Open-source software developed/modified by the Meidai Autonomous Driving (MAD) group (Nagoya University, Japan) for our participation at Tsukuba Challenge 2017.
Tsukuba Challenge 2017 couse and scenes:
Our NUIV robot:
The software includes:
- Autoware: fork of CPFL/Autoware modified for our robot
- E2E (end-to-end): software and models
- RobotControl: SSM and SHSpur libraries
- Maps: point-cloud maps and vector maps (indoor and outdoor)
This software was developed and tested on the following environment:
- Mobile robot with SH2 board running SH-Spur software
- CPU Intel Xeon E3-1545M 8 cores
- 32GB RAM
- NVIDIA M1000M with 4GB GPU RAM
- ROS kinetic (Ubuntu 16.04)
- TensorFlow
- Get the software
$ cd $HOME
$ git clone https://github.com/MAVRG/TC2017.git
- Autoware
$ cd ~/TC2017/Autoware/ros/src
$ catkin_init_workspace
$ cd ../
$ ./catkin_make_release
- RobotControl
$ cd ~/TC2017/RobotControl/ssm-shspur-ros/SSM-0.32
$ make
$ sudo make install
$ cd ../sh_spur-0.11/
$ make
$ sudo make install
$ cd ~/TC2017/RobotControl/ssm-shspur-ros/catkin_ws/src
$ catkin_init_workspace
$ cd ..
$ catkin_make
- E2E
$ cd ~/TC2017/E2E/tsukuba_ws/src
$ catkin_init_workspace
$ cd ..
$ source ~/TC2017/Autoware/ros/devel/setup.bash
$ catkin_make
- Autoware
$ cd ~/TC2017/Autoware/ros
$ ./run
- RobotControl: joystick operation
$ cd ~/TC2017/RobotControl/ssm-shspur-ros
$ ./AutoKerberos.sh [optional argument: path to record rosbag]
- RobotControl: autonomous operation from Autoware
$ cd ~/TC2017/RobotControl/ssm-shspur-ros
$ ./Automove-log.sh [optional argument: path to record rosbag]
- E2E
$ cd ~/TC2017/E2E/Cyclops
$ ./Cyclops-log.sh [optional argument: path to record rosbag]