Two ECE Senior Design Teams in Intel-Cornell Cup Finals

in NEWS, Students

Projects Address Everyday Problems with Embedded Technology

By Mark Dwortzan 

Team2
Omar Rana (CE), Luke Poitras (EE), Caroline Wu (EE), Jason Tow (CE, Jeremy Bensabat (CE). Photo by Gabriella McNevin

Two Electrical & Computer Engineering senior design teams have been named finalists in the Intel-Cornell Cup 2015 competition, which challenges science and engineering college students to conceive of, design and demonstrate the next great embedded technology application. One team’s project, C.A.R.R. System (Cyclist Alert Real-time Response), notifies drivers of potential collisions with approaching cyclists. The other, GrowBox, is an automated hydroponic device that enables users to grow an edible plant, virtually carefree.

A high-level illustration of the C.A.R.R. team's prototype. The system uses two cameras (one mounted on each side-view mirror) that feed a detection algorithm housed within an alert hub mounted on the dashboard. The alert hub processes live video of the road, determines if a cyclist is present and delivers feedback to the driver.
A high-level illustration of the C.A.R.R. team’s prototype. The system uses two cameras (one mounted on each side-view mirror) that feed a detection algorithm housed within an alert hub mounted on the dashboard. The alert hub processes live video of the road, determines if a cyclist is present and delivers feedback to the driver.

The C.A.R.R. System and GrowBox teams will attend talks, network with leading engineering firms and showcase their work along with 20 other finalists from across the country on May 1-2 at NASA Kennedy Space Center. They’ll vie for the competition’s grand prize, $10,000 or one of up

to seven $2,500 awards, all of which include an invitation to exhibit in Intel’s booth at the Maker Fair in New York City or San Mateo, California.

Having survived an hour-long, online semifinal round in February to make it to the finals, both teams subsequently received

$1,500, Intel Atom boards and other equipment, and access to technical experts at Intel and other sponsoring companies to develop their systems.

“Both teams are passionate about their projects and are dedicated to using their engineering skills for the betterment of society,” said Associate Professor of the Practice Alan Pisano(ECE), the lead faculty member for the ECE Senior Design Project course. “They are continually seeking ways to improve their designs, and it’s rare not to see them in the lab working on aspects of their projects.”

Concerned about the rising number of annual bicycle accidents in Boston and other metropolitan areas, the C.A.R.R. System team aims to equip motorists with a bike detection system that consists of cameras attached to both side-view mirrors and a real-time image-processing algorithm. When the system pinpoints a potential or impending collision through the algorithm, it displays and announces a warning on an alerting device that’s easily mountable on the dashboard. Issued within about 200 milliseconds from the moment of detection, the warning indicates which side of the vehicle is on a collision course with an approaching cyclist. In a recent test producing one hour of sample footage, the system successfully identified 92.55 percent of cyclists present, with an overall accuracy of 83.65 percent.

team5
Alexandru Rosca (CE), Patrick Crawford (CE), Mark Barrasso (CE), Ahmed Alfuwaires (EE), Jesse Fordyce (EE). Photo by Gabriella McNevin

Testing out several designs and detection algorithms, the team settled on a dual camera system with a single, centralized alert hub, and an algorithm that provided the most accuracy and fastest response time.

“After living in Boston for four years, we are very aware of the dangers that exist on the road for drivers and cyclists alike,” said C.A.R.R. team member Omar Rana (CE). “We wanted to create a product that could fit both old and new vehicles, be easy to install and remove, and theoretically reach the market at an affordable price.”

Conceptual model of GrowBox prototype
Conceptual model of GrowBox prototype

Seeking to help would-be vegetable gardeners who lack the space, time or requisite green thumb to grow their own food, the GrowBox team has designed an automated system that can see to a plant’s needs and report on its status through a combination of sensors, actuators and image processing software. If human intervention is required, an iOS app will notify the user with instructions. GrowBox consists of a hydroponic subsystem that periodically floods the plant with water and nutrients; red, white and blue LEDs tuned to provide optimal lighting conditions for the plant; a backend subsystem that controls all sensors, actuators and lights; and a cloud-based database that backs up all sensor and image data. GrowBoxes are designed to be stackable so a user could grow a column of vegetables in a compact space.

The team’s biggest challenge has been to create and maintain the water/nutrient solution that’s needed to sustain the plant. To solve the problem, they found a nutrient mixture that keeps pH, electrical conductivity and other essential GrowBox parameters constant.

“Together, we developed ideas for the GrowBox and the reduction of the user’s role in the growth of a plant,” said Sasha Rosca (CE), who came up with the idea for the project. “Once the automation technology is developed, it can be implemented in large grow houses around the world to provide people with food year-round.”