Team 54

In current factory and warehouse settings ensuring workers consistently use Personal Protective Equipment (PPE) is a daily challenge. Machine operators and engineers occasionally need to walk into the hazardous manufacturing section of the plant. There are several potential hazards in this part of the factory. The loud machines can lead to hearing damage overtime. Overhanging machines may drop parts from time to time, so hard hats are required. Some machines may launch metal chips, so protective eye wear is a must. Current methods of PPE enforcement include manual checks or signage which can lead to inconsistent PPE monitoring. On a larger scale, many injuries can result from inadequate PPE usage.  Our team will be tasked with creating an intelligent vending machine system to combat this issue. The hardware consists of the vending machine itself, a camera to observe the employee obtaining PPE, an NVIDIA Jetson for processing power, a security gate, and the PPE inventory. The software consists of a neutral network utilizing computer vision. The goal is for the neural network to detect and determine if the worker is wearing or missing required PPE. If any required articles of PPE are missing, the vending machine will automatically dispense the missing article. Once the worker is wearing all the required PPE, an attached security gate will open and allow the worker to walk into the factory. A manual override will be installed in case the model makes a mistake and the worker. The team will primarily focus on the software and hardware aspects of the AI vending machine project. The client needs a system that can detect any missing PPE and dispense them automatically to the user. The hardware for this project will involve a camera, development board, and the vending machine computer. The camera and the development board need to be selected by the team. The camera should be able to provide consistent and clear images for the computer vision algorithm. The development board needs to have enough computer power to be able to handle complex and large computer vision models and additional middleware software. The hardware will directly interface with the software architecture that the team has chosen. The team will research and develop computer vision models for detecting missing PPE through learning and heuristic means. Additionally, the team will develop communications and middleware software so that each piece of hardware and software can effectively communicate with one another. Finally, the team will design a logic system that will interface with the vending machine to dispense the appropriate items.