Project description

This projects implements autonomous racing car, in scale 1/10 of Formula 1. The car is equipped with Lidar and encoders for localization and supports various path planning and tracking algorithms.

Hardware

1. Traxxes Slash 2WD (chassis)
2. Pololu magnetic encoder
3. Slamtec S2 Lidar
4. Pololu 5V regulator
5. Raspberry Pi 4 Model B 8GB (on-board computer)
6. Arduino Unu (low level controller)
7. Lipo battery

The Raspberry Pi runs Ubuntu 22.04 Mate, with ROS2

Installtion

Raspberry Pi

download imager to burn OS to PI download ubuntu mate for PI 64 from https://ubuntu-mate.org/download/ install ros:https://docs.ros.org/en/humble/Installation/Ubuntu-Install-Debians.html

skip: set locale
setup sources
install ros2 packages , desktop version
echo "source /opt/ros/humble/setup.bash" >> ~/.bashrc

After installtion update TIME and DATE before sudo apt update and upgrade

Arduino

Burn the low-level code to the Arduino. the code implements PID controller for drive motor, steering commands, and reading of the encoder for odometry.

Install Dependencies:

sudo apt install ros-humble-xacro ros-humble-joint-state-publisher-gui ros-humble-joint-state-publisher ros-humble-rplidar-ros ros-humble-ros2-control ros-humble-ros2-controllers ros-humble-gazebo-ros2-control joystick jstest-gtk evtest ros-humble-slam-toolbox ros-humble-navigation2 ros-humble-nav2-bringup ros-humble-turtlebot3*  ros-humble-twist-mux ros-humble-twist-stamper

sudo apt-get install python3-pip
pip install setuptools==58.2.0
sudo apt install python3-serial 
sudo apt install libserial-dev
sudo apt install net-tools
sudo apt install openssh-server
sudo adduser $USER dialout

Find robot IP: connect robot to wifi on robot computer - ifconfig -a -> wlan -> inet
or
hostname -I
or
how to find robot ip - on robot computer ->wifi -> connection information

setting up ssh:

1. sudo apt install openssh-server
2. service ssh status
3. sudo ufw allow ssh
4. ssh 127.0.0.1
5. ip addr
6. ssh robot1@172.20.10.5

Low level control packages:

pyserial-miniterm -e /dev/ttyACM0 57600 with o - open loop, e- encoders, m-closed loop run in terminal pyserial-miniterm -e /dev/ttyACM0 57600

find path to some serial device: ls /dev/serial/by-path/

Original low level controller(for differential drive): https://github.com/joshnewans/diffdrive_arduino/tree/humble

Serial communication: https://github.com/joshnewans/serial

Ros arduino bridge: https://github.com/joshnewans/ros_arduino_bridge.git

sudo chmod 666 /dev/ttyACM0

low level calibration:

1. drive the car stright, adjust the offset of the steering
2. measure the turning radius and adjust the maximal steering angle
3. steering angle = atan(wheelbase/turning radius)

Usage

Build

source /opt/ros/humble/setup.bash       
colcon build 
source install/setup.bash       

Run

Access the robot via SSH, make sure to source:

ssh robot1@192.168.77.6
ssh robot1@192.168.0.102
cd car_hardware
source /opt/ros/humble/setup.bash
source install/setup.bash

Terminal 1: General communication and low level bring-up

ros2 launch rc_car launch_robot.launch.py

Terminal 2: Lidar communication bring-up

ros2 launch rc_car rplidar_s2.launch.py 

Terminal 3: Rviz, Rviz runs on remote computer and NOT via ssh!

ros2 run rviz2 rviz2 -d src/rc_car/rviz_config/localization.rviz --ros-args -p use_sim_time:=false

Terminal 4: Mapping

ros2 launch rc_car online_async_launch.py use_sim_time:=false

1. panels -> add new panel -> slam tool box plugin
2. in rviz add map and set /map to topic, change durability policy TransientLocal
3. save the map with rviz, serialize the map
4. To store map as numpy array:
   restart
   run with localization and to save the map:

ros2 run rc_car ros_Get_map_client.py --ros-args -p map_name:=lab_demo

After Mapping restart the the robot

Terminal 4: Localization

ros2 launch rc_car localization.launch.py map:=demoe.yaml use_sim_time:=false

1. in rviz, manually write "map" in fixed frame
2. change set initial position
3. add map
4. topic->durability policy->transient local

Terminal 5: Path planning service

ros2 run rc_car PathPlanning_service_prm.py --ros-args -p map_name:=map_world -p use_sime_time:=false

Terminal 6: Path planning client

ros2 run rc_car PathPlanning_client.py 0 0 6.22 -4.5 path1  --ros-args -p use_sime_time:=false

Optional: Smooth the path

Use the script py_path_loops_generator.py

Terminal 7: Path tracking

NOTE - the path_name without the '_meter'

ros2 run rc_car ros_Path_Tracking.py --ros-args -p use_sime_time:=false -p show_path:=true -p show_marker:=true -p path_name:=demoe_path

General

ros2 ros2_control list_hardware_interfaces

How to stop the lidar?
ros2 launch rc_car rplidar.launch.py
ros2 service call /stop_motor std_srvs/srv/Empty "{}"
ros2 service call /start_motor std_srvs/srv/Empty "{}"

How to remap ros2 run teleop_twist_keyboard --ros-args -r /cmd_vel:=/diff_cont/cmd_vel_unstampted

Video

Performance - OptiTrack - Planned VS Executed path

[]