Roshan Parmar

CASE STUDY FOR Electronic-Nose Device www.5exceptions.com - Project Overview Electronic-nose devices have received considerable attention in the field of sensor technology during the past twenty years, largely due to the discovery of numerous applications derived from research in diverse fields of applied sciences. We have made a leading platform for our client to sell futuristic electronic nose devices. It converts the raw data fetched from the multiple sensors of electronic nose devices into 2-dimensional graphs, which can be easily understood and analyzed by device owners and researchers to resolve the problems under investigation. www.5exceptions.com - Project Requirement Our client emphasized developing a solution that can simplify to analyze and understand the data fetched from multiple electronic nose devices and utilize them in a meaningful way for future research. The client was looking for a fast, reliable and user-friendly platform that can be easily accessible by people. The client wanted us to develop a solution leveraging emerging technologies to streamline the whole process of using electronic nose devices www.5exceptions.com - Our Solution After understanding the requirement of our client, we started formulating the basic framework of our solution which included advances in sensor design, material improvements, software innovations and progress in microcircuitry design, systems integration, etc. q We had planned to build a web application on AWS using Python as a core programming language and Flask as its web framework www.5exceptions.com - Features of the Solution q We have applied the Principal Component Analysis (PCA) algorithm, which reduces the dimensionality of a data set while keeping as much information as possible by eliminating the less required variables q To execute PCA algorithms in the cloud, we have used Amazon Sagemaker, which is capable of handling millions of data q We also gave the option for the users to view raw data in the form of graphs in the sample graph section. After filling in required data and uploading files of raw data, they can opt to view sample graph or PCA graph, as per the requirement www.5exceptions.com - Basic Framework www.5exceptions.com - Functionality Summary Framework Admin 1. Add application 2. Upload training data 3. Add ml model definition 4. Train ml model with training data 5. Assign ML training model to the application used by mobile app www.5exceptions.com - Functionality Summary Framework Mobile App 1. Connect mobile app to i-nose 2. Select application on the mobile app for detection 3. Get sample data from i-nose for the application 4. Send sample data to server via text file 5. Server runs inference from saved machine learning model for sample data 6. Server sends push notification to mobile app informing detection results www.5exceptions.com - Screenshot of the Solution www.5exceptions.com - Technology Used § § § § § § § § § § § § Python Flask Framework Principal Component Analysis (PCA) Algorithm AWS Sagemaker as cloud machine learning platform AWS EC2 Instance AWS S3 Google Graph HTML5 JavaScript CSS3 jQuery Bootstrap 4 www.5exceptions.com - Future Enhancement In the further advancement of this platform, we are planning to directly fetch raw data from the IoT devices and then store it in the database. After fetching raw data, the platform will apply PCA algorithms and show graphs to the users. www.5exceptions.com - Thank You Contact Us: www.5exceptions.com -