Thank you for the introduction, Daniel.

And welcome everyone to my talk.

So today, the title of the talk is Biomechanics

Meets Optimal Control, Predictions of Human Motion.

So I'm going to talk a bit about predictions

of human movement.

But then I'm also going to explain

how these could be improved with AI

and how these can improve database or machine

learned models.

So first of all, why would you want

to predict human movement?

And so one big field of interest for this

is design of wearables, such as shoes, running shoes,

or prosthetics, or exoskeletons.

And so currently what happens with these,

when these are designed, it's a process

where someone comes up with a smart idea,

designs it, prototypes it, and then it's

tested on human participants.

And then based on those tests, the design is improved,

and this circle goes on and on until a nice product is

designed.

However, this takes participants, this takes time,

this takes material.

So in fact, if we could add simulation to this cycle,

we could already optimize prototypes a lot better,

and we could reduce the time required, possibly injuries

on human participants.

And also save a lot of money on materials.

And this is also already done in many fields,

such as the automotive or aerospace industry,

where these simulations are used to optimize engine designs,

for example.

Another possible thing where this can be useful

is state estimation, such as here,

the Adidas Gamer system.

And so these are systems like Fitbit,

where you have a single sensor, for example, on your wrist

or in your phone or on your soccer shoe.

And what we could do with these simulations

is actually extract a lot more information

about the biomechanics of the movement

from these single sensors, which could reduce the chance

of developing injury.

And then the final, yeah, another thing

is just to understand human movement.

Because we know from Monty Python's Ministry of City

Walks that there's many different ways

that we could move, that we could walk from A to B.