Scribd
Upload
64 views45 pages

Artificial Intelligence in Robotics-51-95 - Compressed

Uploaded by

Mustafa Alhumayre
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
64 views45 pages

Artificial Intelligence in Robotics-51-95 - Compressed

Uploaded by

Mustafa Alhumayre
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 45

Example:

3
4
Figure. Delta learning rule.

5
Example:
Learn the continuous type bipolar sigmoid neural network one epoch using Delta
learning rule. The set of input training vectors is as follows:

and the initial weight vector is:

the learning constant is c=0.1 and λ=1. The teacher’s desired responses for x1, x2, x3
are d1=-1, d2=-1, and d3=1 respectively. Draw the schematic diagram of this neural
network. Calculate the Root Mean Square (RMS) normalized error for this epoch.

Solution:
According to the size of the input training vectors (3*1, this means three inputs) and
the size of the desired output (1*1, this means one output), the schematic diagram of
this neural network is:

The bipolar sigmoid activation function is:


( ) -1,

Moreover, for linear activation function f(net)=1.

6
….
Note:
The training algorithm stops when either the number of required epochs reached or
the training error in each epoch reaches the required (minimum) error (Goal).

7
Artificial Intelligence in Robotics
(Lecture 5)

4thyear, Mechatronics and Robotics


Engineering
1st Semester

By:
Dr. Mohammed Y. Hassan

Control and Systems Engineering Department


University of Technology - Iraq

1
Thus, the RMS error for the previous example is (P=3, K=1):

√∑

(H.W. Calculate the second and third epochs for the previous example).

2
Delta Learning Rule for Multi Preceptron layer

3
4
and

Generalized Delta learning Rule:

5
6
7
Single-Layer Feedback Neural Networks:

Mathematical Foundations of Discrete Hopfield Networks

8
9
Artificial Intelligence in Robotics
(Lecture 6)

4thyear, Mechatronics and Robotics


Engineering
1st Semester
By:
Dr. Mohammed Y. Hassan

Control and Systems Engineering Department


University of Technology - Iraq

1
Associative Memory:

2
Tutorials

3
Artificial Intelligence in Robotics
(Lecture 7)

4thyear, Mechatronics and Robotics


Engineering
1st Semester
By:
Dr. Mohammed Y. Hassan

Control and Systems Engineering Department


University of Technology - Iraq

1
Fuzzy Logic Control
Introduction:

2
3
Example: (surface of rules)

4
5
Types of Membership functions (MF):
The following represent some of the types of Membership functions;
there are other different shapes of membership functions:

6
7
8
Artificial Intelligence in Robotics
(Lecture 8)

4thyear, Mechatronics and Robotics


Engineering
st
1 Semester

By:
Dr. Mohammed Y. Hassan

Control and Systems Engineering Department


University of Technology - Iraq

1
Fuzzy Logic Control

2
Example:

3
4
Example:

5
Example:
For the following implied Fuzzy set, calculate the defuzzified value using COG
method.

Solution:

6
Example:
For the following implied Fuzzy set when the product is used to represent
the implication, calculate the defuzzified value using Center Of Average
(COA) method.

Solution:

7
Example:
For the previous example, calculate the defuzzified value using MAX
criterion method:

Solution:

8
Artificial Intelligence in Robotics
(Lecture 9)

4thyear, Mechatronics and Robotics


Engineering
1st Semester

By:
Dr. Mohammed Y. Hassan

Control and Systems Engineering Department


University of Technology - Iraq

1
Fuzzy Logic Control

2
Fuzzy Logic Control

3
4
Types of Fuzzy Logic Controllers
The FLC can be used to design different types of con trollers as in the following:

P-Like Fuzzy Controller:

, Where Ku is the output scaling factor.

PD-Like Fuzzy Controller:

5
PI-Like Fuzzy Controller

You might also like