ECE3009 Neural Networks and Fuzzy Control L T P J C
3 0 0 4 4
Pre-requisite ECE2006 - Digital Signal Processing Syllabus version
1.0
Course Objectives:
1. To summarize basic learning laws and architectures of neural networks.
2. To describe supervised and unsupervised learning laws of Neural Networks.
3. To introduce Fuzzy Logic, Fuzzy relations and Fuzzy mathematics for designing a Fuzzy
logic controller.
4. To discuss neuro fuzzy approaches like ANFIS and CANFIS.
Course Outcomes:
1. To translate biological motivations into various characteristics of artificial neural networks
2. To comprehend and analyze basic learning laws of neural networks and activation functions
3. To interpret associative memories for storing and recalling the input patterns
4. To learn and implement supervised and unsupervised learning algorithms for various
applications.
5. To learn fuzzification and de-fuzzification methods for developing Fuzzy inference systems
6. To apply and integrate various neuro-fuzzy techniques for designing intelligent systems
using ANFIS and CANFIS.
7. To design a model using neural networks and fuzzy logic for various applications.
Student Learning Outcomes (SLO) 1,2,5
Module:1 Introduction to Artificial Neural Networks 3 hours
Artificial neural networks and their biological motivation, terminology, models of neuron,
topology, characteristics of artificial neural networks, and types of activation functions.
Module:2 Learning methods 7 hours
Error correction learning, Hebbian learning, perceptron – XOR problem– perceptron learning rule
convergence theorem – adaline.
Module:3 Supervised Learning 9 hours
Introduction to ANN architecture, multilayer perceptron, back propagation learning algorithm,
momentum factor, radial basis function network. Associative memory: Auto association, hetero
association, recall and cross talk. Recurrent neural networks - Hopfield neural network.
Module:4 Unsupervised Learning 9 hours
Introduction, competitive learning neural networks, max net, Mexican hat, hamming net,
Kohonenself organizing feature map, counter propagation, learning vector quantization, adaptive
resonance theory, performance of SOM.
Module:5 Fuzzy Sets and Fuzzy Relations 4 hours
Introduction, classical sets and fuzzy sets, classical relations and fuzzy relations, membership
function.
Module:6 Fuzzy Inference Systems 6 hours
Fuzzification, fuzzy arithmetic, numbers, extension principle, fuzzy inference system,
defuzzification, fuzzy rule based systems, fuzzy nonlinear simulation, fuzzy decision making, fuzzy
optimization.
Module:7 Neuro-Fuzzy Systems 5 hours
Introduction, ANFIS, ANFIS as universal approximator, CANFIS.
Module:8 Contemporary issues 2 hours
Total lecture hours: 45 hours
Text Book(s)
1. J.S.R. Jang, C.T. Sun, E. Mizutani, “Neuro Fuzzy and Soft Computing - A computational
Approach to Learning and Machine Intelligence”, 2012, 1 st edition, PHI learning Private
Limited, New Delhi.
2. Timothy J. Ross, Fuzzy Logic with Engineering Applications, 2016, 4 th edition, John Wiley
and sons, USA
Reference Books
1. Jacek. M. Zurada, “Introduction to Artificial Neural Systems”, 2014, 11 th edition, Jaico
Publishing House, Mumbai.
2. Simon Haykin, “Neural Networks and Learning Machines”, 2016, 3 rd edition, Pearson
Education Inc. India
3. Samir Roy, Udit Chakraborthy, “Introduction to Soft Computing Neuro - Fuzzy and Genetic
Algorithms”, 2013, 1 st edition, Pearson education, Noida.
Mode of Evaluation:Internal Assessment (CAT, Quizzes, Digital Assignments) & Final
Assessment Test (FAT)
Typical Projects
1. Adaptive filtering for Medical (ECG) signals.
2. Adaptive Neuro Fuzzy Inference System
3. Automation of Traffic signal using Raspberry Pi
4. Cardiac Image Diagnostic System
5. Cryptographic System using Neural Networks
6. Design and Development of Biometric Recognition and Matching System
7. Digital Audio Watermark Embedding System
8. Electrical load forecasting using Neural Networks
9. Electronic Music System using ANN
10. Face Identification System using ANN
11. Feature Extraction of EEG Signals
12. Image Decryption using Neural Networks
13. Internal Fault identification using Artificial Neural Network
14. Signature Forgery and Handwriting Detection System
15. Smart Driver Assist System using Raspberry Pi
16. Speaker Recognition using Soft Computing
17. Speech Separation Using ICA Based Neural Networks
Mode of evaluation:Review I, Review II and Review III
Recommended by Board of Studies 13/06/2015
Approved by Academic Council No. 37 Date 16/06/2015
