robots.net blog for tim.holt

5 May 2004

Wed, 5 May 2004 16:21:58 GMT

Interesting meeting last night here at Oregon State University. Professor Belinda King, department chair in Mechanical Engineering called an "all engineering" meeting to explore interest in setting up an Autonomous Vehicles research cluster here. A cluster is not exactly a single department or research group, but a collaboratory cross of multiple departments.

Hopefully this first gathering of mechanical engineering, oceanographic, electrical, computer science and other disciplines can lead to something interesting here.

28 Apr 2004

Wed, 28 Apr 2004 23:43:13 GMT

Looking at pontoons for prototype vehicle. Have decided to just use larger diameter (8") PVC pipe to create pontoons, as this should provide a lot of strength, plus is easy to work with. With a bit of calculation I figure a 6' pontoon made from 8" diameter PVC will support about 130 lbs. Multiply that by two (two pontoons) and we're talking about 260 lbs buoyancy, which should be more than sufficient for the initial prototype boat.

A bit of web searching gave me this project powerpoint which shows a similar setup to how our prototype vehicle will be layed out.

26 Apr 2004

Mon, 26 Apr 2004 20:30:54 GMT

Starting to create the control software for the project. Implementation is in Perl, and is designed around a client/server (TCP/IP) architecture.

We'd like to have a manual control for the vehicle, for a user to steer it during learning phases and situations where the AI is unable to properly respond. An example might be bringing it into the dock. GPS will get the ASV close to the dock (within the < 15m error range of GPS), but can't deliver it right to someone ready to grab it.

Right now we're using the ActiveState Perl Win32::Multimedia::Joystick module. It is able to work with conventional joysticks such as one from Logitech as well as game controllers such as this one also from Logitech. I'm also implementing a mouse based controller where a user clicks and drags inside a target window to simulate joystick motion, for situations when a joystick is not available.

20 Apr 2004

Tue, 20 Apr 2004 23:57:17 GMT

Acquired a Vantec CDFR23 dual DC motor controller.

Next step is to acquire motor components and pontoon system. Pontoons will either be built (rigid foam with fiberglass skin and embedded Unistrut on top) or purchased (Kelly pontoons).

Image of current prototype design at http://www.nacse.org/~holtt/images/boatpic2.gif