Lego car-like robot stuff

The project idea is a mobile car-type robot capable of mimicking human-like trajectories to reach its assigned position while avoiding obstacles.

This repository includes the source code for both its control system and computer vision system (CV). The CV was designed for robot localization and obstacle detection.

Please take a loot at the video first.

Table of Contents

• Description
• Installation
• Usage
• Support
• Contributing

Description

The repository consists of three parts:

• The control_part folder includes the source code of the control system
• The cv_part folder includes the source code of the computer vision systel
• The scilab folder includes files associated with the simulation of the control system in Scilab

The Car-type robot was built using the LEGO MIDSTRMS EV3 set. The general view of the robot and its kinematic model are shown in the figures below.

Car-like robot	Kinematic model

The whole structure of the project is illustrated on the figure below.

Installation

Robot

1. Build the car-type robot (Ackermann Steering tutorial) using LEGO mindstorms EV3

2. Creating a bootable Micro SD card with Linux

3. Connect the EV3 to a WiFi network using a WiFi adapter (eg. a doodle)

4. Connect via SSH to your robot (user: robot, password: maker)

   ssh [email protected] (IP adress of the robot)

5. Clone the repository or copy only the control_part folder to your robot

   cd ~
   git clone https://github.com/red-itmo/lego-car-project.git     

6. Change permissions for Python scripts

   cd control_part/sources/carsystems/
   sudo chmod +x *.py

7. Intall numpy library for Python3

   pip3 install numpy

Computer

1. Install OpenCV library to your computer (installation on Linux)

2. Install kivy library(installation on Linux)

3. Install numpy

   pip3 install numpy

Usage

1. Mount your camera on the ceiling so that the optical axis is directed as a normal to the floor. Measure the distance from the camera to the floor. Calculate the length of the diagonal of the frame along the floor using the diagonal field view of camera. This value has to be inserted instead of the "D" variable in the file cv_part/kivy_gui.py.

   px_to_m = D / math.hypot(cam_res[0], cam_res[1])

2. Creat a color marker for the robot and place it on its upper part

3. Calibrate the colors using this script. Detection of the robot's position and orientation must be stable. The color ranges for the markers need to be written in the script cv_part/libs/Mapping.py. More precisely, inside the constructor of the class "Mapping" in the following lines:

   self.blueLower = (96, 110, 126)
   self.blueUpper = (105, 255, 247)
   self.yellowLower = (0, 89, 122)
   self.yellowUpper = (94, 150, 255)

4. Clone this repository to any directory

   git clone https://github.com/red-itmo/lego-car-project.git

5. Launch the CV part of the project

   $ python cv_part/kivy_gui.py

You will see a window showing a user friendly interface as seen in the figure below.

Support

Please open an issue if you've found a bug.

Contribution

Please contribute using Github Flow. Create a branch, add commits, and open a pull request.

Authors

Evgeniy Antonov

Aleksandr Kapitonov

Aleksandr Karavaev

Egor Zamotaev

Kirill Artemov

Oleg Souzdalev

Rami Al-Naim

Dmitrii Dobriborsci