px4tools

Anaconda

Linux: OSX: Windows: Version: Downloads:

A log analysis toolbox for the PX4 autopilot written in python.

Features

• Flight Plotting using standard python libraries.
• Automatic System Identification from log data.
• Automatic Control Design from log data.
• Cross-Platform deployment, testing, and support (Windows/OSX/Linux).
• Well integrated with Jupyter notebook and Pandas.
• Natively uses pandas CSV format so easily integrated with all log formats.

Usage

1. px4log format

   First use the sdlog2_dumpy.py program to convert the px4log to csv:

   wget https://github.com/PX4/Firmware/raw/master/Tools/sdlog2/sdlog2_dump.py
   python sdlog2_dumpy.py your_log.px4log > your_log.csv

   Now start jupyter notebook in the directoy of your_log.csv:

   jupyter notebook

2. ulog format

   No preprocessing required, see ulog example.

Examples:

1. Automatic System Identification and Control Design
2. General Flight Data Plotting
3. ULOG basic use.
4. ULOG noise analysis.

Install

Dependencies

1. See requirements.txt

   For pandas, to fix time series plotting memory issue with time delta index you need this branch:

   . conda_env (See instructions below for setting up conda_env script)
   git clone [email protected]:jgoppert/pandas.git
   cd pandas
   git checkout tdi_plot_fix
   python setup.py install

Using Anaconda (Recommended)

1. Install anaconda

   • Python 3 version recommended
   • Do not select add to path if you use ROS. ROS expects the standard python to be installed. You can create a script to source and add anaconda to your path. This is similar to setup.bash for ROS users.

   ~/bin/conda_env:

   #!/bin/bash
   export PATH=$HOME/anaconda3/bin:$PATH

   Now you can source the script to start using anaconda instead of the sytem python:

   . conda_env

2. Install px4tools via conda

   conda config --add channels conda-forge
   conda install px4tools jupyter

3. Upgrading px4tools using pip in conda

   The conda-forge px4tools package usually lags behind master. If you need the latest code, use pip within your conda env.

   . conda_env
   pip install px4tools

4. Building px4tools form source and installing to conda environment

   If you want to do development and edit some of the source, follow this example:

   . conda_env
   git clone [email protected]:dronecrew/px4tools.git
   cd px4tools
   python setup.py build install

Using PyPi

pip install px4tools jupyter --user