Biomedical Engineering
School of ITEE

Researcher Profile

What is your area of research?

Algorithmic robotics. Basically, I develop the "brains" of autonomous robots. It's the software that enables robots to deliberately decide what to do (such as what strategy to take, what control parameters to use, what measurements are needed, what information to extract from sensing data, etc..) so that it can accomplish the tasks given by its "master" (e.g., by me : ) )

What are the major scientific challenges facing your area of research?

Handling uncertainty. Uncertainty is everywhere and is the most likely culprit when robots fail. For instance, GPS signals are often blocked and the robot's perception is far from perfect, causing them to have difficulties assessing their operating environment; predictions of wind and water currents are often wrong, which cause problems for most flying and marine robots whose motion is easily disturbed by these currents; customers have different needs, requiring a robot to adjust its service accordingly even though it may not know the exact requirements of its current customer, etc. Autonomous robots must decide how to accomplish the given tasks, despite uncertainty and errors in their sensing, perception, and control, and despite lack of information about their operating environment. Existing methods for handling uncertainty are quite conservative, in the sense that they exaggerate the risk of operation and confine the robots operation to tasks where the exaggerated risk is very low (almost zero). As a result, we have something like "bubble" robots (ref. to the Bubble Boy movie), which operate far below their capacities. Opposite to the conservative approach, I try to use probability theory to carefully quantify uncertainty and risks, to enable robots leave their "bubbles" and be less fearful, yet safe to be around. The probabilistic approach as a theoretical approach is not new, but, because it is notorious for its high computational complexity, it has been deemed impractical for robotics. However, it turns out if we exploit structures abundant in the physical world where robots operate, we can turn this theoretical approach into a practical application ( for more info on this).

Why did you become a computer scientist?

Algorithmic robotics is part of computer science, which is the only science where we are not limited by anything else other than ourselves. For instance, it is fine and doable to invent a world with a totally different law of physics (as long as the law is consistent within the system). Mathematics has this nature too, but computer science allows such a "made up" world to become something more concrete (e.g., in a game or a simulator). Applying computer science to robotics allows me to bring any "made up" world to be even more concrete, something tangible in the physical world, and it also helps me to be a bit more grounded to reality :)

Who has been/is the biggest influence on your career and why?

I would say is a 50-50 split between my dad and my junior high Arithmetic teacher. My dad taught me to see hurdles as nothing more than constraints and focus on "what to do to get around constraints". I think it's a very practical advice to avoid making a mountain out of a molehill and to save time. My Arithmetic teacher used to say "you need to put your dreams as high as possible, so that if you fail you'll still end up with something good". To me, the clause after "so that" makes this advice better than the various clich├ęs that teachers usually tell their students, and so it has stuck in my mind to this day. I've been using the combination of these two pieces of advice to make many decisions that lead to my current career, they have the biggest influence on my career.

If you could choose one person to sit next to on a plane, whom would it be?

Actually, if I could choose, I prefer no one sits next to me. This way, I get more space :) If I have to choose one person, I prefer a stranger who has a very different experience & background from the people I usually meet in my daily life, can speak a language I understand, and has a medium level of chattiness.

School of Information Technology
& Electrical Engineering