PI based ROv utilising..Node.Js..socket.io and HTML5 canvas

Anything to do with the Raspberry Pi
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

The problem I have now is as I'm a newbie on this html and Java script stuff I'm liek a kid in a sweet shop. Every time I read something I look under another stone and there's something else new to learn. My ROV front end has got more and more complex then back to basics and I think a clear simple one looks best.. I've now started to learn three.js and looked a bit at bootstrap but as I'll only be using this on probably the same PC or similar I don't think I need the facilities bootstrap offers.
I'm doing 3d type carousels for my axis setting at present.. If you've read my other thread on my 3d printer yo'll see my programming learning has stalled a bit trying to get that to work in the few minutes I have spare every day..
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

Image

This is my simplified version but soon to change....
User avatar
KR2_Diving
Posts: 391
Joined: Aug 30th, 2012, 11:43 am
Location: Currently: NW Suburbs of Chicago. Originally: NE Wisconsin

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by KR2_Diving »

Sorry... first... full disclosure... have not read the whole thread...

LOVE the mock up HUD! A little bit of extra fun is why this is called a hobby... not a job! :)
<><
rossrov
Posts: 383
Joined: Feb 28th, 2013, 5:01 pm
Location: Australia

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by rossrov »

Like the GUI, especially the buttons. Depth and heading hold should be no problem. I see PIDs in your future ;)
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

I'm working on the PIDs at the moment. I'm building a test screen where I can graph the output and simulate water current coming from the side and stuff to see what PID terms seem to be the most stable...
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

Ok scrap the PId simulator thing.. It is too difficult to try guess and replicate the real world. The max rate of change of the ROV I can guess at but how much it will go over the set point due to inertia and stuff ends up so much guess work that I'm sure my model wouldn't mean anything. unfortunately I need to finish the build on the ROV then plop in some values and tune the PId to get my desired results. I think it is a given that a PID is the best controller for this sort of thing so again my inaccurate simulator wouldn't tell me any thing new. Out of interest I did get a crude simulator up and running and got some graphs on the screen and stuff so it looked ok but was indeed useless. Getting my 3d printer to work decently is taking up too much ROV building time (and I thought it would speed it up !!).
rossrov
Posts: 383
Joined: Feb 28th, 2013, 5:01 pm
Location: Australia

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by rossrov »

A simulator would be great in that development and coarse-tuning could be achieved at the desktop, anywhere. But yes, a simulator would be a major job just on it's own. Alternatively a test tank (swimming pool?) in the back yard that one see from the PC inside the house, with some means of generating real-world currents etc. Lots of time with mathematics and programming or hire an excavator and build some pumps! On-the-fly real-world tweaking the go provided consideration is given for that from the outset.
Are the accelerometers even sensitive enough for this sort of navigation? Guess that is what you are trying to find out. Maybe a test carriage on a sloping track/ long piece of timber or tensioned wire, with movement damped to simulate water, and fit the accelerometer to that. Use a model aircraft prop instead of a "water" prop??
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

I'm fortunate in that I have a pool in the garden tis only small and unheated so I'll not be going anywhere near it till summer but I will be chucking the ROV in.
Last edited by perfo on Apr 17th, 2015, 2:15 pm, edited 1 time in total.
a_shorething
Posts: 289
Joined: Sep 10th, 2013, 5:26 pm
Location: New Jersey Shore

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by a_shorething »

I think this is the big question:
rossrov wrote:<snip>Are the accelerometers even sensitive enough for this sort of navigation? <snip>
Essentially the question is: At what point does it acclimate to 'stasis' In humans, balance adjusts very quickly to the environment we're in and I think the 9-DOF boards they have work more or less the same way. Assume you're in a current that is going N>S at 10 knots and you have no ability to use visual clues. Will the board be able to know that it's drifting relative to the bottom but is basically sitting still with respect to the water it's in? Will it ever adjust and reset to 'still' or will it always know that it's drifting?
perfo
Posts: 151
Joined: Jan 27th, 2015, 12:05 am

Re: PI based ROv utilising..Node.Js..socket.io and HTML5 can

Post by perfo »

Yep setting a heading or setting an end location are different things.. Just because you maintain a certain degrees of heading then you may not end up at the expected destination as the craft may have been crabbed by the current. I am expecting though that the current won't act on the whole craft at the same time and thus unless it's a prevailing or head on current it will try and rotate the ROV and this should be detectable and correctable. I'm not expecting it to be pin point accurate but as long as I know the direction it was heading when it left the Buoy and the distance from the buoy then I should be able to plot it's rough position on a screen. In any case what I'm trying to achieve is a replacement for a direct joystick connection.
A joystick would be constantly adjusted by the pilot to keep the ROV heading towards something. If that thing is visible then great but if not then it is probably going to be done against the returned compass data. So id being doing the same with the exception of I don't need to constantly adjust the joystick. I just aim the ROV at the object/location (if it's visible) or adjust it so its facing a certain heading and press a button and the on-board controller would look after any adjustments. So if the compass feed back is so vague I have to end up correcting every now and then, I'm no worse off than doing it all manually anyway...so really the accuracy of the sensors don't affect my plan any more than a manually operated one but stand a chance to be a bit better.
Post Reply