"ardrone_autonomy" is a ROS driver for Parrot AR-Drone quadrocopter. This driver is based on official AR-Drone SDK version 2.0 and supports both AR-Drone 1.0 and 2.0. "ardrone_autonomy" is a fork of AR-Drone Brown driver. This package has been developed in Autonomy Lab of Simon Fraser University by Mani Monajjemi.

• September 5 2012: Experimental automatic IMU bias removal.
• August 27 2012: Thread-safe SDK data access. Synchronized navdata and camera topics.
• August 20 2012: The driver is now provides ROS standard camera interface.
• August 17 2012: Experimental tf support added. New published topic imu.
• August 1 2012: Enhanced Navdata message. Navdata now includes magnetometer data, barometer data, temperature and wind information for AR-Drone 2. Issue #2
• July 27 2012: LED Animation Support added to the driver as a service
• July 19 2012: Initial Public Release

This driver has been tested on Linux machines running Ubuntu 11.10 & 12.04 (32 bit and 64 bit). However it should also work on any other mainstream Linux distribution. The driver has been tested on both ROS "electric" and "fuerte". The AR-Drone SDK has its own build system which usually handles system wide dependencies itself. The ROS package depends on these standard ROS packages: roscpp, image_transport, sensor_msgs, tf, camera_info_manager and std_srvs.

The installation follows the same steps needed usually to compile a ROS driver.

• Get the code: Clone (or download and unpack) the driver to your personal ROS stacks folder (e.g. ~/ros/stacks) and cd to it. Please make sure that this folder is in your ROS_PACKAGE_PATH environmental variable.

    ```bash
    $ cd ~/ros/stacks
    $ git clone https://github.com/AutonomyLab/ardrone_autonomy.git
    $ rosstack profile && rospack profile
    $ roscd ardrone_autonomy
    ```
  

• Compile the AR-Drone SDK: The driver contains a slightly patched version of AR-Drone 2.0 SDK which is located in ARDroneLib directory. To compile it, execute the ./build_sdk.sh. Any system-wide dependency will be managed by the SDK's build script. You may be asked to install some packages during the installation procedure (e.g daemontools). You can verify the success of the SDK's build by checking the lib folder.

    ```bash
    $ ./build_sdk
    $ [After a couple of minutes]
    $ ls ./lib
  
    libavcodec.a   libavformat.a    libpc_ardrone_notool.a  libvlib.a
    libavdevice.a  libavutil.a      libsdk.a
    libavfilter.a  libpc_ardrone.a  libswscale.a
    ```
  

• Compile the driver: You can easily compile the driver by using rosmake ardrone_autonomy command.

The driver's executable node is ardrone_driver. You can either use rosrun ardrone_autonomy ardrone_driver or put it in a custom launch file with your desired parameters.

The driver's main loop is being executed in 50Hz, however the data publish rate is device and configuration dependent. Basically, the data will be published when a new data arrives. For example, navdata and imu update frequencies is 15Hz if navdata_demo parameter is set to true, otherwise it will be 50Hz.

Information received from the drone will be published to the ardrone/navdata topic. The message type is ardrone_autonomy::Navdata and contains the following information:

• header: ROS message header
• batteryPercent: The remaining charge of the drone's battery (%)
• state: The Drone's current state: * 0: Unknown * 1: Inited * 2: Landed * 3,7: Flying * 4: Hovering * 5: Test (?) * 6: Taking off * 8: Landing * 9: Looping (?)
• rotx: Left/right tilt in degrees (rotation about the X axis)
• roty: Forward/backward tilt in degrees (rotation about the Y axis)
• rotz: Orientation in degrees (rotation about the Z axis)
• magX, magY, magZ: Magnetometer readings (AR-Drone 2.0 Only) (TBA: Convention)
• pressure: Pressure sensed by Drone's barometer (AR-Drone 2.0 Only) (TBA: Unit)
• temp : Temperature sensed by Drone's sensor (AR-Drone 2.0 Only) (TBA: Unit)
• wind_speed: Estimated wind speed (AR-Drone 2.0 Only) (TBA: Unit)
• wind_angle: Estimated wind angle (AR-Drone 2.0 Only) (TBA: Unit)
• wind_comp_angle: Estimated wind angle compensation (AR-Drone 2.0 Only) (TBA: Unit)
• altd: Estimated altitude (mm)
• vx, vy, vz: Linear velocity (mm/s) [TBA: Convention]
• ax, ay, az: Linear acceleration (g) [TBA: Convention]
• tm: Timestamp of the data returned by the Drone

The linear acceleration, angular velocity and orientation from the Navdata is also published to a standard ROS sensor_msgs/Imu message. The units are all metric and the reference frame is in Base frame. This topic is experimental. The covariance values are specified by specific parameters.

Both AR-Drone 1.0 and 2.0 are equipped with two cameras. One frontal camera pointing forward and one vertical camera pointing downward. This driver will create three topics for each drone: ardrone/image_raw, ardrone/front/image_raw and ardrone/bottom/image_raw. Each of these three are standard ROS camera interface and publish messages of type image transport. The driver is also a standard ROS camera driver, therefor if camera calibration information is provided either as a set of ROS parameters or appropriate ardrone_front.yaml and/or ardrone_bottom.yaml, the information will be published in appropriate camera_info topics. Please check the FAQ section for more information.

• The ardrone/* will always contain the selected camera's video stream and information.

• The way that the other two streams work depend on the type of Drone.

  • Drone 1

  Drone 1 supports four modes of video streams: Front camera only, bottom camera only, front camera with bottom camera inside (picture in picture) and bottom camera with front camera inside (picture in picture). According to active configuration mode, the driver decomposes the PIP stream and publishes pure front/bottom streams to corresponding topics. The camera_info topic will include the correct image size.

  • Drone 2

  Drone 2 does not support PIP feature anymore, therefore only one of ardrone/front or ardrone/bottom topics will be updated based on which camera is selected at the time.

The Navdata message also returns the special tags that are detected by the Drone's on-board vision processing system. To learn more about the system and the way it works please consult AR-Drone SDK 2.0's developers guide. These tags are being detected on both drone's video cameras on-board at 30fps. To configure (or disable) this feature look at the "Parameters" section in this documentation.

The detected tags' type and position in Drone's camera frame will be published to the following variables in Navdata message:

• tags_count: The number of detected tags.
• tags_type[]: Vector of types of detected tags (details below)
• tags_xc[], tags_yc[], tags_width[], tags_height[]: Vector of position components and size components for each tag. These numbers are expressed in numbers between [0,1000]. You need to convert them back to pixel unit using the corresponding camera's resolution (can be obtained front camera_info topic).
• tags_orientation[]: For the tags that support orientation, this is the vector that contains the tag orientation expressed in degrees [0..360).

By default, the driver will configure the drone to look for oriented roundels using bottom camera and 2D tags v2 on indoor shells (orange-yellow) using front camera. For information on how to extract information from tags_type field. Check the FAQ section in the end.

The drone will takeoff, land or emergency stop/reset by publishing an Empty ROS messages to the following topics: ardrone/takeoff, ardrone/land and ardrone/reset respectively.

In order to fly the drone after takeoff, you can publish a message of type geometry_msgs::Twist to the cmd_vel topic.

    -linear.x: move backward
    +linear.x: move forward
    -linear.y: move right
    +linear.y: move left
    -linear.z: move down
    +linear.z: move up

    -angular.z: turn left
    +angular.z: turn right

The range for each component should be between -1.0 and 1.0. The maximum range can be configured using ROS parameters discussed later in this document. Publishing "0" values for all components will make the drone keep hovering.

The driver publishes two tf transforms between three reference frames: ${tf_prefix}/${base_prefix}_link, ${tf_prefix}/${base_prefix}_frontcam and ${tf_prefix}/${base_prefix}_bottomcam. The ${tf_prefix} is ROS standard way to handle multi-robot tf trees and can be set using tf_prefix parameters, by default it is empty. The ${base_link} is the shared name prefix of all three reference frames and can also be set using parameters, by default it has the value of ardrone_base. Using default parameters, the three frames would be: ardrone_base_link, ardrone_base_frontcam and ardrone_base_bottomcam. By default the root frame is ardrone_base_link. Therefor ardrone_base_frontcam and ardrone_base_bottomcam are children of ardrone_base_link in the published tf tree. This can be changed using root_frame parameter.

The frame_id field in header of all published topics (navdata, imu, cameras) will have the appropriate frame names. All frames are ROS REP 103 compatible.

Calling ardrone/togglecam service with no parameters will change the active video camera stream.

Calling ardrone/setledanimation service will invoke one of 14 pre-defined LED animations for the drone. The parameters are

• uint8 type: The type of animation which is a number in range [0..13]
• float32 freq: The frequency of the animation in Hz
• uint8 duration: The duration of the animation in Seconds.

The type parameter will map [in order] to one of these animations:

    BLINK_GREEN_RED, BLINK_GREEN, BLINK_RED, BLINK_ORANGE,
    SNAKE_GREEN_RED, FIRE, STANDARD, RED, GREEN, RED_SNAKE,BLANK,
    LEFT_GREEN_RIGHT_RED, LEFT_RED_RIGHT_GREEN, BLINK_STANDARD`

You can test these animations in command line using commands like rosservice call /ardrone/setledanimation 1 4 5

If do_imu_caliberation parameter is set to true, calling this service would make the driver recalculate the biases in IMU data based on data from a short sampling period.

The parameters listed below are named according to AR-Drone's SDK 2.0 configuration. Unless you set the parameters using rosparam or in your launch file, the default values will be used. These values are applied during driver's initialization phase. Please refer to AR-Drone SDK 2.0's developer's guide for information about valid values.

• drone_frame_id - The "frame_id" prefix to be used in all tf frame names - default: "ardrone_base"
• bitrate_ctrl_mode - default: DISABLED
• max_bitrate - (AR-Drone 2.0 only) Default: 4000 Kbps
• bitrate - Default: 4000 Kbps
• outdoor - Default: 0
• flight_without_shell - Default: 1
• altitude_max - Default: 3000 mm
• altitude_min - Default: 100 mm
• control_vz_max - Default: 850.0 mm/s
• control_yaw - Default: 100 degrees/?
• euler_angle_max - Default: 12 degrees
• navdata_demo - Default: 1
• detect_type - Default: CAD_TYPE_MULTIPLE_DETECTION_MODE
• enemy_colors - Default: ARDRONE_DETECTION_COLOR_ORANGE_YELLOW
• enemy_without_shell - Default: 0
• detections_select_h - Default: TAG_TYPE_MASK(TAG_TYPE_SHELL_TAG_V2) (The macro is defined in ardrone_api.h)
• detections_select_v_hsync - Default: TAG_TYPE_MASK(TAG_TYPE_BLACK_ROUNDEL) (The macro is defined in ardrone_api.h)
• root_frame - The default root in drone's tf tree (0: _link, 1: _frontcam, 2: _bottomcam) - Default: 0
• cov/imu_la, cov/imu_av & cov/imu_or: List of 9 covariance values to be used in imu's topic linear acceleration, angular velocity and orientation fields respectively - Default: 0.0 for all members (Please check the FAQ section for a sample launch file that shows how to set these values)
• do_imu_calibration: [EXPERIMENTAL] Should the drone cancel the biases in IMU data - Default: 0

The Parrot's license, copyright and disclaimer for ARDroneLib are included with the package and can be found in ParrotLicense.txt and ParrotCopyrightAndDisclaimer.txt files respectively. The other parts of the code are subject to BSD license.

• Rachel Brindle - Enhanced Navdata for AR-Drone 2.0

github offers a nice and convenient issue tracking and social coding platform, it can be used for bug reports and pull/feature request. This is the preferred method. You can also contact the author directly.

The ARDrone 2.0 SDK has been patched to 1) Enable the lib only build 2) Make its command parsing compatible with ROS and 3) To fix its weird main() function issue

The driver has been configured by default to use the maximum bandwidth allowed to ensure the best quality video stream possible (please take a look at default values in parameters section). That is the reason why the picture quality received from Drone 2.0 using this driver is far better than what you usually get using other softwares. If for any reason you prefer the lower quality* video stream, change bitrate_ctrl_mode, max_bitrate and bitrate parameters to the default values provided by the AR-Drone developer guide.

(*) Please note that lower quality does not mean lower resolution. By configuring AR-Drone to use bitrate control with limits, the picture gets blurry when there is a movement.

Drone 1: 320x240@15fps UVLC Codec Drone 2: 640x360@20fps H264 codec with no record stream

tag_type contains information for both source and type of each detected tag. In order to extract information from them you can use the following c macros and enums (taken from ardrone_api.h)

#define DETECTION_EXTRACT_SOURCE(type)  ( ((type)>>16) & 0x0FF )
#define DETECTION_EXTRACT_TAG(type)     ( (type) & 0x0FF )

typedef enum
{
  DETECTION_SOURCE_CAMERA_HORIZONTAL=0,   /*<! Tag was detected on the front camera picture */
  DETECTION_SOURCE_CAMERA_VERTICAL,       /*<! Tag was detected on the vertical camera picture at full speed */
  DETECTION_SOURCE_CAMERA_VERTICAL_HSYNC, /*<! Tag was detected on the vertical camera picture inside the horizontal pipeline */
  DETECTION_SOURCE_CAMERA_NUM,
} DETECTION_SOURCE_CAMERA;

typedef enum
{
  TAG_TYPE_NONE             = 0,
  TAG_TYPE_SHELL_TAG        ,
  TAG_TYPE_ROUNDEL          ,
  TAG_TYPE_ORIENTED_ROUNDEL ,
  TAG_TYPE_STRIPE           ,
  TAG_TYPE_CAP              ,
  TAG_TYPE_SHELL_TAG_V2     ,
  TAG_TYPE_TOWER_SIDE       ,
  TAG_TYPE_BLACK_ROUNDEL    ,
  TAG_TYPE_NUM
} TAG_TYPE;

It is easy to calibrate both cameras using ROS Camera Calibration package.

First, run the camera_calibration node with appropriate arguments: (For the bottom camera, replace front with bottom)

rosrun camera_calibration cameracalibrator.py --size [SIZE] --square [SQUARESIZE] image:=/ardrone/front/image_raw camera:=/ardrone/front

After successful calibration, press the commit button in the UI. The driver will receive the data from the camera calibration node, then will save the information by default in ~/.ros/camera_info/ardrone_front.yaml. From this point on, whenever you run the driver on the same computer this file will be loaded automatically by the driver and its information will be published to appropriate camera_info topic. Sample calibration files for AR-Drone 2.0's cameras are provided in data/camera_info folder.

<node name="ardrone_driver" pkg="ardrone_autonomy" type="ardrone_driver" output="screen">
    <param name="max_bitrate" value="2000" />
    <param name="bitrate" value="2000" />
    <param name="do_imu_caliberation" value="true" />
    <param name="tf_prefix" value="mydrone" />
    <!-- Covariance Values (3x3 matrices reshaped to 1x9)-->
    <rosparam param="cov/imu_la">[0.1, 0.0, 0.0, 0.0, 0.1, 0.0, 0.0, 0.0, 0.1]</rosparam>
    <rosparam param="cov/imu_av">[1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0]</rosparam>
    <rosparam param="cov/imu_or">[1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 100000.0]</rosparam>
</node>

• Make the tf publish optional.
• Add separate topic for drone's debug stream (navdata_demo)
• Add the currently selected camera name to Navdata
• Make the togglecam service accept parameters
• [DONE] Enrich Navdata with magneto meter and baro meter information