Metadata-Version: 2.1
Name: flightanalysis
Version: 0.1.0
Summary: A package for analysing flight data
Home-page: https://github.com/PyFlightCoach/FlightAnalysis
Author: Thomas David
Author-email: thomasdavid0@gmail.com
Description-Content-Type: text/markdown
License-File: LICENSE

# FlightAnalysis

This package contains tools for analysing flight log data. It is based around the State and Section objects, which handle state information for an instant and a section of time. It also contains tools for constructing template Section and State instances based on serialised aerobatic manoeuvre definitions. Temporal alignments can be performed to automatically label flight data based on these templates. New scaled templates can then be constructed based on measurements taken of the labelled elements. 

The Main Idea
1. Read the sequence defintion from the serialisation of it (currently schedule/p_21.py)
2. Read, trim and rotate flight recorded flight data
3. generate a roughly scaled template sequence
4. run a dynamic time warping algorithm to align the flight to the template
5. generate scaled templates based on the rules and the flown manoeuvre
6. Expose an environment for grading criteria to be encoded based on rules of the aerobatic competition disciplines

# Side Projects
- Estimate winds from flight data without airspeed sensor. see examples/wind.ipynb
- Construct a FD model from the flight data with MLP regression or something.
- The templates are constructed in some kind of 'Judging' axis, similar to the aircraft wind axis but with atmospheric wind removed. Consider accounting for some aerodynamic characteristics here and constructing templates at some alpha and beta.
- Compare logged control inputs to ideal control inputs calculated using the templates and the FDmodel.


# External Dependencies:
python (3.6+?)
numpy
pandas
scipy
fastdtw

## This package also depends on other PyFlightCoach packages:
ardupilot_log_reader
flightdata
geometry
