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 setup.py install_requires

   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