Traditional computer vision systems perform real-time object detection by using abundant computational resources in the cloud. Our project explores running these detection models on edge devices such as the Raspberry Pi to perform human detection in a live video stream. Implementing this detection system on an edge device requires design optimizations to overcome the lack of network connectivity and perform object detection in a way that is computationally feasible at the edge without cloud resources.

Parq is an IOT network that will help you find parking downtown or in parking garages. It consists of Bluetooth low energy (BLE) and LiDar devices. Parq is able to give you real time data on parking spot availability. Through Parq’s android app users are able to view parking spot availability as well as get directions to a selected spot. With Parq, users will save time and money trying to find parking.

This project was intended to develop a means of viewing and understanding bearing currents that result from the induction motor within an Electrical Submersible Pump (ESP) system. Beyond that, it is meant to provide a way to test and modify the use of various bearing current mitigation methods to better understand how to avoid minimize them within these systems.

The point-of-use water purification system was developed with the ideas of portability, affordability, and ease of use for specifically populations in sub-Saharan Africa where access to clean water is limited. The driving technology behind our system is ultraviolet (UV) radiation. UV LEDs are used to kill bacterial strains as water is passed through the two stage system to be filtered and purified of impurities.

Compas is personalized awareness system that keeps you from ever losing your valuables. This is done by alerting the user when their item is getting too far away from their phone. This is accomplished with Bluetooth signal strength testing along with geofencing and an accelerometer to determine more accurately whether the user’s item is in danger.

he project involved predicting the downhole working parameters of an oil rig pump system given it's uphole characteristics through the use of a simulator and various deep learning techniques. The problem was solved through the use of a convolutional neural network, regression and decision trees, and a recurrent neural network (LSTM).

Our team is building a video streaming system which allows multiple robot teleoperators to use VR headsets to independently inspect a robot's environment. The system will operate in an Ubuntu Linux environment and a Robot Operating System (ROS) environment. This system will be useful in inspecting material conditions of facilities that are too dangerous for humans to enter, such as certain areas of nuclear facilities.

Swivalry is excited to introduce the first smart residential door opener. Our solution is an easy, no-drill installation product that is fully controlled through a simple iPhone app allowing users to open/close the door and securely share/revoke temporary access. The product uses a Raspberry Pi that communicates with AWS to process job requests sent through the 2FA-protected mobile app to open/close the door. We also provide a flexible payment plan and add additional layers of digital security through device certificates issued by AWS.

The goal of this project was to create an online device manager that owners of Dell EMC products could use to grant and revoke access to their device. This device manager allowed authorized users to have access to all of the management tools they would have if they were on-site with the server (without access to the data itself). These tools include virtual Keyboard-Video-Mouse (vKVM) and electronic file transfer.

The lack of extensive, labelled data sets representative of dynamic multi-agent environments currently hinders our ability to use machine learning techniques to develop autonomous driving systems. In order to conduct experiments on alternative algorithms for real-time collaborative sensing, there is a need for programmable and repeatable dynamic scenarios.