This page is a record of the experimental run of the vision scout mapping four landmarks as it drives across the floor in a prearranged straight line.
Experiment of October 25, 2002.
This experiment was arranged in SBP lab 2 here at Carnegie Mellon.
This is the first experiment showing the vision scout conducting mapping with vision system using a Kalman filter.
To carry out the experiment, the robot was commanded to travel 300 inches forward at a speed of 4 inches per second. As the Scout drove across the room, the camera took images which were processed for landmarks based on intensity of the light landmarks as in the bearing-range experiment. The results of the experiment can be seen in the animated gif below. The state vector and covariance matrix were saved as the Kalman filter ran online. The dotted line shows the path of the robot with the start as the green circle and end as a red circle. Each frame of the gif shows a snapshot of the state: the robot's position is shown by a black asterisk, the landmark locations by blue plusses, and the covariance of each landmark by a magenta 2-sigma covariance ellipse.
The covariance can be seen to quickly decrease, only problem is it's probably too fast. The next step will need to be some gain adjustment to more accurately represent the landmarks and errors associated with them.
Go back to: H-SLAM Experiments Page | H-SLAM page
Jump to: People | Robots | Research | Events | Projects | Education | Papers | Software | Links
© Copyright 2002 Sensor Based Planning Lab, Carnegie Mellon University. All Rights Reserved.
 |
 |
 |