want to do is not new in any way, its been done on all sorts of other vehicles, its just that all of the vehicles have another input source to tell the system when you are turning…
You sound like my grand pa that complains about expensive over complicated self driving tesla . I get it if you are saying control is hard and doesn’t worth it. My argue is it worth it if it can make it easy for normal riders to enjoy this efoil even in wavy waters. This actually can bring more users and average consumers. Which will expand the market drastically and thst means lower price. How? Price of a product mostly depends on demand not the complexity of it. Dji mavic drone’ 10 years a go you could not even get something close to it for $100k. Efoil can easily get close even below 3000$ if tje market cap is big enough.
And the logic is not flawed. You are just looking at it from a wrong angle. It doesnt need a joystick it has all information from imu and user input throttle. Any inputs from user or externally injected ( waves or winds or whatever) it will get smoother as it gets close to the limits of going out of balance.
3 Likes
Please tell me how your algorithms will distinguish between a minor turn ie. a very slight lean from the user OR a very slight twitch in balance OR a wave interference? All of this with the throttle is set to cruise control mode…
I say it’s going to see all three of those the same way, unless you have an additional control input other than throttle to tell it not to.
Smoothing foiling means slowing down any natural reaction in the case of emergencies… Therefore making it more dangerous!
I think I answered it already! My. Initial theory is control loop will react to both in a same way and will smooth and damp it stronger as it gets to the tripping point. But we won’t know the whole scope until we simulate it and test it. I see your point but for me with an engineering brain it is another problem That I have to find a solution for it.
1 Like
Assuming I am not an engineer shows your nativity to the problem…
I have a masters degree in Control Theory (an undergrad degree in electronics and I also taught control at Uni while doing the masters). I have also designed and built boats, surfboards, foil boards etc… I know my stuff
I know what you are trying to do with the smoothing, but all a smoothing algorithm will do is slow down user (balance based) input and thus make the foil less reactive. This in turn will force the user to put in much greater input than needed when they actually desire to turn, leading to being off balance when the algorithm eventually does let them turn. Imagine an algorithm telling a car to brake less than what the user has inputted. It’s going to lead to a crash every time unless you are the only car in the middle of a flat desert…
I’m not trying to stop you, I’m just warning you that you are complicating a problem that has already been widely solved in a much simpler manner…
Well It can be a nice feature , but I think I will be hard to evaluate if the balance moves are done on purpose or not and in 3D , no?
On a road yes , but on moving water
Guys, I totally agree with @samisin, let’s not dismiss and rule out ideas before we test and try. It’s a hard problem yes. I don’t think everything here has been widely solved yet. Many smart engineers here and I am a strong believer that together we can make progress on stabilization. There are many stages in between today’s pure weight-shift control to full self stabilized obstacle avoiding autonomy. Not everyone here is interested in all this and that’s fine, but let’s please keep the momentum and discussion going.
First, let’s start with pitch stability. But regarding roll stability, intuitively I would start with a low pass filter. Twitch and wave interference result in high frequency inputs. Making a turn (= shifting weight left or right) is a low frequency input. Active user inputs will be dampened but with the right tuning I think they can still be responsive enough to ensure full control and safety.
2 Likes
Thats why I never got along with university teachers , so religious in to their books. I am surprised how come someone that expert in control theory could not understand a control system that for example controls the car to not flip over in the curves even if either user puts too much steering or car by itself inters the curve so fast! When you said it is impossible I should have known there is a teacher behind it. you lost me right there!
A possibility I can think of, that may be a little advanced but help with the stability control is an IMU that the rider wears on their chest. Would have to be battery powered and wirelessly communicate to the board - definitely doable. This IMU data would be compared with the IMU data on the board and this would make it possible to distinguish between unwanted board movement that should be corrected and movements the rider wants the board to do. I too think that this stability control would be great for the sport and making efoils easier to ride and also making it possible to push them into more serious manoeuvres like sharp turns and jumps etc.
1 Like
To start with anyway, I think simple pitch control that regulates motor speed based off board pitch angle would be a very good help. A lot of first timers I’ve put on the boards struggle with controlling the throttle to help stay balanced while also balancing on the foil simultaneously.
1 Like
Hahaha…so many assumptions…
You claim I lost you but then you bothered replying so clearly I didn’t lose you
There’s a super easy way to solve both these issues. They been solved a looong time ago. A wider wing and a longer fuselage. They also have the advantage of being cheaper with less to fail…
If you want to do more serious manoeuvres like sharp turn and jumps etc, you need to maximise weight loss and optimise your wings. Look at the stuff Kai does and rides and then another great rider is https://instagram.com/mizofernando?igshid=1i7g9uqj5gakd
Unifoil have some fantastic wings which are super stable but still pretty quick Performance Hydrofoils on Instagram: "#Repost @misterbennetts • • • • • • Byron Bay, New South Wales Legit foil porn, cannot wait to ride these bad boys over the coming days... thank you @unifoilaustralia 🦅 @unifoil @foiltheworld @eastcoastfoilclub_aus #foiltheworld #foilsurfing #loyaltothefoil #hydrofoil #surffoil #byronbay #unifoil"
Why should he not try his way?
I got traction control at my skateboard, so where is your god now? 
The point is, theoretically you got a 3 degree of freedom system (1 direction, 1 speed, 1 height) which you can’t control with only 1 user input (the throttle at the remote).
So you have to make assumptions and add sensors (or do a 3 channel remote, but nobody wants that because it is even more difficult than foiling in the first place).
So one assumption would be e.g. constant height.
You are right, that wont work all the time, e.g. not in small waves, as well as your traction control in your car will fail in lose sand (it will make the car digging a hole instead helping you out).
You just have know when to use which assistance system and how to stop it. E.g. a simple dead man button on the remote could solve all of these issues…
Maybe we add computer vision and machine learning to train and optimize for different wave or obstacle conditions. Something like the NVidia Jetson Nano 
Lol I don’t see a reason to not doing it. At least obstacle avoidance with sonar seems very easy to do