IoT device built with Arduino and Raspberry Pi. Cardboard box models real-world cargo shipping container.
Cargo ships are the primary ways of transporting produce across oceans. However, their sensitive nature makes them sensitive to degrading in transit -- resulting in unnecessary waste. About 33% of global fresh produce is thrown away due to their quality degrading during shipment. Additionally, every year, at the US-Mexican border, 35-45 millions pounds of fruits and vegetables are thrown away due to not meeting standards. This hurts both consumers and suppliers alike.
What it does
Uses sensors, Computer Vision, and ML to improve the efficiency of current supply chain management. Using IOT, we build smart containers that can detect if a produce is fresh or not and then creates a bidding system based on how fresh a produce is and uses it to distribute it.
[Input]: Suppliers create a product page on their shipment and sync the device to it. [Bidding]: Prospective buyers can bid on the product shipment by inputting two parameters: their bid amount and their maximum freshness threshold. After the bid winning, it is locked to them. [Monitoring/Rebidding]: The order is shipped and monitored by the hardware to provide interested parties with details such as location, humidity, temperature, CO2, and the like. If it falls below a set freshness threshold, the customer can back-out and re-open bidding. Otherwise, it works like a typical B2B ordering site and remains locked to the customer.
The freshness score is calculated by using an Ensemble Machine Learning approach that incorporates multivariate Ordinary Least Squares and Computer Vision to predict how fresh a produce is. The image is then updated onto the database after ever hour.