How can I find the instantaneous directional vector of 3D dubins vehicle?

Artificial Intelligence Programming

Started by lucky6969b October 25, 2016 04:33 AM

3 comments, last by lucky6969b 7 years, 5 months ago

lucky6969b

1,334

Author

October 25, 2016 04:33 AM

http://www.et.byu.edu/~beard/papers/preprints/BeardMcLain__.pdf

I've figured out how to find out the instantaneous position of the vehicle.

But Will the direction be the delta of the current position and next position?

I am not sure, is there a reliable way to do it?

Of course, I already have the start direction and some ending directions

in the spirals and straight lines.

For example, when the vehicle starts at radian 4.7, with a left turn

can I assume at time 0, it is at 4.7 and at time 1, it will be at 4.6.....(left-handed system)

The elevation isn't taken into account anyways....

Thanks

Jack

ferrous

6,164

October 25, 2016 04:48 PM

If you're not taking the elevation into account, the direction would be the tangent line of the circle. I'd have to poke around more on how you'd get the Yaw from a helix.

ApochPiQ

23,138

October 25, 2016 09:42 PM

It should just be the first-order derivative of the equation you use to get positions. There are online tools to do differentiation if your calculus is rusty :-)

Wielder of the Sacred Wands
[Work - ArenaNet] [Epoch Language] [Scribblings]

alvaro

21,604

October 26, 2016 09:54 AM

As a first approximation, take the positions at times t+0.001 and t, subtract them and multiply the resulting vector by 1000. This is an inelegant approximation, but it's really easy to implement and to get right. If you end up using a fancier method, you can still use the inelegant method to verify that your fancier code is correct.