University of Notre Dame
Aerospace and Mechanical Engineering
ME 469: Introduction to Robotics
Homework 6
B. Goodwine
Fall, 1998 |
Issued: November 13, 1998
Due: November 20, 1998 |
Do the following two problems as a group.
Consider the following two images. The purpose of this homework
is to develop software that can distinguish blurry photographs
of objects that may appear similar. You can write the programs
using any programming language that you want.
- (100 points)
Write a program to determine the object properties for each
airplane:
- the area;
- the coordinates of the centroid of the object;
- the length of the perimeter;
- a measure of the objects compactness;
- the mean distance from the centroid to the shape
boundary and its standard deviation.
- (100 points)
Write a computer program to implement an artificial neural
network to approximate the following function.
f(x1,x2,x3) = 1 if any two of
the three input variables are 1 and the other is zero and
f(x1,x2,x3) = 0 otherwise.
Assume that the inputs can only be 1 or zero. Using the
back propagation method for training the network is
recommended, but not required.
Networks to make other predictions are acceptable with
prior permission from the instructor. Possibilities include
predicting the outcome of sporting events, predicting the
weather, etc. A program to automatically download a summary
of professional football scores can be found here. You may have to ask the
instructor for more details regarding this program.
Last updated: November 9, 1998.
B. Goodwine (jgoodwin@nd.edu)