Engineering at Illinois Students Enter Drone in National Challenge

A team of Engineering at Illinois students, including six Aerospace Engineering undergraduates, have designed and built a small drone aircraft to participate in a national Unmanned-Aerial-Surveillance (UAS) competition.

Majoring in Electrical and Computer Engineering (ECE), Mechanical Sciences and Engineering (MechSE), and Computer Science (CS), as well as in AE, the 13 team members have been working since February to create an autonomous aircraft and control system to compete in the UAS Video Tracking Challenge. Organized by the Air Force Research Laboratory, Texas A&M University, and Raytheon, the challenge seeks to address current and future technologies for intelligence, surveillance and reconnaissance (ISR) missions.

Team members

Unmanned aerial vehicles and video motion tracking have been common in the U.S. military for a decade, but the vehicles have been controlled from the ground. “We’re trying to put all the control and computations within the aircraft, itself,” said team leader Xichen Shi, an AE senior. That way, he said, “If we start the mission and lose communication (with the aircraft), we could finish the mission.”

The students have developed a solution they’ve named Aerial-Based Intelligent Surveillance System (ABISS) to make their electrically-powered aircraft autonomous.

Human monitoring is not necessary with ABISS once surveillance tracking starts, although human operators can continue to command the aircraft if the communication link remains open. Equipment on board the craft will do essential computations.

Deploying ABISS will be a plane weighing less than 15 pounds. Shi said the smallness of the plane and onboard tracking system have made it more feasible for university groups to participate in this type of challenge. “Sensors and computational equipment are now so miniature, they can be held in a cell phone-sized box, so we can build a smaller plane,” he said. “You can’t keep a big system in a lab. If it’s over 50 pounds, you need more storage and would have to run the plane on gas rather than electric.”

Although sized to fit for the small aircraft, the ABISS can be adapted to any size airframe with compatible hardware. The onboard operating system is equipped with advanced parallel scheduling to increase the system robustness. The main system supervises each subsystem to ensure minimal waste of resources. The system also employs hardware redundancy to increase its survivability.

The ABISS motion video tracking capability utilizes cutting-edge research algorithms involving a three-stage process:

• The tracker will follow known features from frame to frame using vision descriptors and optical flow.
• The detector will match up features from processed frames and ensure target visibility.
• The learner will organize and distinguish the target to ensure algorithm robustness. Some experimental works also includes adapting target shape to improve tracking efficiency.

Using an industrial-grade PCIe/104 series embedded computer, ABISS will be able to expand its functionality through numerous compatible PCIe/104 stackable modules, as well as normal PCI/PCIe devices. This feature also allows external hardware to be added onto ABISS without major modifications.

Separately, both the aircraft and the tracking system have tested successfully, with the team using a camera on a video of a moving car to test the tracking system. The two components will be put together and tested this summer, Shi said. The UAS Challenge will take place in October at Texas A&M University, with 11 teams competing.

The challenge will include these phase:

• The aircraft must be able to track and follow a vehicle on the ground, moving on a straight path.
• The tracking task will then become more difficult when buildings and shadows are added as obstacles.
• The aircraft then must track a person carrying a suitcase as the person exits the ground vehicle and moves away from it.
• Although the team’s work began in response to the Challenge, Shi said the project will not end when the competition is over. “There’s more to be done afterward,” he said. “The industry of personal drones is just getting started, and what we’re doing here would be ideal for that kind of work.”
• AE Assistant Prof. Soon-Jo Chung and ECE Prof. Seth Hutchinson have advised the team, along with AE postdoctoral research associate Ashwin Dani. Rockwell Collins has helped sponsor the team’s work.
• Team members are:
• Xichen Shi, AE senior
• Sunil Patel, AE junior
• Mario Suarez, AE senior
• Patrick Chia, AE junior
• Mathias Kristian, ECE freshman
• Elias Waddington, AE freshman
• Guanyang Luo, MechSE graduate student
• Yangyang Yu, ECE sophomore
• Rasheed Ibrahim, AE senior
• Shubham Gupta, ECE sophomore
• Shaun Toomey, CS junior
• Hussain Alzeera, CS junior
• Shravan Gupta, ECE freshman