Contour of model predictive control
“How can we predict how a system will react based on how it reacts right now?”
When working with control systems, we often have some desired output in our mind. However, frequently the actual performance of our systems diverges greatly from what we want. So how can we use our engineering mindset to correct this problem? Well, let’s think about it. We can tell a computer how we want a certain system to behave. And we can also create a log of its outputs. So what if every time we gave an output, we took its data, compare it to our desired, and try to minimize the difference with the next iteration? Well, this is the fundamental idea behind model predictive control and is used in industries spanning from building controls to renewable energy to intelligent transportation systems!