http://etiennechassaing.com/teaching/epfl/EPFLHugo Blox Builder (https://hugoblox.com)en-usSun, 01 Oct 2023 00:00:00 +0000http://etiennechassaing.com/media/sharing.jpghttp://etiennechassaing.com/teaching/epfl/http://etiennechassaing.com/teaching/epfl/legged-robots/Sun, 01 Oct 2023 00:00:00 +0000http://etiennechassaing.com/teaching/epfl/legged-robots/<p><img src="featured.png" alt="Legged robots"> <em>Legged robots</em></p> <p>I was a teaching assistant for this Master course at EPFL.</p> <h2 id="about-the-course">About the course</h2> <p>The design, control and applications of legged robots: a review of two-, four- and multi-legged robots and of the control methods for legged locomotion. Students also learn to analyse key papers in the field critically, and to design their own models and locomotion controllers in simulation.</p> <h2 id="topics">Topics</h2> <ul> <li>History of legged robotics: two-, four- and multi-legged robots</li> <li>Mechanical structures: passive and dynamic walkers</li> <li>Background concepts: static versus dynamic stability, stability criteria (zero-moment point, capturability…), energy consumption and cost of transport, state estimation</li> <li>Simple models of locomotion: rimless wheel, inverted pendulums (LIP, SLIP), template versus anchor models</li> <li>Control approaches: trajectory-based methods, virtual leg and virtual model control, hybrid zero dynamics, optimal control, planning, reinforcement learning and bio-inspired approaches</li> <li>Reading and presenting key papers of the field</li> <li>Numerical exercises: students build their own controllers for simulated legged robots, with weekly sessions with the assistants and the professor</li> </ul> <p><a href="https://edu.epfl.ch/coursebook/en/legged-robots-MICRO-507">Course description in the EPFL coursebook</a></p>http://etiennechassaing.com/teaching/epfl/foundations-of-ai/Fri, 01 Sep 2023 00:00:00 +0000http://etiennechassaing.com/teaching/epfl/foundations-of-ai/<p><img src="featured.png" alt="Taxonomy of artificial intelligence"> <em>AI taxonomy, source Google</em></p> <p>I was a teaching assistant for this course at EPFL.</p> <h2 id="about-the-course">About the course</h2> <p>The theoretical concepts behind machine learning, and a set of tools to apply it to problems from mechanical engineering.</p> <h2 id="topics">Topics</h2> <p><strong>Tools</strong></p> <ul> <li>Supervised learning: regression and classification</li> <li>Unsupervised learning: singular value decomposition, K-means</li> <li>Deep learning: a short introduction to neural networks</li> <li>Reinforcement learning: a short introduction to policy-gradient methods</li> </ul> <p><strong>Theory</strong></p> <ul> <li>Optimisation: the role of convexity, gradient descent, least squares</li> <li>Statistics: the Bayesian approach, the bias-variance trade-off</li> </ul> <p><a href="https://edu.epfl.ch/coursebook/en/foundations-of-artificial-intelligence-ME-390">Course description in the EPFL coursebook</a></p>http://etiennechassaing.com/teaching/epfl/mpc/Fri, 01 Sep 2023 00:00:00 +0000http://etiennechassaing.com/teaching/epfl/mpc/<p><img src="featured.png" alt="Diagram of the model predictive control principle"> <em>MPC principle</em></p> <p>I was a teaching assistant for this Master course at EPFL.</p> <h2 id="about-the-course">About the course</h2> <p>An introduction to the theory and practice of Model Predictive Control (MPC). MPC lets you specify time-domain objectives flexibly, optimise the performance of complex multivariable systems, and enforce constraints on the system&rsquo;s behaviour explicitly.</p> <h2 id="topics">Topics</h2> <ul> <li>Convex optimisation and the optimal control theory MPC builds on</li> <li>Receding-horizon control of constrained linear systems</li> <li>Practical issues: tracking and offset-free control of constrained systems</li> <li>Theoretical properties: constraint satisfaction, invariant sets, stability of MPC</li> <li>An introduction to advanced topics in predictive control</li> <li>A simulation-based project giving hands-on experience with MPC</li> </ul> <p><a href="https://edu.epfl.ch/coursebook/en/model-predictive-control-ME-425">Course description in the EPFL coursebook</a> · Figure: <a href="https://fr.mathworks.com/help/mpc/gs/what-is-mpc.html">MathWorks</a></p>