Overview / Usage

A cargo management system that uses RFID technology to track packages and find if this package was stolen or misdirected, some prediction AI algorithms run over it allowing risk prediction to increase and time average to be read.
This project was academically based and concept proof, winning the First prize in Brazil SBESC Embedded Systems competition in 2017 and was a First prize in Intel Embedded Systems Design Competition 2018 in Shanghai China.

Methodology / Approach

We use a RFID reader to detect Tags, also a 2 tags system to allow the detection to a package violation and to a package theft even if the package isn't there anymore.
A LSTM archtecture to time prediction.
A Statistical analysis to risk prediction.
A Dieff Hellman encryption technique to grant security in the IoT channel.
We made PHP and AJAX based webserver and webclients to make the system show the data and update everything in the system.
Them we built a Flask Python application to serve as an AI API to be consulted by the client.

Technologies Used

RFID, Python, Keras, Flask, PHP, Javascript, UPsquared, FPGA Cyclone V

Repository

https://github.com/MiniMarvin/markme_ai