drake-crazyflie-tools

A collection of tools to develop controllers for the Crazyflie using Drake

Installation

• install the lcm types

    cd lcm
    sudo python setup.py install

• install the client

    cd crazyflie-clients-python
    sudo bash setup.sh

• install the estimator

    cd estimation
    sudo python setup.py install
    sudo pip install filterpy

• flash the firmware

    cd crazyflie-firmware-2
    make cload

• install vicon

    cd vicon
    mkdir bin
    mkdir obj
    cp lib/libViconDataStreamSDK_CPP.so /usr/lib/libViconDataStreamSDK_CPP.so
    make

Main Usage

• run the vicon client to get sensor data

    vicon/bin/vicon_lcm_client objectname

• run the crazyflie client

    ./simpleclient

• You can send lcm commands from the nanokontrol board

    ./nanokontrol

• You can also use drake to send commands ("crazyflie_input" channel), and therefore build pretty sophisticated controllers. See Drake's documentation for more details.

LCM Types

You can regenerate the lcm types if you have lcm installed.

• Install LCM (lcm-gen)
• add lcm.jar to CLASSPATH
• run lcm/make_types

Handling Logs

How to use LCM logging with MATLAB:

• Take your logs. On a computer that is recieveing the relavent LCM messages, run

    $ lcm-logger

in the directory you want your log files. This will produce a file that contains all your LCM messages.

• Export to .mat format

    $ python log_to_mat.py -f -l lcmtypesmodule,anotherlcmtypemodule lcmlogfile

Running the visualizer

After running make -j in the drake-distro/director folder:

drake-distro/build/bin/ddConsoleApp -m crazyflieviewer

you can play a log back

lcm-logplayer -s .5 lcmlog-2015-04-02.00

and visualize it in matlab

addpath_crazyflie
cf=Crazyflie()
cf.visualizeStateEstimates()