Support remote update of lighthouse calibration

[geonnave]

The plan:

1. swarmit -c ../swarmit-argus.toml -n A001 calibrate ~/.dotbot/calibration.out
2. the device receives it and stores on flash
3. once it's done, the device needs to be rebooted for the new calibration to apply

Regarding the transfer, my approach is to send the homography matrices one by one, the packets will be like this:

• count: the total count of homography matrices
• index: the index of the current homography matrix
• homography: the actual matrix data with 9 int32_t items

The calibration is stateless on the device side: every time it receives a calibration packet, it writes count to the flash config area, as well as the the homography matrix (at the correct ìndexas indicated in the packet). This means that the samecount` will be written several times (one for each packet), but there is no harm on that.

---

Update: this now works! Remains to be tested with a real lighthouse though.

How to test the remote calibration update:

swarmit -c ../swarmit-argus.toml calibrate-lh2 <calibration file>

I also used these snippets to generate some dummy calibration files:

python3 -c "import struct; m=[[1,1,1],[111000,9,-111000],[1,999000,-1]]; \
  open('calib111.bin','wb').\
  write(bytes([1]) + b''.join(struct.pack('<i',v) for r in m for v in r))"
# generates a matrix like [[1,1,1],[111000,9,-111000],[1,999000,-1]]

python3 -c "import struct; mats=[[[k,k,k],[k,k,k],[k,k,k]] for k in (1,2,3,4)]; \
  open('calib1234.bin','wb').\
  write(bytes([len(mats)]) + b''.join(struct.pack('<i',v) for m in mats for r in m for v in r))"
# generates 4 matrices like [[k,k,k],[k,k,k],[k,k,k]]

[geonnave]

With the busy days of IETF meeting, visit and talk at HKUST-GZ, adjusting the Mari paper, contributing to OpenSwarm deliverables... I could not advance this PR. I think this is really important as @AmyMcMahon2 has a demo for early/mid April, however I will not be able to work on it next week as I will be on PTO. @aabadie if you have some time, could you please take a look? All the plumbing is here and the principle works, the only missing part is finding out why the device reboots/crashes when I write to the flash.