monoDrive Clients

Collection of monoDrive client software for different languages.

• C++
• ROS

monoDrive C++ Client

Windows Prerequisites

• Windows 10
• Visual Studio 2019 Community Edition
• VSCode

Windows Library Dependencies

NOTE: Extract or install these libraries to C:\local so cmake can find them.

• Boost

  • Add C:\local\boost_1_73_0\lib64-msvc-14.2 to your PATH variable

  • Create the BOOST_ROOT environment variable and set it to C:\local\boost_1_73_0\

    

• CMake

  • When installing, make sure to choose the option to CMake to your Windows Path variable.

To build the examples the following are required:

• OpenCV Extract to C:\local\opencv and add C:\local\opencv\build\x64\vc15\bin to your PATH environment variable.
• Eigen Extract to C:\local\Eigen3 and add C:\local\Eigen3 to your PATH environment variable.

  • NOTE: You will need to move the extracted contents out of the version subfolder (eg .\eigen-3.3.7\) into the parent directory, such that C:\local\Eigen3\Eigen is a valid directory path.

    

Figure: Environment Variables

Ubuntu 18.04 Prerequisites

• Ubuntu 18.04
• VSCode
• Install packages:

  sudo apt-get update && sudo apt-get install libboost-dev libboost-system-dev build-essential libeigen3-dev

Setup Using VSCode

Build

1. Open VSCode.

2. Add the following VSCode extensions:

   • CMake
   • CMake Tools
   • C/C++

3. Select File -> Open Folder and navigate to this folder to build the cpp-examples or simulator-cpp-client to build just the client library.

4. Use the CMake extension to configure and build

   1. Click the Configure All Projects icon:

      

   2. If prompted to Scan for Kits select Yes.

      Windows: Use Visual Studio Community 2019 Release - amd64.

      Linux: Use the compiler of your choice. Tested with g++ 7.5.0.

   3. Build the client by clicking the Build All Projects icon:

      

Set Launch Target

NOTE: Launching automatically fires off a build check in cmake so just setting the launch target will suffice.

Run Example

Windows

Example configuration with fisheye_camera_equidistant as a build + launch target.

After you've set your build configuration and target, you can run by hitting F5 to run in debug or Ctrl+F5 to run without debugging.

Installation with CMake

You can include the monoDrive Simulator client in your existing CMake project by adding the following lines to your CMakeLists.txt

# add monodrive client library from local repo
add_subdirectory(path/to/monodrive-client/monodrive mdclient)

# link targets as needed
target_link_libraries(<mytarget> monodrive)

Installation with Bazel

You can include the monoDrive Simulator client in your existing Bazel project by adding the following lines to your WORKSPACE file

local_repository(
    name = "monodrive",
    path = "path/to/monodrive-client"
)

and the following to your BUILD files as needed

cc_library(
    name = "my_lib",
    srcs = [...],
    hdrs = [...],
    deps = [
        ...,
        "@monodrive//monodrive/core:monodrive"
    ]
)

Installation to system on Ubuntu

You can build and install the monoDrive Simulator client to your Ubuntu system using CMake

mkdir build
cd build
cmake ..
make
sudo make install

This will install the client library under the prefix /usr/local/monodrive/client. It can now be included or linked as needed.

For example, to compile your own executable with the monoDrive client library

g++ main.cpp -I/usr/local/monodrive/client/include/ -L/usr/local/monodrive/client/lib -lboost_system -lmonodrive -o my_program

Make library available for dynamic loading

export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/usr/local/monodrive/client/lib

Finally run

./my_program

monoDrive ROS Client

There are two ways to integrate the monoDrive simulator with ROS.

The first is to configure the sensors to publish data directly to ROS instead of usng TCP streams. This is supported for IMU, Waypoint, State, Camera and Lidar sensors. The simulator_control, vehicle_control, andwheel_vehicle_control examples demonstrate how to do this.

The second is using TCP/IP streams to receive sensor data, converting that data on a frame by frame basis to ROS messages, and publishing those on the desired ROS topic. To support this approach, there is a message conversion utility class in monodrive/ros/src/monodrive_msgs/include/MessageFactory.h that facilitates converting raw sensor frame data to ROS messages and vice-versa. The distributed ROS example demonstrates how to use this approach.

Ubuntu 18.04 Prerequisites

• monoDrive c++ client
• ROS Note: Tested with melodic
• ROS Bridge:

sudo apt-get install ros-melodic-rosbridge-suite

• ROS Joy:

sudo apt-get install ros-melodic-joy

monoDrive ROS Packages and example build

1. Execute the following to build the monoDrive ROS library with:

cd ./monodrive/ros
catkin_make
source devel/setup.bash

2. Add the monoDrive ROS library setup file to your .bashrc to add the packages to your ros path on terminal load:

echo "source <path/to/monodrive/ros/devel/setup.bash>" >> ~/.bashrc

3. Execute the following to build the ROS packages:

cd ./examples/ros
catkin_make
source devel/setup.bash

4. Add the examples setup file to your .bashrc to add the packages to your ros path on terminal load:

echo "source <path/to/examples/ros/devel/setup.bash>" >> ~/.bashrc

monoDrive Simulator and Client network setup for the the direct-to-ROS examples

If you are running both the client and simulator on the same machine you can skip this section as the networking defaults are for local host.

If you are running the simulator and client on separate machines the following networking settings must be configured.

1. Set the IP address and port IDs for the machine running the simulator and the machine running the ros bridge

   • In the configuration file, simulator_control/confg/simulator.json, set the IP and port (default is 9090) of the machine that will host the ros bridge:

   "ros": {
       "port": 9090,
       "server": "192.168.86.167"
   },

   • and the IP and port of the simulator:

   "server_ip": "192.168.86.168",
   "server_port": 8999,

2. Forward the ports on both machines (9090 and 8999) from step 1 or disable the firewalls on both machines.

Launching the direct-to-ROS examples

3. To launch the monoDrive examples create 3 tabs and run each command in a separate terminal:

   1. Launch rosbridge, you can leave this running:

   roslaunch rosbridge_server rosbridge_tcp.launch bson_only_mode:=True

   2. Start the vehicle control node which will subscribe to the state sensor topic and publish vehicle controls (the simulator does not need to be running)

   rosrun vehicle_control node

   Note: The vehicle_control example only requires the monodrive_msgs package and provides an example of how to connect your code to monoDrive through ROS messages.

   3. Make sure the monoDrive simulator is running since the next command will connect to and start the simulator scenario running.

   rosrun simulator_control node

Launching the message conversion to ROS example

4. To launch the monoDrive example create 3 tabs and run each command in a separate terminal:
   1. Make sure the monoDrive simulator is running since the run command will connect to and start the simulator scenario running.

   rosrun distributed node

Using the G920 Logitech wheel

Configuring the Joystick

1. Connect your joystick to your computer. Check if Linux recognized your joystick.

   ls /dev/input/

From the list verify that you can see jx0 or jsX in the list, this will mean Linux recognize your joystick.

2. Make the joystick accessible to the ROS joy node. Where jsX is js0 from the example above.

sudo chmod a+rw /dev/input/jsX

3. Setup your the joystick device to use on the joy node . Assumig your joystick is js0

rosparam set joy_node/dev "/dev/input/js0"

Optional: You can also specify the deadzone as a parameter

rosparam set joy_node/deadzone 0.05

Launching the example

To launch the monoDrive examples create 4 tabs and run each command in a separate terminal:

1. Launch rosbridge, you can leave this running:

roslaunch rosbridge_server rosbridge_tcp.launch bson_only_mode:=True

2. Start the joy_node.

rosrun joy joy_node

3. Start the wheel vehicle control node which will subscribe to the state sensor topic and publish vehicle controls (the simulator does not need to be running)

rosrun wheel_vehicle_control node

Note: The vehicle_control example only requires the monodrive_msgs package and provides an example of how to connect your code to monoDrive through ROS messages.

4. Make sure the monoDrive simulator is running since the next command will connect to and start the simulator scenario running.

rosrun simulator_control node

Note: The following table show how the buttons in the G920 wheel map to the ROS message.

AXES	BUTTONS
Function ROS message Steering joy->axes[0] Throttle joy->axes[1] Brake joy->axes[2] Clutch joy->axes[3] Directional pad RIGHT joy->axes[4] Directional pad LEFT joy->axes[4] Directional pad UP joy->axes[5] Directional pad DOWN joy->axes[5]	Button ROS message A joy->buttons[0] B joy->buttons[1] X joy->buttons[2] Y joy->buttons[3] RB joy->buttons[4] LB joy->buttons[5] Menu button joy->buttons[6] View button joy->buttons[7] RSB joy->buttons[8] LSB joy->buttons[9] Xbox button joy->buttons[10]