MOBILE ROBOTICS RESEARCH GROUP

teaching

people

research

publications

teaching

robots

robot movies

links

Contact: Bridget Hallam
Last updated: Fri Jan 12 13:56:57 2001

Intelligent Sensing and Control (ISC)

The aims of this course are to introduce the fundamental problems associated with producing real world intelligent behaviour in artificial systems, together with some of the different kinds of information processing techniques and control architectures that have been developed.

The objectives are to provide students with a sound understanding of the underlying principles and assumptions adopted by different approaches to intelligent sensing and control, and to equip them to pursue further details and related topics in the literature.

To expose these problems in real terms the course is structured around a practical-based programme involving the construction of a series of small mobile vehicles of increasing sensorimotor sophistication. A series of lectures will present and discuss taught material covering related sensing and control ideas, approaches, organisational architectures, technologies, and current controversies from the literature and active robotics research projects within the Division.

Lego vehicles

We use a basic Lego car kit with a few extras (such as large gears), and some in-house electronics to drive motors and sensors. In 2000 the sensors available are:

• microswitches -- robust, reliable, but only giving very local, binary, information
• infra-red sensors -- which give an idea of the distance to the nearest object, but the value returned also depends upon the object's reflectance
• light-dependent resistors -- which return the total light intensity falling on them
• whiskers -- which vibrate as the robot moves so give many false positive readings
• a hall-effect sensor -- which detects if the vehicle is stuck but requires very careful placement and gives false negatives

Both sensors and Lego are imperfect, producing strong constraints on vehicle design and strategy. The electronics also contribute constraints, in that the boards are sizeable and relatively fragile. We hope to teach students something about working within constraints, the inter-relationship between hardware and software, the effect of the environment on design, the effect of robot morphology on behaviour, the frustrations of working with real hardware, the complications of working with parallel (or rather, pseudo-parallel) code, and how to incorporate unplanned effects into desired strategy.

The practical sessions are extremely challenging, frustrating, rewarding, ingenuity-developing, swearing-inducing, time-consuming, and fun. A robot capable of fulfilling the basic requirements of the assignments can be made relatively easily, but the element of competition drives people to improve their robot beyond this. The practicals are designed to be a good introduction to the real world where, although the equipment may be more sophisticated, the same basic frustrations of it not working as intended apply.

Related courses which may be of interest to mobile robotics students include

• Intelligent Assembly Systems
• Computational Vision
• Machine Vision
• Genetic Algorithms and Genetic Programming
• Neural Computation

ISC is just one of about 45 specialised modules available to Master's, 4th year, and 3rd year students. Some modules are compulsory, some have restricted access, the details vary depending upon the degree for which you have registered.

If you have further enquiries (or would like a hardcopy version of the prospectus) then please e-mail phd-admissions@inf.ed.ac.uk.

Finally if you have any specific enquiries, you might try mailing the mobile robotics group directly.