UBC Science’s Post

Transcript

Broadly what I do falls under the umbrella of what we call real time systems because you're talking about these tiny computers, which we call embedded computers, which are embedded inside a device. And that might be controlling a significant part in your car. in an airbag when a crash occurs, it triggers the airbag deployment system. Now the airbag must be deployed at a very specific point of time. So even a slight delay in terms of a few milliseconds here and there could be fatal for the passengers. Our goal is essentially to come up with hard mathematical proof that something like this would never happen. Almost all the times when you're dealing with complex software, you need to then start taking help of rigorous mathematical techniques. I see mathematics as a tool to kind of formalize what you do in computer science and to design software more rigorously and to make claims about your software that you cannot do by just testing or running the software for a few minutes. And that's the only way to convince a safety critical system like a car developer or an airplane manufacturer that your software is correct. So when I talk about airbags, drones, airplanes, even robots, so all of these are examples where there is very tight integration between the physical world and the software world, which is computing all the values. And because of this tight integration, these timing requirements arise because physics is not going to stop for anyone. So your software better keep up with the physical world.