High-Flying Dreams

DRONES TODAY ARE COMMONPLACE. Wedding photographers use them to take aerial videos, some 300 of them took to the skies to dance in unison during last year’s National Day Parade, and they were mentioned by Prime Minister Lee Hsien Loong during his recent May Day speech to illustrate how technology is disrupting industries. Given how you can walk into a computer store and find camera-equipped drones starting from $50, it may then seem strange to think NUS has a research group looking into this.

Professor Ben Chen of the Department of Electrical and Computer Engineering takes pains to point out, though, that he is not doing research into toys. Part of the problem is that the term “drone” is used loosely. Prof Chen, who is also a Provost Chair, is leading research into unmanned aerial vehicles (UAVs), with an emphasis on “unmanned”. While a remote-controlled plane or quad-copter relies on a human pilot to operate, UAVs fly autonomously, relying on onboard computers to complete a mission.

LOOK, NO HANDS!

Today, UAVs are used for a variety of tasks: Some are routine functions like stocktaking, surveillance and the spraying of pesticide on crops, while others are literally lifesaving because the UAVs are used for firefighting and search and rescue.

Prof Chen is one of the advisers of the Unmanned Systems Research Group at NUS. The group is on the cutting edge of UAV research and it has done well in various local and global competitions. Just last year, the team was named the indoor and outdoor champions at the International Micro Air Vehicles (IMAV) Competition held in Toulouse, France. IMAV is a yearly event aimed at fostering key technologies for the development of micro-air vehicles.

Apart from continuously refining the different systems needed for autonomous mission completion, the group also develops innovative hardware to handle different environments. Two PhD students in the group spent four years developing a hybrid UAV that is both a helicopter and an aeroplane. The U-Lion, which was unveiled last year, can lift off the ground like a helicopter. In mid-air, the wings unfold and the craft changes orientation so that the rotors of the helicopter become the propellers of the plane. All this is done autonomously. The hybrid UAV combines the advantages of both helicopters and aeroplanes. Because it can do vertical takeoff and landing, it eliminates the need for a runway. In addition, it is able to hover. At the same time, it can fly greater distances when configured as an aeroplane.

The research being done by the group is being commercialised through a spin-off company called AeroLion Technologies where Prof Chen is also an adviser. AeroLion Technologies builds UAVs for both indoor as well as outdoor use. Its drones have been employed in projects such as sewage tunnel inspections as well as warehouse management. The project with the national water agency PUB to use a UAV to inspect sewage tunnels was a particularly challenging one because they could neither use a Global Positioning System (GPS), as it was underground, nor LIDAR, due to weight issues. The UAV had to rely entirely on computer vision systems to autonomously get around within the tunnel. AeroLion has also worked with YCH Singapore to use UAVs for inventory management in warehouses. Using an autonomous drone to do inventory management in a large warehouse is faster, safer and more accurate that sending a human up on a cherry-picker to do the laborious job of stocktaking.

REACHING FOR THE SKIES

Interestingly enough, although he is a well-known name in the field of autonomous flying technology today, Prof Chen is not an aeronautical engineer by training. His undergraduate degree was in mathematics and computer science, which he obtained from Xiamen University in China. He did his masters in electrical engineering in Gonzaga University in Spokane, Washington and completed his PhD in electrical and computer engineering at Washington State University. (His time in Washington State has left its mark on him and till today, he is a fan of the Washington State Cougars, the American football team of his alma mater.) After a year teaching at the State University of New York in Stony Brook, he came to Singapore to teach at NUS. This year marks the 25^th anniversary of his time here.

When he first started doing research on UAVs 15 years ago, the field was in its infancy. Armed with a grant, Prof Chen and his team began by looking for a remote-controlled helicopter to buy. Yamaha offered them one that could carry a 50kg payload. The first catch was that it would cost US$150,000. The second catch was they would have to buy two of these helicopters. “Instead, we bought remote-controlled helicopters from a hobby shop for S$2,000 each.”

Dr Chen and his team then set out to build all the systems needed from scratch. They also wrote their own software to integrate all the systems and sensors. Because they were forced to do everything, both hardware and software, they learned how to handle the whole spectrum of systems needed to operate a UAV. And while the lack of an aeronautical background meant that it took time to get things off the ground in the beginning, there were also upsides to being unschooled. “The good thing is that if you don’t have preconceived notions, you have broader ideas,” says Prof Chen. “The way we do the automatic control systems is quite different from others.” The team also spent six years building a mathematical model of a helicopter. “If you have a very accurate mathematical model of a UAV, you know its behaviour, its limitations and its capacities. You don’t have to go for actual flight test. Normally we do a simulation first on the computer, then we go out and test it out.” There are now over 10 people working in the research group, consisting largely of post-graduate students. In addition to doing research on UAVs, they are now branching out into unmanned surface vehicles and unmanned underwater vehicles.

Looking further ahead, Prof Chen believes that artificial intelligence (AI) will have a big impact on UAVs in future. He is now looking at how to incorporate AI to improve the autonomy of UAVs. “We want to make unmanned systems as smart as possible.” Big data is another field that is important. UAVs can gather large amounts of information so the challenge is to find a way to extract usable information from large data sets.

Research into UAVs continue apace. Driven by passion, Prof Chen and the Unmanned Systems Research Group will undoubtedly continue to keep the NUS flag flying. Autonomously, of course

AUTONOMOUS AND AIRBORNE

To operate without human intervention, an unmanned aerial vehicle (UAV) has an automatic control system that manages the various avionics and sensors on board. Among other things, a UAV has a measurement system so it can detect its position and velocity, a way to identify objects around it using computer vision or LIDAR (Light Detection and Ranging, which uses light pulses to measure distance), and an inertia measurement unit (IMU) to measure heading angle, rotating angle and acceleration.