Robot Learning & Sensorimotor Control (RLSC)

Course Description

Control of complex, compliant, multi degree of freedom (DOF) sensorimotor systems like humanoid robots or autonomous vehicles have been pushing the limits of traditional control theoretic methods. This course aims at introducing adaptive and learning control as a viable alternative. The course will take the students through various aspects involved in motor planning, control, estimation, prediction and learning with an emphasis on the computational perspective. We will learn about statistical machine learning tools and methodologies particularly geared towards problems of real-time, online learning for sensorimotor control. Issues and possible approaches for learning in high dimensions, planning under uncertainty and redundancy, sensorimotor transformations and stochastic optimal control will be discussed.This will be put in context through exposure to topics in human motor control, experimental paradigms and the use of computational methods in understanding biological sensorimotor mechanisms.
This MSc course (designed as a follow up to the introductory course on robotics (R:SS) in Semester 1) will gear students towards specialized topics in robot control and planning as well as human motor control from a machine learning perspective and is a must for students looking to pursue a post-graduate degree in robotics or human motor control.

Level 11 SCQF Official Course Descriptor

When and Where?

Semester: 2 Time: 10:00-10:50
Location: Monday (Appleton Tower AT2.4 ) and Thursday (AT4.12)

First Class: Jan 14, 2013

Lecturer

Prof. Sethu VIjayakumar

Professor of Robotics and Director, IPAB, School of Informatics.

Syllabus

(Refer to the Lecture Notes tab on the left for the detailed schedule and resources - updated weekly)

Machine Learning Tools for Robotics

- Regression in High Dimensions
- Dimensionality Reduction
- Online, incremental learning
- Multiple Model Learning
Adaptive Learning and Control
Predictive Control
Movement Primitives
- Rhythmic vs Point to Point Movements
- Dynamical Systems and DMPs
Planning and Optimization
- Stochastic Optimal Control (2)
- Bayesian Inference Planning
- RL, Apprenticeship Learning and Inverse Optimal Control
Understanding Human Sensorimotor Control
- Force Field and Adaptation
- Optimal control theory for Explaining Sensorimotor Behaviour

- Cue Integration and Sensorimotor Adaptation

Assessment

One Homework Assignment - 30%
One Oral Presentation - 10%
Final Exam - 60%

Background and Prerequisites

The student taking this course should ideally have had some prior exposure to robotics basics (such as R:SS) and/or machine learning basics.
The student should have some feel for the formulation and use of mathematical models and possess sufficient mathematical maturity in order to be able to follow some readings from the research literature. On the practical side, one of the assignments will require programming, in an environment such as MATLAB.

Personal tools