Matthew Downing (me):
– Andrew began looking into the Turtlebot block demo.
– Nathaniel is still mapping the lengths of the linear actuators in the upper arms so they can be represented as AX-12 servos.
– Alan has completed the URDF for the head of PR2 Lite. Here is an image of the model with the head.
We used models of the Kinect from the Turtlebot. Other designs in the head were based off of the images created from the 3D models in Autodesk Inventor. These images are talked about and linked to in the second half of the linked blog post.
I got quite a bit of help from him, but I’m learning so it’s all right.
Here is a link to the game, called Catch Bigfoot which is hosted on my brother’s website.
Use WASD to move.
Matthew Downing (me):
– I am soldering up a new daisy chain for the wheels, with Frank teaching me how to solder and helping me hold the wires together.
– I cut out three wooden cubes for basic testing in the future for the robot to move around.
– I’m also writing this blog.
– Andrew is doing some documentation for code involved in the multi-calibration of Kinects.
– Nathaniel is trying to map the lengths of the linear actuators in the upper arms so they can be represented as AX-12 servos.
– He has made the upper arms mapped to angles of depression.
– Robert spent most of his time wasting the batteries on my laser pointer.
– Robert is modifying simavr to emulate the AVR 485 hardware. This is used for communication between the different microprocessors on the robot. He is mainly debugging the simulator.
– Alan is trying to fix the URDF to display the robot and work on the head.
– Frank is teaching me how to solder.
After having to tweak a bit of the arm navigation code in ROS electric, the upper arm now works. The problem with the upper arm was that it is made in a parallelogram shape, so that the elbow portion is always horizontal, or parallel to the ground. In URDF, objects cannot loop back around and reattach like in a parallelogram, so the elbow’s angle had to be changed to be the opposite of that of the shoulder’s.
The next update for our URDF for PR2 Lite is that we will start working on making the head in our model, including the Kinect and Hokoyo. I made images based off of my Autodesk Inventor 2012 files to help aid in the process (labeling dimensions and where objects attach).
Here is a link to those images.
On May 19 & 20, 2012 (Saturday & Sunday), we brought PR2 Lite to its first Maker Faire. Luckily, there were no problems in transport like those that happened at RoboGames 2012, mainly because of our new way of loading the robot in and out of the car.
Maker Faire was great, and most of the exhibits were innovative and creative. PR2 Lite got publicity, but unfortunately the teleop demo (PR2 Lite would mirror someone’s arm movements) was getting confused by people walking behind the person it was meant to mimic, even though the demo was working at the time.
PR2 Lite, among other projects at Maker Faire.
We have now attached the neato lidar to neato motherboard on the shelf. We replaced the orignal motherboard’s instantaneous “ON” switch with an LED so that the lidar will turn on whenever the ROS node is started. We kept the whole motherboard for ease of development. We may at some time expand the functionality to take advantage of Neato’s sonar and edge detection sensors. Still, the batteries must be attached for the motherboard to boot, even though we have the AC recharger permanently attached to the motherboard.
Below is a screenshot of the portion of the lidar in our kitchen and its corresponding picture, as well as the two superimposed. To see the full sized image, click on the image.
Screenshot of what the Neato is detecting.
Corresponding photo of what the Neato is detecting
Overlaid Neato image and photo.
As of June 9, 2012, we are trying to get constraint-aware arm navigation in ROS Electric to work with PR2 Lite’s upper arms. PR2 Lite’s upper arms have joints configured similar to a parallelogram. This configuration allows the lower arms’ bases to always be flat and movable by lower-end smart servos like the AX12. Unfortunately, ROS can’t handle URDFs with closed loops like a parallelogram, so we’re in the process of modifying the arm navigation code to handle PR2 Lite’s upper arm. Here’s a cropped screenshot of PR2 Lite in the ROS planning components visualizer.
Screenshot of PR2 Lite in the ROS planning components visualizer running in Linux. Click for a larger image.