safe_rtl_viz

This repository is the first steps towards my 2017 GSoC project, "Safe Return-to-Launch for ArduPilot". The python script, visualizer.py, takes a standard .log file and animates the copter flying, and the result of the Safe-RTL algorithm. To test it out, run:

./visualizer.py ./logs/robert_lefebvre_octo_PM.log

To try it out with new log files, find a .bin file and use mission planner to convert the .bin file to a .log file (there's a button for that).

How the current version works:

• create an array that represents the path that we will fly to get back home.
• the first item in this array should be the home position
• Every time a new position comes in, check if our current position is more than two meters from the last item in the array
  • if it is, append the current position to the array
• If the array is almost full, run the following cleanup algorithm:
  • check if the path overlaps (or almost overlaps) with itself (pruning step)
    • if so, slice out any points between those.
  • If a point B lies between points A and C, remove B (simplification step)
    • uses the Ramer–Douglas–Peucker algorithm
• When RTL is triggered, run the cleanup algorithm once more, and then fly the path back home.