I welcome you all to this presentation On:

Neural Network Applications

Imran Nadeem & Naveed R. Butt

220504 & 230353

Systems Engineering Dept. KFUPM

Part I:Part III: Control Applications Introduction to Neural Networks Part II: LMS & RBF

Part I: Introduction to NNs

There is no restriction on the unknown Neural networks are parameterized nonlinear functions whose parameters can function to be linear. Thus, neural be adjusted to achieve different shaped networks provide a logical extension to nonlinearities. create nonlinear adaptive control schemes. In essence, we try to adjust the neural Universalto serve as an approximator for network Approximation Theorem: neural an unknown function that we know only networks can reproduce any nonlinear through for a limited outputs function its inputs and input set.

Part I: Introduction to Neural Networks

Human Neuron

Part I: Introduction to Neural Networks

Artificial Neuron

Part I: Introduction to Neural Networks

Adaptation in NNs

Part I: Introduction to Neural Networks

Single Layer Feedforward NNs

Part I: Introduction to Neural Networks

Multi-Layer Feedforward NNs

Part I: Introduction to Neural Networks

Recurrent (feedback) NNs

A recurrent neural network distinguishes itself from the feed-forward network in that it has at least one feedback loop. For example, a recurrent network may consist of a single layer of neurons with each neuron feeding its output signal back to the input of all input neurons.

Part I: Introduction to Neural Networks

Recurrent (feedback) NNs

The presence of feedback loops has a profound impact on the learning capability of the network and on its performance.

Part I: Introduction to Neural Networks

Applications of NNs
Neural networks are applicable in virtually every situation in which a relationship between the predictor variables (independents, inputs) and predicted variables (dependents, outputs) exists, even when that relationship is very complex and not easy to articulate in the usual terms of "correlations" or "differences between groups

Part I: Introduction to Neural Networks

Applications of NNs
Detection of medical phenomena Stock market prediction Credit assignment Condition Monitoring Signature analysis Process control
Part I: Introduction to Neural Networks

Nonlinear Identification & Adaptive Control

End of Part I

Part II: LMS & RBF

LMS: The Adaptation Algorithm

RBF: Radial Bases Function NN

Part II: LMS & RBF

LMS: The Adaptation Algo.

Actual Response Estimated Response

Estimation Error

e[n] d[n] ~[n] s

Cost Function

J E{| e[n] |2 }

Mean Square Error

Weight Updates

w[n 1] w[n]

w J

Adaptation Step Size

Part II: LMS & RBF

RBF-NNs
Radial functions are a special class of functions. Their characteristic feature is that their response decreases (or increases) monotonically with distance from a central point and they are radially symmetric.

Part II: LMS & RBF

RBF-NNs

Gaussian RBF

Part II: LMS & RBF

RBF-NNs
Neural Networks based on radial bases functions are known as RBF Neural Networks and are among the most commonly used Neural Networks

Part II: LMS & RBF

RBF-NNs
Two-layer feed-forward networks. Hidden nodes: radial basis functions.

Output nodes : linear summation.

Very fast learning Good for interpolation, estimation & Classification

Part II: LMS & RBF

Part III: Control Applications

Nonlinear System Identification

Adaptive Tracking of Nonlinear Plants

Part III: Control Applications

Nonlinear System Identification

Part III: Control Applications

Nonlinear System Identification

y (t 1) 0.8606 y(t ) 0.0401y 2 (t ) 0.0017 y 3 (t ) 0.000125 y 4 (t ) 0.0464u (t ) 0.045 y(t )u (t ) 0.0034 y 2 (t )u (t ) 0.00025 y 3 (t )u(t ) 0.0012u 2 (t ) 0.0013 y(t )u 2 (t ) 0.0001458 y 2 (t )u 2 (t ) 0.00002083u 3 (t ) 0.00002083 y (t )u 3 (t )

Continuously Stirred Tank Reactor

Part III: Control Applications

Nonlinear System Identification

0.6 Plant Output NN Output 0.4

Simulation Results Using SIMULINK

0.2

-0.2

-0.4

-0.6

-0.8

20

40

60

80

Part III: Control Applications

100 Time

120

140

160

180

200

Adaptive Nonlinear Tracking

Part III: Control Applications

Adaptive Nonlinear Tracking

y (t ) 0.5 y (t 1) x(t 1) 0.1x(t 2) x(t ) 1 u (t ) u 2 (t ) 0.2u 3 (t )

Hammerstein Model

Part III: Control Applications

Adaptive Nonlinear Tracking

2 Reference Plant Output Model Output 1.5

Simulation Results Using SIMULINK

0.5

-0.5

-1

-1.5

-2 0

10

20

30

40

50 Time

60

70

80

90

100

Part III: Control Applications

Adaptive Nonlinear Tracking

1 0.8 0.6 0.4

Simulation Results Using SIMULINK

Control Signal

0.2 0 -0.2 -0.4 -0.6 -0.8 -1

10

20

30

40

50 Time

60

70

80

90

100

Part III: Control Applications

Thank you