F1TENTH

Setup | Startup Instructions | Resources

Hello, this repo functions to hold the f1tenth workspace on ROS2 Foxy as well as any evaluations made on the algorithms used.

f1tenth
├── dep //DEPRECATED
│   └── ...
├── f1tenth_ws //ROS2 PACKAGES
│   └── ...
├── f1tenth_gym_ros //ROS2 simulator
│   └── ...
└── performance //EVALUATION
    └── ...

Setup

1. Clone the repository.

git clone --recurse-submodules https://github.com/bentjh01/f1tenth.git

2. Build the containers

docker build -f f1tenth_bt.Dockerfile -t bentjh01:f1tenth-foxy .

Startup Instructions

Mac

1. Start containers in detached mode, and exec into f1tenth_gym_ros-sim-1.

./run_f1tenth_sim_mac.sh

2. Start the simulator from the interative termial of f1tenth_gym_ros-sim-1.

./start.sh

3. Open the novnc from the browser for display
   localhost:8080/vnc.html

4. Exec into f1tenth_bt

./run_f1tenth_mac.sh

Timeline/ Plan

1. Code a simple algorithm to complete rounds around a track to obtain starting data

   • Requirements
     • algorithm outputs in Twist
     • node to convert
     • able to make consecutive laps minimum 5

2. Create library to analyse alogorithm performance

   • Metrics:

     • fastest lap,
     • consecutive laps.
     • consistent path

   • Rules:

     • starts from 5m away from a lap counter line.
     • said line is perpendicular to the wall of the track
     • said line is located at a stright.

   • Visualisation:

     • starting and ending points of the recording,

3. Code more algorithms with the same requirements as before and collect data on them in a rosbag

   • List of Algorithms
     • Reactive Navigation
       • Minimum Filter
       • Search Algorithm e.g. A* or Dijkstra
     • Global Planning
       • SLAM
         • gmapping
         • etc
       • Optimisation
         • geometric
         • time
         • etc
       • Controller
         • PID
         • MPC
     • ROS2 Nav

4. use best performing algorithm to collect data set on different maps

5. train neural network on dataset

6. train reinforment learner on environment maybe Isaac Sim

Setup

1. This repo

1. Clone this repo

git clone https://github.com/bentjh01/BT_f1tenth.git

2. Setup a python virtual environment

# Install python venv
pip install virtualenv
# Creates new virtual environment
python -m venv .venv
# Activate virtual environment
source .venv/bin/activate
# Install requirements
pip install -r venv_requirements.txt

2. Setup F1TENTH Simulator

https://github.com/bentjh01/f1tenth_gym_ros.git

Creating a new package

ros2 pkg create --build-type ament_cmake <package name>
cd <package name>
mkdir <package name>
touch ./<package name>/__init__.py
mkdir scripts

Build Docker image

docker build -t [maker:name] .

Resources

• PlotJuggler
• Python Virtual Environments
• MarPlotLib Img
• f1tenth/f1tenth_gym_ros.git