Optical flow support

[digitalentity]

This is to discuss and review the code. @martinbudden I'd really apreaciate your comments and suggestions on this one.

Tasks to do:

• OpFlow detection code in sensor autodetection code
• OpFlow task to periodically read and collect the data
• ADNS3080 driver skeleton for SoftSPI
• ADNS3080 detection code
• ADNS3080 initialisation code
• ADNS3080 read code
• Add OpFlow support to MSP
• Add OpFlow support to position estimation code and navigation functions

[martinbudden]

Perhaps deltaX and deltaY? Or a union of the two?

[digitalentity]

A union is good. Will do.

[martinbudden]

Well this all looks pretty straightforward so far.

The interesting task is, of course, "Add OpFlow support to position estimation code and navigation functions".

And the soft SPI fixes are a bit hacky, but I guess we can live with that for the moment.

[digitalentity]

You mean when doing delays? SoftSPI by itself is hacky 😄 Until we have a proper abstraction of Bus layer we'll probably have to live with hacky stuff.

Yes, OpFlow navigation is gonna be a tough one, especially knowing that we need sonar to measure real altitude and calculate correct velocities. Lack of position data from OpFlow is also a problem, so estimated position will probably be locked to (0;0) w/o GPS, but velocity control will allow PosHold and CRUISE mode as well.

[martinbudden]

I think we should get into the habit of forward declarations in header files, eg

struct opflow_s;
bool opflowADNS3080Detect(struct opflow_s *opflow);

[jflyper]

Any good source of ADNS3080?

[digitalentity]

eBay - http://www.ebay.com/itm/-/141282083588
AliExpress - http://www.aliexpress.com/item/-/32224579616.html
BangGood - http://www.banggood.com/APM-2_5-2_6-2_8-Optical-Flow-Sensor-Optical-Flow-V10-No-GPS-Point-p-971679.html

[jflyper]

Cool. Thanks!

[martinbudden]

@digitalentity , can you rebase this?

[digitalentity]

Sure.
At the moment I don't have much time to spend on this but the project is not abandoned 😄

[digitalentity]

@martinbudden rebased. I would appreciate if you would look at initialisation part of drivers\opflow_adns3080.c - this is the most hacky part of code at the moment.

[johnsilvester]

Hi guys! Is this still in progress?

[digitalentity]

This is a very slo-mo task, sorry. Eventually we'll see opflow support, but probably not very soon.

[johnsilvester]

Wish I could help. I'll try to do a little research and maybe speed up the progress

[martinbudden]

@digitalentity , can you rebase this. I know it's on hold at the moment, but I think it's a good idea to rebase it now otherwise rebasing will get increasingly difficult.

[johnsilvester]

Could someone explain to me how far into the development we are? I see the check marks but I'm a little confused as to what needs to be done. Thanks!

[martinbudden]

Typo, presumably this should not be PITOT_NONE.

[digitalentity]

OMG, after PGs are done this will be one hell of a rebase.

[digitalentity]

@johnsilvester I've barely scratched the surface here. The hardware was detected and read properly (at least several rebases ago), but no actual sensor fusion is performed.

[johnsilvester]

@digitalentity Cool! Thanks for all of the progress so far!

[MengHaoyu]

does this patch work？I want to use it

[digitalentity]

@MengHaoyu no, it doesn't. I have this on low priority.

[MengHaoyu]

ok,thanks,i bought an a3080 sensor, i thought i could use it 😬

[digitalentity]

Will be replace with new PR