Unit name | Intelligent & Adaptive Systems (UWE) |
---|---|
Unit code | MENGM0009 |
Credit points | 15 |
Level of study | M/7 |
Teaching block(s) |
Teaching Block 1 (weeks 1 - 12) |
Unit director | |
Open unit status | Not open |
Pre-requisites |
None |
Co-requisites |
None |
School/department | Department of Mechanical Engineering |
Faculty | Faculty of Engineering |
This unit is provided by UWE.
Introduction: Review of the links with other disciplines, e.g. classical AI, psychology, robotics, ethology, neuroscience and classical control. Scope and limitations of this module, especially with respect to classical control and AI. Learning and adaptive systems: Working definitions of intelligence, adaptive systems and learning. Adaptation through learning versus design. Basic Architectures: Neural networks. Fuzzy systems. Evolutionary computation. Supervised, unsupervised and reinforcement learning. Compound Architectures: Classifier Systems. Behaviour-based systems. Agent-based systems. Multi-agent systems. Example applications: Review of work carried out in this Faculty, and at other establishments, in order to demonstrate the major strengths and weaknesses of the techniques. For example; intelligent multiple agents for fault diagnosis in electrical power distribution systems, fuzzy control of an automated underground transportation system, co-operative behaviour in multi-agent mobile robotics, neurocontrol of an industrial robot manipulator, fuzzy classifier systems for telecommunications network routing, evolutionary computation as an aid to engineering design, human face and handwriting recognition using neural networks.
On completion of this module a student will typically be able to:- Show a detailed knowledge and understanding of
Demonstrate subject specific skills with respect to
Show cognitive skills with respect to
Demonstrate key transferable skills in
Lectures will introduce the fundamental concepts. Tutorial case study sessions will be used for two purposes. They will be used to expose students to demonstrations of the basic architectures in action. They will also be used to discuss real implementations of these new techniques, each designed to illustrate the essential details of a particular concept or technique, and especially its strengths and weaknesses in both technical and business contexts. At all times specific examples will be used to "ground" the theory.
Assessment Weighting between components A and B A: 50% B: 50%
ATTEMPT 1 First Assessment Opportunity Element Description Element Type % of Component % of Assessment Component A (Controlled Conditions) Examination (180 mins) Exam 100% 50% Component B Assignment 1 Coursework 50% 25% Assignment 2 Coursework 50% 25%
Second Assessment Opportunity (further attendance at taught classes is not required) Element Description Element Type % of Component % of Assessment Component A (Controlled Conditions) Examination (180 mins) Exam 100% 50% Component B Written Assignment Coursework 100% 50%
SECOND (OR SUBSEQUENT) ATTEMPT Attendance at taught classes is not required.
Indicative Reading List The following list is offered to provide validation panels/accrediting bodies with an indication of the type and level of information students may be expected to consult. As such, its currency may wane during the life span of the module specification. However, CURRENT advice on readings will be available via other more frequently updated mechanisms. Internet sources: There is a very wide range of up-to-date information on these subjects, at the appropriate introductory level for this module, available on the internet. Students will be provided with a Compact Disk that guides their access to approved and publicly available sites (such as the Evonet “Flying Circus” site for evolutionary computation). There are also some older books that still represent excellent introductions to this subject area.
White & Sofge (1992). The Handbook of Intelligent Control, Van Nostrand-Reinhold
Beal, R & Jackson, T (1990). Neural Computing - an introduction, Adam Hilger
Rao & Rao (1995). C++ Neural network and Fizzy Logic, 2nd Edition, MIS (ISBN 15585515526)
Brown & Harris (1994). Neurofuzzy Adaptive Modelling and Control, Prentice Hall (ISBN 0131344536)
Miller, Sutton & Werbos (1991). Neural Networks for Control, MIT Press
Arbib, M.A (1995). The Handbook of Brain Teory and Neural Networks, MIT Press
Design tool user manualse.g. MATLAB 6.2 Fuzzy, Neural Network, and Simulink Toolboxes,