Skip to content

Latest commit

 

History

History
167 lines (128 loc) · 7.16 KB

README.md

File metadata and controls

167 lines (128 loc) · 7.16 KB
Raw

Supermarket Cleaning Robot

Build Status License

Overview

According to a study done in Morrisville, North Carolina, the Walmart Supercenter located in the town receives about 10,000 people per day. Unquestionably, the actual foot traffic depends on a variety of factors, but we cannot disregard that supermarkets are one of the busiest places in a town. The more the number of people, the more likely it is for the store to become dirty. It always begets frustration among workers to maintain the store at its most pristine level. Our supermarket cleaning robot can remove the stress of cleanliness by performing the tasks of an employee.

Currently, most of the robots are only capable of executing a single task. It turns out to be expensive for a store owner to buy a robot that can perform a single task. We propose to develop a robot that can perform various maintenance tasks. The robot will be able to maintain cleanliness as well as make supermarkets autonomous. The robot will able to clean aisles, stack up empty rows, and collect fallen items.

For prototyping, we are focusing on only one task that is identifying and collecting the items using the robot. The robot will roam in a supermarket like environment in Gazebo and identify the type of items that it needs to collect. It identifies the item using a camera and HSV Color Detection algorithm, mounted on its base, and moves towards the fallen item. Here, we are considering objects such as food, soft drinks cans and it is assumed that the robot will already know the type of item that it needs to pick. As the robot reaches the location of the item and touches it, the item will vanish depicting that the item is collected using a suction cup. The robot will traverse randomly in the supermarket and keep on collecting a can. We are focusing on the detection of cans using the OpenCV to improve the processing of the detection feature. In addition to this, the robot has an obstacle avoidance feature that is used to prevent the robot from colliding from obstacles such as humans, uninteresting items and walls/shelves.

Figure 1 - Robot approaching towards the cans lying on the ground to collect them

Team Members

  • Umang Rastogi - Pursuing masters in Robotics at University of Maryland | B.Tech in Electronics & Communication Engineering
  • Naman Gupta - Grad Student at University of Maryland, pursuing M.Eng. in Robotics.

Demo

In this demo, the turtlebot is our cleaner robot which traverses in the supermarket environment and keeps on scanning for the cans. As the can is detected, a bounding box is created and then the turtlebot apprached towards the can to collect it.


Figure 2 - Demo of Robot scanning and detecting items

Presentation

Click on the link to preview presentation

AIP and Sprint Documents

  • Click on this link to access our AIP Google Sheet.
  • Click on this link to access our Sprint notes document.

Accessing the UML Diagrams

  • Open the UML directory of the project.
  • Access UML diagrams from uml sub-directory.

API Documentations

Known Bugs and Issues

  • Right now we are facing the object collection issue, where our idea is to remove the object from the world as the robot reached the object.
  • The Cola Can should be vanished representing that the object has been collected by the robot.
  • We are working to solve this issue.

Dependencies

  • Ubuntu 16.04/18.04
  • ROS Kinetic/Melodic
  • Gazebo
  • Turtlebot-3 Packages

Install Dependences

  • It is highly recommended that ROS Kinetic/Melodic is properly installed on your system before the use of this project.
  • Follow the instructions on the ROS Kinetic install tutorial page or the ROS Melodic install tutorial page to install Full-Desktop Version of ROS Kinetic.
  • The full-version would help you install Gazebo as well. If you have ROS Kinetic pre-installed on your machine, use the following link to just install Gazebo on your machine.
  • Ensure successful installation by running Gazebo via your terminal window:
gazebo
  • An empty window of Gazebo Simulator should be launched.
  • Create your ROS workspace by following instructions on the create ROS workspace tutortial page.
  • Make sure that Turtlebot-3 packages have been installed on your machine using the following commands:
cd <ROS_WORKSPACE>/
source devel/setup.bash
export TURTLEBOT3_MODEL=burger
roslaunch turtlebot3_gazebo turtlebot3_world.launch
  • A window of Gazebo Simulator with various items with a Turtlebot-3 burger robot should be launched.
  • If an error pops up upon launching the Turtlebot-3 world, then install the necessary Turtlebot-3 packages:
cd <ROS_WORKSPACE>/
git clone https://github.com/ROBOTIS-GIT/turtlebot3_msgs
git clone https://github.com/ROBOTIS-GIT/turtlebot3
cd ../ && catkin_make

Build

Switch to your src sub-directory of your ROS workspace to clone this repository.

<ROS Workspace>/src
  • Run the following commands to clone and build this project:
git clone --recursive https://github.com/urastogi885/supermarket-cleaning-robot
cd ../
catkin_make

Test

Close and terminate everything including rosmaster. In a new terminal, switch to the ROS workspace and build the tests. Type

cd catkin_ws
source devel/setup.bash
catkin_make run_tests_supermarket_cleaning_robot

Run

Now, we use launch file to run. In a new terminal, type

cd catkin_ws
source devel/setup.bash
roslaunch supermarket_cleaning_robot object_collection.launch

Doxygen

To install doxygen run the following command:

sudo apt-get install doxygen

Now from the cloned directory run:

doxygen Doxyfile

Generated doxygen files are in html format and you can find them in ./docs folder. With the following command

firefox docs/html/index.html