Parkotron

Autonomous Space Detection with Robot Operating System

Project Overview

Parkotron is an autonomous robot that patrols a lane looking for available spaces, detects valid spaces, and parallel parks itself into a valid space.

The algorithm was written using Python and ROS and implemented on a Turtlebot3 Burger model using the onboard Raspberry Pi camera and LiDAR sensor.

My Contributions

I was responsible for implementing the computer vision for valid space recognition via AR tags, available space detection via LiDAR sensing, and proportional controls for robot turning.

In addition, I established the state machine architecture for the robot's behavior and created the final demonstration video for the project.

Parkotron finding a parking spot

Video demo

State machine diagram

Parkotron's simulated parking environment

Detecting a valid spot: Parkotron point-of-view

FULL Project Website ↗