GVG SLAM N200 Lab 1 Experiment 1

GVG N200 Lab 1 Experiment 1

   Experiment of August 27, 2002.
   Conducted by Deryck Morales (deryck@alumni.cmu.edu)

   This experiment was arranged in SBP lab 1 here at Carnegie Mellon.
   Most of the room is not engineered for the test: this is a real-world indoor experiment.
   The red chair/pair of tables serves as the central obstacle. Fortunately, the height of the sonar
   sensors on the N200 is that of the tables, though we will see that the surrounding desks are lower,
   such that the objects on the desks are detected instead, and create a noisy sonar environment.

   Here are some pictures of the room with the N200 and a few cardboard walls to enclose the space:

From corner: notice cardboard walls at bottom

   Our exploration program successfully drove the robot through the space and recorded the GVG map.
   Using the disp log display utility (included in the software package) we can view the result of
   the mapping run. On the right is the final graph structure. On the left is an animated image that
   is composed of 5-step intervals of the exploration as displayed by the disp utility.

   You can download the log files generated by this experiment here and on the downloads page.
   Four recorded clips of the exploration are available for download as well.  
   These are in .mov format (Quicktime) and are also listed on the downloads page:

   Clip #1  : The N200 tracing GVG edge from meetpoint 2 to meetpoint 4.
   Clip #2  : Homing in on meetpoint 5, exploring to boundary node 6.
   Clip #3  : Tracing GVG edge from meetpoint 5 to Meetpoint 1.
   Clip #4  : Edge tracing from meetpoint 1 and homing in on meetpoint 2.

   Finally, here is the terminal log display of the detailed graph information:

Go back to: H-SLAM Experiments Page | H-SLAM page