Adds unit test for imu preintegration of a single step

[raabuchanan]

Related to issue #627

This adds a unit test which integrates simulated data from an IMU in a foot while to foot takes a single step. See issue for related dataset.

Currently the unit test actually passes which is in contradiction to what I am seeing in the issue. I need to investigate the problem further but wanted to put up this unit test to keep everyone in the loop. Also someone may spot something wrong that I'm doing. I will write more unit tests to try and understand the problem.

[varunagrawal]

I believe we can merge all 3 files into 1 so we have a comprehensive test file that is easy to modify. While I like the structure a lot, it seems like a lot of added complexity.

[varunagrawal]

@raabuchanan I've made some updates to the code to make it agnostic to users. I'd love to work with you on this PR and figure this issue out.

Currently, I am a bit confused on what are we testing for specifically? Is there a specific assertion you want to test for?

[dellaert]

@varunagrawal , please help land? Also, don’t force-push :-)

[varunagrawal]

Since @raabuchanan got the Imu impact working, what should we be testing for? The same thing as before?

@raabuchanan please could you refresh my memory on the goal of this test? I believe you had a slide on it.

[raabuchanan]

Hi @varunagrawal yes I fixed the issue on my side and this passes now. This test checks a sequence of simulated IMU data that the relative Z position is zero. It also checks that there are no rotations. I will push a small cleanup soon but I can add more documentation if you want.

[varunagrawal]

Awesome. I'll clean it up a little bit and then we can merge.

[varunagrawal]

Very minor comments, but LGTM.