lucianzhong / eagleye Goto Github PK

(Alpha version)

Demo Video

Eagleye is an open-source software for vehicle localization utilizing GNSS and IMU[1]. Eagleye provides highly accurate and stable vehicle position and orientation by using GNSS Doppler[2][3][4][5][6]. The flowchart of the algorithm is shown in the figure below. The algorithms in this software are based on the outcome of the research undertaken by Machinery Information Systems Lab (Meguro Lab) in Meijo University.

GNSS receiver

• Ublox NEO-M8T / EVK-M8T
• Ublox ZED-F9P / C099-F9P

GNSS Antenna

• Ublox ANN-MB Series
• Tallysman TW2710

IMU

• Tamagawa Seiki TAG300 Series
• ANALOG DEVICES ADIS16475

Wheel speed sensor

Eagleye uses vehicle speed acquired from CAN bus.

1. Clone and Build MapIV's fork of RTKLIB. You can find more details about RTKLIB here.

    sudo apt-get install gfortran  
    cd $HOME  
    git clone -b rtklib_ros_bridge https://github.com/MapIV/RTKLIB.git
    cd $HOME/RTKLIB/lib/iers/gcc/  
    make   
    cd $HOME/RTKLIB/app  
    make   
   

2. Clone and build rtklib_ros_bridge.

    cd $HOME/catkin_ws/src  
    git clone https://github.com/MapIV/rtklib_ros_bridge.git  
    cd ..  
    catkin_make -DCMAKE_BUILD_TYPE=Release  
   

3. Clone and build nmea_navsat_driver.

    cd $HOME/catkin_ws/src  
    git clone https://github.com/MapIV/nmea_navsat_driver.git  
    cd ..  
    catkin_make -DCMAKE_BUILD_TYPE=Release  
   

4. Installing dependent packages

In the case of Ubuntu18.04 melodic.

	sudo apt-get install ros-melodic-geodesy   
	sudo apt-get install ros-melodic-can-msgs  

In the case of Ubuntu16.04 kinetic.

	sudo apt-get install ros-kinetic-geodesy   
	sudo apt-get install ros-kinetic-can-msgs  

5. Clone and build eagleye.

    cd $HOME/catkin_ws/src  
    git clone https://github.com/MapIV/eagleye.git  
    cd ..  
    catkin_make -DCMAKE_BUILD_TYPE=Release  
   

6. RTKLIB settings.

Change inpstr1-path of $HOME/RTKLIB/app/rtkrcv/conf/rtklib_ros_bridge_sample.conf according to the serial device you use.

ie)

inpstr1-path =/serial/by-id/usb-u-blox_AG_-_www.u-blox.com_u-blox_GNSS_receiver-if00:230400:8:n:1:off

7. nmea_navsat_driver settings.

Change arg name="port" of $HOME/catkin_ws/src/nmea_navsat_driver/launch/f9p_nmea_serial_driver.launch according to the serial device you use.

ie)

<arg name="port" default="/dev/serial/by-id/usb-FTDI_FT232R_USB_UART_AG0JNPDS-if00-port0" />

8. GNSS receiver settings. Configure the receiver settings using u-center.

• UART1(Connect to RTKLIB) Enable UBX message (output rate 5Hz, baudrate 230400) ※ Set to output only RAWX and SFRBX
• UART2(Connect to nmea_navsat_driver) Enable NMEA message (output rate 1Hz, baudrate 115200) ※ Set to output only GGA and RMC

This file is a sample configuration file for F9P.
Open u-center.
Tools/Receiver Configuration.../Load configuration "Transfer file -> GNSS"

9. IMU settings.

• Output rate 50Hz

10. Check the rotation direction of z axis of IMU being used. If you look from the top of the vehicle, if the left turn is positive, set "reverse_imu" to true in eagleye/launch/eagleye_localization.launch.

    param name="/eagleye/reverse_imu" type="bool" value="true"
    

1. Check if wheel speed (vehicle speed) is published in /can_twist topic.

• Topic name: /can_twist
• Message type: geometry_msgs/TwistStamped twist.liner.x

2. Check if the IMU data is published in /imu_raw topic.

3. Start RTKLIB.

    cd $HOME/RTKLIB
    bash rtklib_ros_bridge_single.sh
   

4. Check if RTKLIB is working by execute the following command in the terminal. If the RTKLIB is working correctly, positioning information is appeared continuously in the terminal.

    status 0.1  
   

5. Start rtklib_ros_bridge.

    roslaunch rtklib_bridge rtklib_bridge.launch   
   

6. Start nmea_navsat_driver.

    rosrun nmea_navsat_driver f9p_nmea_serial_driver.launch   
   

7. Start eagleye.

    roslaunch eagleye_rt eagleye_rt.launch
   

The 3D maps (point cloud and vector data) of the route is also available from Autoware sample data.

1. Play the sample data.

    rosparam set use_sim_time true
    rosbag play --clock eagleye_sample.bag
   

2. Launch eagleye.

    roslaunch eagleye_rt eagleye_rt.launch
   

The estimated results will be output about 100 seconds after playing the rosbag. This is because we need to wait for the data to accumulate for estimation.

1. J Meguro, T Arakawa, S Mizutani, A Takanose, "Low-cost Lane-level Positioning in Urban Area Using Optimized Long Time Series GNSS and IMU Data", International Conference on Intelligent Transportation Systems(ITSC), 2018 Link

2. Takeyama Kojiro, Kojima Yoshiko, Meguro Jun-ichi, Iwase Tatsuya, Teramoto Eiji, "Trajectory Estimation Based on Tightly Coupled Integration of GPS Doppler and INS" -Improvement of Trajectory Estimation in Urban Area-, Transactions of Society of Automotive Engineers of Japan 44(1) 199-204, 2013 Link

3. Junichi Meguro, Yoshiko Kojima, Noriyoshi Suzuki, Teramoto Eiji, "Positioning Technique Based on Vehicle Trajectory Using GPS Raw Data and Low-cost IMU", International Journal of Automotive Engineering 3(2) 75-80, 2012 Link

4. K Takeyama, Y Kojima, E Teramoto, "Trajectory estimation improvement based on time-series constraint of GPS Doppler and INS in urban areas", IEEE/ION Position, Location and Navigation Symposium(PLANS), 2012 Link

5. Junichi Meguro, Yoshiko Kojima, Noriyoshi Suzuki, Eiji Teramoto, "Automotive Positioning Based on Bundle Adjustment of GPS Raw Data and Vehicle Trajectory", International Technical Meeting of the Satellite Division of the Institute of Navigation (ION), 2011 Link

6. Yoshiko Kojima, et., al., "Precise Localization using Tightly Coupled Integration based on Trajectory estimated from GPS Doppler", International Symposium on Advanced Vehicle Control(AVEC), 2010 Link

Eagleye is provided under the BSD 3-Clause License.

If you have further question, email to [email protected].