Research
Using a Proportional Controller(Winter 2002)
Experiments Using a Proportional Controller(Spring 2002)
The Coupling Algorithm(Summer 2002)
Coupling Algorithm Simulation(Fall - Winter 2002)
Coupling Algorithm Experimental Results(Spring 2003)
The FaceTrack with a Proportional Controller(Summer 2003)
SHORT DESCRIPTION: To gain confidence, a simple proportional controller was first implemented using FaceTrack and some experiments were done. This
program was used to make a comparison with the feedforward controller.
Modeling the PTU(Summer 2003)
SHORT DESCRIPTION: This page contains the block diagram and the pan transfer function of the PTU-FaceTrack (by saying PTU-FaceTrack I mean
the PTU with FaceTrack program running). The values of the Ki, Kp and Kd are also offered (those that were established
by TRUST Automation).
The Feedforward Controller(Summer 2003)
SHORT DESCRIPTION: In this page, the block diagram for Feedforward controller is presented by comparison with a Feedback block diagram.
To implement it I was using the
alpha-beta-gamma filter for target's
velocity prediction in the image plane.
Experiments with Feedforward Controller(Fall 2003)
SHORT DESCRIPTION: In this page, I present some videos from some experiments (tracking my toytruck, tracking a rotating target and
the tracking the Cye robot). In the first part, only the PTU was moving (no booming). Some plots are also shown as a comparison between the feedforward with proportional
control as feedback compensation and feedforward with PID control as feedback compensation. It can be seen that the
difference is not that big.
Tracking a human with Feedforward Controller(November 2003)
For the Cye robot experiment, the boom was moving as well as the target. Three videos are showing what the boom-camera
is seeing, what can be seen on the PC monitor and a view from another camera.
SHORT DESCRIPTION: The feedforward controller was implemented for both, pan and tilt. As in case of proportional controller, an experiment
of tracking a human being was performed. Three videos are showing what the boom-camera is seeing, what can be seen on
the PC monitor and a view from another camera.
A Comparison between Feedforward and
Proportional(November 2003)
An experiment was set up to assess the performance of feedforward controller versus the proprtional one. A Mitsubishi
robotic arm was programmed to sinusoidaly move the target. The camera was tracking the target while booming in the
same way. Some data were recorded.
The Maximum Camera-Target Relative Velocity
(January-February 2004)
To determine the maximul camera-target relative velocity that cam be tracked usign condensation algorithm
an experiment was set up. The target was mounted on a turntable. The program was launched and the maximum
velocity that was trackable was recorded for two zoom constant values.
A Model For the Boom(February-April 2004)
I am currently working to develop a model for my boom. I am using Lagrange method to determine the equation
of motion. For simulation, I am using Mathematica, matlab and Simulink.
What is New(Summer 2004)