Design Fabrication and Control of a Product Listing Machine

Project Title

Design Fabrication and Control of a Product Listing Machine

Project Area of Specialization Artificial IntelligenceProject Summary

Product listing is the process in which a product’s images are captured following specific standards and its necessary information (name, category, size, variations and description) is uploaded on an eCommerce website along with images. Product listing provides information of each product to be sold on eCommerce website. When the listing is done, the product is visible/available to customers. If the listing process is incomplete, to product wouldn’t be visible to its customers. 

Usually, a virtual assistant is hired for the listing of products. He fills out more than 40 sections while listing a product on an eCommerce website. It contains information such as the product's size, type, multiple images, and other pertinent information. While eCommerce is expanding rapidly with each passing day. The entire planet is becoming a global marketplace. Every day, new inventions take shape. Almost 4000 products are listed on just Amazon each day. To stay up, it's vital to make the listing process quick, simple, and cost-effective.

The conventional process of listing products is not enough to deal with exponential growth in ecommerce. The problem of conventional listing method can be twofold:

• It is a long-drawn-out process
• It is a high-priced task

A reliable solution is to design a machine that will serve the purpose. When a product is placed on the slider plate of this machine, it will take it to a box where its images, size, and other attributes will be determined using Machine Learning Techniques and the machine will list it to a web page local host. This machine will get the key attributes of a product and list it on eCommerce websites. It will save both time and effort.

An application of artificial intelligence (AI) and computer that allows systems to learn and develop without being explicitly programmed. It is commonly known as machine learning.

Machine learning is concerned with the creation of computer programs that can access data and learn for themselves.

All the necessary information about the product gets listed on a webpage that is visible to our customers and it helps the customer to make is decision. This machine would be time saving and less expensive. I will assist to make the overall listing process easy and quick. We do not claim that this project is a complete substitute for human-role as virtual assistant as artificial intelligence still is not still enough to replace human-being after years of research. This project is only meant to facilitate the human efforts and contribution. An appropriate solution is to design a machine that will serve the purpose. When a product is placed on the slider plate of this machine, it will take it to a black box where its images, size, and other attributes will be determined using Machine Learning Techniques and the machine will list it to a web page local host. This machine will get the key attributes of a product and list it on eCommerce websites. It will save both time and effort.

Project Objectives

The main idea of this project entitled Design, Fabrication and Control of Product Listing Machine is to design a machine using the knowledge of design of machine elements and machine learning techniques for facilitating human efforts in conventional method listing products on ecommerce websites. This will in return provide an alternative method of product listing in ecommerce business. Ecommerce means buying and selling products online. It is one of the fastest growing industries in the world. Its global retails sales in 2021 were almost 4.9 trillion dollars.  Thousands of new products are listed on different ecommerce websites every day. This process of listing products on websites is done manually by virtual assistants. This makes the process of listing expensive. A viable solution would be to design an intelligent machine that will help make the listing process easier. It is a trailblazing effort to design and develop an automated product listing machine that will assist the listing process. This machine intends to extract the key attribute of a product and list it on a website/server page. The design and assembly of different parts of this machine is done on SolidWorks. Different key attributes of product are to be determined using Machine Learning Algorithms and listed on local web page. The machine is not in any way going to replace the human role but will facilitate him. Further efforts will lead to improve the performance and features of machines.

Project Implementation Method

To save time and energy, the whole project is divided into three sections. Each group member was assigned to work on one section. These sections are:

• Design the 3D model of the listing machine
• Complete the software part of the machine
• Control the movement of the slider plate

Design the 3D model of the listing machine

We will discuss the methodology for each section one by one: The 3D design of the Automated product listing machine has been completed on SolidWorks. We went through several iterations to improve our design. Final design of machine is shown here.

Complete the software part of the machine

The materials used by this model to achieve the active monitoring model are as follows:

 YOLO: Object Detection Algorithm

YOLOv5

Data Creation

Model Training

Running Our Trained Model

NVIDIA GPU (Graphics Processing Unit)

CUDA (Compute Unified Device Architecture)

CuDNN

Darknet

Localhost

Django

Django Directory Structure

Control Movement of Slider Plate

We want to control the motion of slider plate such that when it completely enters the box of machine it should stop there for a while so that we could capture the images of objects using camera. For this purpose, we need a motor whose speed can easily be controlled. A stepper motor can serve this purpose.

1. NEMA34 Stepper Motor

For our project, we need a motor of very high torque. We will use a NEMA 34 Stepper motor for our project. It is a bipolar brushless DC motor. Its face plate size is 3.4X3.4 inches. It draws a current of 6A at 9V.  It has a very high holding torque of 42 kg-cm

We will be using “2DM860H” micro-step driver and Arduino UNO to control its speed. TB600 micro-step driver can be used but it will provide less current and low speed. If motor draws current more than 3A, it will damage the driver, for this purpose we are using 2DM860H driver. It can provide an output current of 12A. it has a high range of operating voltage. Moreover, it has 400 steps per revolution. It is safe and perfect for the job.

Benefits of the Project

To complete the listing process a virtual assistant has to fill almost 40 sections on ecommerce websites. In conventional method, the data or information about product is fed manually and it requires a lot of time to list one product. Also, in order to hire VAs for product list, people mostly prefer freelance marketplaces. These virtual assistants charge hourly basis for the task. An experienced virtual assistant charge $15 per hour for listing purpose. Moreover, you also need to hire a photographer for taking images. This makes the listing process expensive.

A reliable solution is to design a machine that will serve the purpose. When a product is placed on the slider plate of this machine, it will take it to a box where its images, size, and other attributes will be determined using Machine Learning Techniques and the machine will list it to a web page local host. This machine will get the key attributes of a product and list it on eCommerce websites. It will save both time and effort.

An application of artificial intelligence (AI) and computer that allows systems to learn and develop without being explicitly programmed. It is commonly known as machine learning.

Machine learning is concerned with the creation of computer programs that can access data and learn for themselves.

All the necessary information about the product gets listed on a webpage that is visible to our customers and it helps the customer to make is decision. This machine would be time saving and less expensive. I will assist to make the overall listing process easy and quick. We do not claim that this project is a complete substitute for human-role as virtual assistant as artificial intelligence still is not still enough to replace human-being after years of research. This project is only meant to facilitate the human efforts and contribution. An appropriate solution is to design a machine that will serve the purpose. When a product is placed on the slider plate of this machine, it will take it to a black box where its images, size, and other attributes will be determined using Machine Learning Techniques and the machine will list it to a web page local host. This machine will get the key attributes of a product and list it on eCommerce websites. It will save both time and effort.

Technical Details of Final Deliverable

We will be able to move a product form one end of the machine to the other end. We will be using a slider plate mounted on linear bearings and a lead screw to move it to and fro. When the product is placed under the cameras, we will be able to capture its images and recognize the product using camera and machine learning.

We are creating a webpage to list all the key attributes of product along with images. There is an important point to mention that our machine would not distinguish between all types of products of same category. For example: we have trained our system to list a scientific calculator. Our machine will capture images of scientific calculator and list it on our website. but it would be able to recognize other types of calculators because our system has not been trained to list them. Our aim is to achieve these goals,

• Completion the 3d design of product listing machine
• Manufacturing / Fabrication of machine
• Installation of cameras to capture images
• A webpage to list and display the product

Final Deliverable of the Project HW/SW integrated systemCore Industry OthersOther Industries Manufacturing Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	67170
ball screw	Equipment	1	17850	17850
linear slider rails with bearings	Equipment	1	12400	12400
stepper motor	Equipment	1	16500	16500
side pannels	Equipment	4	1730	6920
camera	Equipment	1	3500	3500
manufacturing cost	Miscellaneous	1	10000	10000