Improve Decawave localization

[JWCS]

@MOLLYBAS Did you have experimental data on the performance, std dev over different ranges? @MTlachac For R&D, did you want to buy more? They're not the cheapest as dev boards, but they're less problematic. I'm not where where they all are or how many we have.

[libBAS]

TODO:

• fix documentation on installing serial

4.3.9 API

• robust interfacing

• math

[JWCS]

At this point, given the popularity of decawaves, we could make it its own repo, I suppose, like the navx. I think if serial is installed as a submodule, it should work just fine?? Haven't tested. (Compare to when we weren't using ros, and had to install it globally as a normal library)

[libBAS]

The plan for setting up the class:

[JWCS]

I advised him against going a la C list, since the implementation is in C++; either to use a C++ container, or the (very) default vector, which pretty much removes the tag type entirely, returning a vector of anchors. It'll be cool when this is working smooth.

[libBAS]

TODO: Contain segfault when decawave disconnected/not connected. Node respawning/launch file

[JWCS]

Note the respawn argument: <node pkg="decawave" type="decawave_node" name="decawave_node" output="screen" respawn="true"/>

[libBAS]

Current status:

• Lower level node - publishes a group of distances at correct frequency, but cuts out for large amounts of time. Could be signal or software.

• Upper level node - math is written, need to put into node and test

[libBAS]

Through testing, the issue with the lower level node is likely the signal cutting out, can only mitigate so moving on to upper level.

[rfavour]

Update for feb 8th meeting: 1) position calculation is complete. correctly calculates x,y 2) publishes a stamped_pose_with_covarience msg, not sure which topic to publish to, but the formatting is good.
3) decawave node also publishes a "range msg" was going to be internal for calculation, but now does nothing 4) at present both tags anchors need to be updated at the same time for an estimate to be sent. change? 5) covarience numbers are a bit off, need to check matrix ops. 6) include dist from center of robot to dw into covar? 7) most of the arithmatic for position calculations occurs in the do_math function.