Warehouse Robot (Sokoban Solver)

ITU - Autumn 2022 - Advanced Robotics - Group 3

Team members:

• Balázs Tóth ([email protected])
• Boris Karavasilev ([email protected])
• Michele Imbriani ([email protected])

Introduction

This repository contains our code, unit tests and the report that we created during our mini-project. The goal of this project was to build a robot out of LEGO Mindstroms EV3 and implement a controller with a hierarchical architecture. The controller consists of two subsystems. First, is the low level controller running on the robot that is able to execute a plan by following the edge of a line and detecting intersections. Second, is a high level planner running on a PC that solves the "Sokoban" type of challenge and generates a plan for the low level system.

To run the solver:

> python solve_map.py

To upload the micro python code to the LEGO EV3 robot install the LEGO Mindstorms extension, open the file "main.py" and press Ctr+F5 in VS Code.

Solver output example:

Used symbols

X - wall
# - wall
$ - can/diamond
. - goal
* - can/diamond on a goal
@ - robot/man
a - robot on a goal

Video

👇 Video of our robot in action during a classroom competition. 👇