XavierAutonomouDriving

ROS-based program to control a 1/3-scale motorized vehicle.

Hardware

[on-board computer] NVIDIA Xavier, Ubuntu 18, JetPack 4.2
[2D LiDAR] RPLIDAR A1 https://www.slamtec.com/en/Lidar/A1
[motor driver] 5V-36V 350W DC Brushless Motor Controller BLDC PWM Driver Board
[motor] 8inch 24V Brushless Hub Motor Toothless Wheel For Electric Scooter Skateboard
[camera1] Logitech C270
[camera2] ELP USB with Camera 2.1mm Lens 1080p Hd Free Driver USB Camera Module,2.0 Megapixel(1080p) Usb Camera,for Linux Windows Android Mac Os
[PWM board] MCP4728 Evaluation Board
[IMU] PCA9685
[Joystick] SONY DUALSHOCK 3 + Mayflash Magic-NS (bluetooth receiver + converter)

Features

Joystick -> Magic-NS -> PWM board -> motor driver boards -> motors
2 front wheels + 1 auxiliary rear wheel. Only front wheels will be motorized
SLAM using Google Cartographer

Getting started

These instructions will get you a copy of the project up and running on your local Ubuntu 18 machine for development and testing purposes.

Prerequisites

1 Installing ROS melodic on your Ubuntu 18 machine

http://wiki.ros.org/melodic/Installation/Ubuntu

2 Install RPLIDAR_ROS

sudo apt-get install ros-melodic-rplidar-ros

or

https://github.com/robopeak/rplidar_ros

3 Install usb-cam ROS package

sudo apt-get install ros-melodic-usb-cam

usb-cam package may require some workarounds to function properly in ROS melodic.

https://github.com/ros-perception/image_pipeline/issues/201

4 Install SMBUS2 for Python2 ROS currently works with Python 2 only. To let non-root users to access i2c port of Xavier, you may need to add your user to i2c group. If necessary, copy troubleshoot/30-i2c-tools to /etc/udev/rules.d and reboot.

5 Install IMU packages A tool for debugging IMU RTIMULib2 A ROS interface for the debugging tool mentioned above https://github.com/jeskesen/i2c_imu

6 Follow the instruction below to install Google Cartographer. Because cartographer uses ninja as a compilation tool, it is recommended to create an independent ROS workspace for Google Cartographer and source the directory in .bashrc.

https://google-cartographer-ros.readthedocs.io/en/latest/

Wiring/connection

1 Xavier GPIO <-> MCP4728 GPIO 2 <-> VDD GPIO 6 <-> GND GPIO 3 <-> SDTA GPIO 5 <-> SCLK

2 MCP4728 <-> motor drivers OutputA <-> left wheel speed OutputB <-> left wheel direction OutputC <-> right wheel speed OutputD <-> right wheel direction

3 Magic-NS Plug Magic-NS in the USB hub. Run in Direct Input Mode (green LED)

4 Cameras Plug two cameras in the USB hub.

5 Lidar Plug in typeC port.

Running the test

roslaunch main main.launch

IMU in Google Cartographer

For 2D LIDAR and SLAM, IMU is not a must. However, it can improve the quality of the map greatly especially at the initial stage of mapping. To enable IMU, cartographer requires an IMU topic (sensor_msgs/IMU). ALso enable imu in "trajectory_builder.lua" in cartographer workspace.

atelier-ritz/XavierAutonomousDriving