Technology explainer

Using dynamic programming, as introduced previously, we eventually found the shortest route to the summit. We are finally at the top of the mountain! Naturally, many of us will directly reach for our phones. Perhaps to take a selfie with friends, to construct a panorama of neighbouring mountain tops, or maybe to record a video of some alpine birds effortlessly gliding around.

We’ve made it up from the valley, thanks to our gondola and optimal control. Finally, we will plan our path to the summit! In this post we will do precisely that, but in a systematic way. It turns out that this connects to reinforcement learning.

After going up in the gondola and appreciating how smooth the ride was (in part, thanks to automatic control!), lets quickly grab a coffee before we ascend the mountain.

In this second stage we leave the valley and move a bit higher up. Of course, we could walk or climb, but let us enjoy a remarkable piece of engineering and catch a gondola (cable car).

We start our hike in a typical Swiss valley: tidy green fields, some cozy chalets, but most importantly, a herd of curious cows that welcomes us.

In our new series of explainers, let Wouter Jongeneel walk you through both Switzerland and automatic control.

If you read about automatic control, you've definitely heard of Model Predictive Control (MPC), but what is it? Let's find out by seeing how some of our researchers are applying it on trains.

You've probably heard of reinforcement learning (or RL for short) if you've spent any time reading about technology or control theory, but what does it really mean? Let us explain.

If you've heard about game theory but don't really know what it is or what it means, this game by Nicky Case is a great starting point. (please note that Nicky Case is not associated with NCCR Automation - we just think this is a great resource).

While neural networks often achieve impressive performance, they sometimes fail to capture simple concepts that humans understand intuitively – a gap that may even put lives at risk. But can we transfer our human knowledge to neural networks to help them learn more meaningful solutions? NCCR Automation researcher Loris Di Natale has a plan to make sure automated machines don’t freeze us out.