Milestone 2 Reached! Now You Can Watch It

We have put together a video of our Milestone 2 run, which we completed this past Monday, June 1, 2009. This particular run had our PR2 alpha robot navigate through eight doors, and plug its power cord into nine outlets. In this video, you can see the various challenges our robot faced, such as a crowded office environment and the abrupt appearance of a human obstacle. We nearly sabotaged the run early on. Folks around the office were eager to track the progress of the robot, so many people ran their own monitoring programs on the PR2. This caused an increase in CPU load, starving the navigation software. Nonetheless, the robot was able to continue with the run, albeit more slowly and cautiously.

As you can see from the video, perhaps the element most critical to our success in the milestone was PR2's robust failure handling, which was managed by TREX. The run was not perfect, but PR2's ability to detect failure and try again allowed the robot to persist until it successfully completed its task. During the run, two attempts at picking up its plug failed, but the TREX executive repositioned the robot to try again, with success. Similarly, the robot had trouble detecting a partially open door, but with a couple of retries, it was able to detect and open the door.

As the run took nearly one hour, we have sped up some of the footage. For Milestone 2, our tester designated ten different outlets that the robot was required to plug into, although only nine of the selected outlets were physically reachable. The robot was expected to successfully detect that one of its ten goal outlets was behind a locked door, abort that effort, and move on to its next outlet. The rest of the doors were placed in various states of closed, partially opened, and fully open.

Comments

I applaud and commend you all

I applaud and commend you all on your great work and progress! I'm not sure if you guys have thought about this, however I guess it wouldn't hurt to throw out an idea or two; but I was wondering if adding some sort of secondary camera(s) in and/or around the hand, so it has greater depth and sensing, when it grasps items or things, such as the door handle, etc. And maybe having the power plug come out of (or near) the hand as well, might make it easier and more convenient for the robot to plug it in? Best wishes and good luck with your great endeavors! :]

Thank you for your comments.

Thank you for your comments. Regarding the secondary cameras, the final design of the robot does feature a wide-angle camera in each forearm. They weren't ready at the time that we were working on Milestone 2, but they are in our current prototypes that we have in the lab. Thankfully our researchers were able to make do with what they were given and performed admirably.

Video

I can't watch the video! It says "This video is not available in your country due to copyright restrictions."...

YouTube mislabelled the video

YouTube mislabelled the video as containing copyrighted music (there is no music track). We are currently disputing this with them.

a couple of minor suggestions

I would like first to express my utmost excitement about your progress. Finally, an open source organization is taking seriously the challenge of creating an intelligent autonomous machine. I've got just a couple of comments with regards to enhancing the experience of the viewers of the site. First of all, would be nice if you could post also a very brief description regarding how the robot achieves the different tasks, e.g. how it detects the plug, the socket, the door handle; how it knows where to find the sockets, etc. For example, I noticed that there was a checker board pattern printed on the top side of the plug, is that how the robot detects the plug. Also, I noticed similar checker board patterns stuck to walls around the building, is that how the robot knows its position. Me being a robot enthusiast, I am very excited about how the stuff works. Would be very interesting also to be able to see the comments other people have posted. Thanks and keep up the excellent work, robot enthusiast

It's all open source

I know, it'd be a lot (a week or so) work to get a grasp of what's going on, but it's all open source. This is the procedure: http://pr.willowgarage.com/wiki/Milestone2/Results_2009-05-29_Integrated_(Trial_Procedure) You'll get all the code and config files. Unfortunately, svn doesn't currently work for the robot itself. Once I got a grasp of ros tools, it was not too painful to find the code that does the actual work. - Alex

Thank you for the suggestions

We will be posting better descriptions of the different aspects of the milestone, but it will take some time to pull that content together (and give people time to catch up on sleep).

We do use the checkboard on top of the plug as part of the detection process. The checkboards that you see attached to walls are not used -- they are from a different experiment. The robot uses AMCL (Adaptive Monte Carlo Localization) with input from our two laser range finders (one tilting) and odometry.

 - Ken

Papers on Milestone 2 are planned for ICRA

We'll be submitting several system's papers giving the technical details behind Milestone 2 to ICRA 2010 http://icra2010.grasp.upenn.edu/ - Gary Bradski