Provide the capability to generate a mission from a flight log

[stronnag]

Provide the capability to generate a mission from a flight log

[stronnag]

It will also be possible just to give the number of mission points required and let the application iterate a few epsilon values to give the desired number of points (subject to some sanity criteria, like greater than say 8, less than 60).

[stronnag]

This is probably the final iteration for now:

log2mission

log2mission will create an inav XML mission file from a supported flight log (Blackbox, OpenTX, BulletGCSS). The mission will not exceed the inav maximum of 60 mission points.

$ log2mission
Usage of log2mission [options] file...
  -end-offset int
    	End Offset (seconds) (default -30)
  -epsilon float
    	Epsilon (default 0.015)
  -index int
    	Log index
  -interval int
    	Sampling Interval (ms) (default 1000)
  -mode-filter string
    	Mode filter (cruise,wp)
  -rebase string
    	rebase all positions on lat,lon[,alt]
  -split-time int
    	[OTX] Time(s) determining log split, 0 disables (default 120)
  -start-offset int
    	Start Offset (seconds) (default 30)

• The start-offset and end-offset compensate for the fact that the start / end of the flight is usually on the ground, and thus is not a good WP choice. The defaults are 30 seconds for the start offset and -30 seconds (i.e. 30 seconds from the end) for the end offset. The end offset may be specified as either a positive number of seconds from the start of the log or a negative number (from the end). Locations prior to the start offset and after the end offset are not considered for mission generation. If the end-offset is specified (0 cancels it), and there is no flight mode filter, then RTH is included in the generated mission.
• The mode-filter allows the log to filtered on Cruise and WP modes, e.g. -mode-filter=cruise, -mode-filter=wp, -mode-filter=cruise,wp. If mode-filter is specified, log entries not in the required flight mode(s) are discarded. Cruise includes both 2D and 3D cruise.

epsilon tuning

The epsilon value is an opaque factor that controls the point simplification process (using the Ramer–Douglas–Peucker algorithm). The default value should be a good starting point for fixed wing with reasonably sedate flying. On a multi-rotor in a small flight area, a much smaller value (e.g. 0.001) would be more appropriate. Increasing the value will decrease the number of mission points generated. log2mission will do this automatically if the default value results in greater than 60 mission points, for example: the log here would generate 77 points with the default epsilon value.

$ log2mission -start-offset 60 -end-offset -120 /t/inav-contrib/otxlogs/demolog.TXT
Flight   : MrPlane on 2021-04-08 13:24:07
Firmware : INAV 3.0.0 (fc0e5e274) MATEKF405 of Apr 7 2021 / 17:02:08
Size     : 19.36 MB
Log      : demolog.TXT / 1
Speed    : 28.0 m/s at 13:54
Range    : 17322 m at 14:22
Current  : 30.6 A at 00:05
Distance : 48437 m
Duration : 43:44
Altitude : 292.8 m at 25:42
Mission  : 56 points (reprocess: 1, epsilon: 0.018)

The output from this example would be demolog.1.mission

multirotor example

Using a old, contributed MR log, in quite a small area, with user specifed epsilon.

$ log2mission -epsilon 0.001 logfs.TXT
Log      : logfs.TXT / 1
Flight   :  on 2019-02-08 15:21:13
Firmware : INAV 2.1.0 (7bdd5967e) OMNIBUSF4V3 of Jan 22 2019 09:39:17
Size     : 32.03 MB
Current  : 23.5 A at 02:21
Distance : 1560 m
Duration : 04:10
Altitude : 52.5 m at 02:50
Speed    : 17.3 m/s at 02:38
Range    : 174 m at 01:22
Mission  : 13 points

13 points is a adequate mission to reproduce the flight.

Using an extreme user defined epsilon results in an excessive number of points:

$ log2mission -epsilon 0.00001 logfs.TXT
...
Mission  : 59 points (reprocess: 8, epsilon: 0.000105)

Whereas, with the default epsilon of 0.015, no useful mission is generated:

$ log2mission logfs.TXT
...
Mission  : 2 points

So some experimentation may be required to get a good mission, particularly for shorter MR flights. In particular, if reprocessing
is indicated and the number of generated points is close to 60, then it's probably worth running again with a slightly larger epsilon than that shown in the output.

pre-release binaries, Linux x86_64 and Win32

fl2kml_0.98-rc3.tar.gz

[stronnag]

Almost the last!
Attempt to resolve short distance flights otherwise giving two point missions.
fl2kml_0.98-rc4.tar.gz