Download as pdf or txt
Download as pdf or txt
You are on page 1of 1

EE659 - Computer Assignment 1

Submission Deadline: 05.10.2015


Each question carry 25 marks. 15 out of 25 will be given based on the report (hard copy), neatness of
the plots and efforts put by the students. 10 out of 25 will depend on the correctness of the results.
You need to generate training and test data separately for each question. Plots of training data,
epoch wise convergence of mean square error and identification result for test data should be attached
for each problem.
You can use C/C++ or FORTRAN as the programming language. Nothing else will be entertained.
Students have to email the soft copy of their programs to the instructor on the day of submitting
the assignment.
Copying from a fellow student will be severely penalized. Depending on the severity, marks will be
deducted. If a student gets zero marks in both assignments because of copying, he/she will be given
an F grade.

Problems
1. Consider a MIMO plant which is governed by the following difference equation:

x
(k)
1

1 + x2 (k)
x1 (k + 1)
u1 (k)
2

+
=

x1 (k)x2 (k)
x2 (k + 1)
u2 (k)
2
1 + x2 (k)
The system is to be identified using the following neural model.

x1 (k + 1)
N [x1 (k), x2 (k)]
u1 (k)

=
+

x2 (k + 1)
N 2 [x1 (k), x2 (k)]
u2 (k)
where N 1 and N 2 represent two multilayered neural networks. Derive the weight update algorithms
for both using back-propagation of errors. Write a computer program to identify the system.
2. Consider the following state space model

x
(k)
1

1 + x2 (k)
1
+
sin(x
(k))
x1 (k + 1)
2
2

+
u(k)
=

x1 (k)x2 (k)
1.5 + sin(x1 (k))
x2 (k + 1)
1 + x22 (k)
You need to learn the above state space model using an RBF network. The inputs to the network
are x1 (k), x2 (k) and u(k) and the desired outputs of the network are x1 (k + 1), x2 (k + 1). Choose
a number of RBFN centers and learn the centers using K-mean clustering. Once learned, fix the
centers and update the weights using
Gradient descent algorithm
Recursive least squares algorithm
Compare the convergence times. Show the initial and final distribution of the radial centers in the
input space.

You might also like