This video is last year's rendition of the introductory lecture. Unfortunately, Zoom crashed during this year's delivery of this lecture, so a new video could not be generated. The contents is quite similar, however. The video and audio quality is not great, will be better on the new videos of the current course. 

This lecture gives a very fast conceptual introduction into key ideas of dynamical systems theory, starting from zero, but going to the Hopf theorem within a single lecture. The conceptual ideas are used throughout the course to design dynamical systems that generated robotic behavior.

This Matlab code illustrates the core idea of the numerical solution of differential equations in a foward Euler procedure. 

This variant of the simple simulator for differential equations reflects how such simulators would more typically be implemented in Matlab. 

This lecture makes use of the concepts of attractors, repellors, and their bifurcations to enable a robotic agent to generate motion toward targets, avoiding obstacles. I also review empirical work from Bill Warren's lab that shows that human movement paths can be well described in these terms.

Supports the lecture on the attractor dynamics approach for vehicle motion planning and the associated Exercise 2. [Read only until page 223 before "(3) Neural field dynamics"]

I review two methods for how to use attractor dynamics to generate paths for autonomous vehicles based on low-level sensory information.

Also a back-ground reading for the associated lecture on the sub-symbolic variant of the attractor dynamics approach. 

In this lecture I define some general concepts to characterize different approaches to motion planning and review a small number of approaches other than the attractor dynamics method of the previous lecture to position the attractor dynamics approach in that landscape. These include the potential field approach, the Borenstein-Koren vector-field method, and the dynamic window approach.

This lecture introduces into the problem of navigation, knowing where you are and how to get places, and links this to the problem of movement planning. I first clarify the concepts involved and then review a dynamical systems based approach to perform navigation, both the building of a map and its use to go to targets. I then provide some neuroscience background for this, reviewing the basic facts of place, grid, and heading direction cells and point to concepts from neuroscience are used to solve technical navigation problems.

In this lecture, Dr. Lei Zhang reviews properties of human voluntary movement of the upper limb and gives a survey over the neurophysiological foundations of human motor control.

This lecture discusses the problem of creating time courses for robotic motion, first in conventional robotic planning, then in autonomous robots. It reviews timing and coordination of human movement, which is a source of inspiration and concepts for conceiving of coordination as emerging from coupled limit cycle oscillators. A few robotic implementations of such dynamic ideas are reviewed, culminating the the landmark work of Aude Billard (which is not explained in any detail, however).

In this short lecture, I review the core idea behind the notion of Dynamic Movement primitives that goes back to Auke Ijspeert's work with Stefan Schaal and others. I only touch the basics here. This lecture builds on the lecture on timing and coordination that provides context.

This is a very brief and not too deep survey over issues in controlling robot arms including the dynamics of kinematics chains, principles of control, with repeated reference and contrast to human motor control. Most topics are only roughly outlined as a guide to this rather large field.

to help with exam preparation