Fix culling and delay implementation

[mxgrey]

During intensive long-run testing, it was discovered that bad inputs from schedule participants can lead to harmful behavior in the traffic schedule. This PR helps in two ways

• Excessive cumulative delays for schedule participants trigger the creation of new routes within the timeline storage so that older route information can be safely culled in a reasonable time.
• The Participant API was changed to make it much easier to correctly set the cumulative delay, even in heavily async code.

[codecov]

Codecov Report

Merging #86 (c79026c) into main (dc17527) will decrease coverage by 0.07%.
The diff coverage is 19.25%.

@@            Coverage Diff             @@
##             main      #86      +/-   ##
==========================================
- Coverage   23.24%   23.17%   -0.08%     
==========================================
  Files         370      370              
  Lines       30751    30768      +17     
  Branches    14369    14372       +3     
==========================================
- Hits         7147     7129      -18     
- Misses      14106    14166      +60     
+ Partials     9498     9473      -25     

Flag	Coverage Δ
tests	23.04% <19.25%> (-0.07%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
..._traffic/include/rmf_traffic/schedule/Database.hpp	0.00% <ø> (ø)
...affic/include/rmf_traffic/schedule/Participant.hpp	0.00% <ø> (ø)
..._traffic/agv/planning/DifferentialDrivePlanner.cpp	44.86% <0.00%> (ø)
.../src/rmf_traffic/schedule/internal_Participant.hpp	100.00% <ø> (ø)
rmf_traffic/test/unit/agv/test_Negotiator.cpp	10.83% <0.00%> (-0.04%)	⬇️
rmf_traffic/test/unit/schedule/test_Mirror.cpp	14.06% <0.00%> (-0.36%)	⬇️
...mf_traffic/test/unit/schedule/test_Participant.cpp	11.27% <3.84%> (-1.26%)	⬇️
rmf_traffic/test/unit/schedule/utils_Database.hpp	17.85% <18.18%> (-3.36%)	⬇️
rmf_traffic/src/rmf_traffic/schedule/Database.cpp	49.12% <20.00%> (-3.52%)	⬇️
...f_traffic/src/rmf_traffic/schedule/Participant.cpp	52.57% <52.63%> (-5.72%)	⬇️
... and 6 more

[luca-della-vedova]

Been doing a series of long high intensity tests and indeed this set of PRs helps increase the reliability a lot, thanks!