Thursday, June 22, 2006

Gamedev 2006


Last night I partied with the Case Gamedev 2006 team. I found out that in an alternate world, I am a "death protein." Go figure... anyway, we went to Aladdins for dinner and partied like the old days at Scott's apartment afterwards. We took a nice group photo too, hopefully I'll get that uploaded once I get my hands on it.

Thursday, June 15, 2006

My MacBook is Sweet

I've been using my MacBook for a couple of weeks so far... here are my feelings on it. Note: I am pretty critical in this review because I am focusing on the negative aspects of the machine. Overall it is an exceptional notebook, and most of the problems have to do with running Windows on it. It is a sad Mac any day of the week when it has to run Windows... anyway:

1. Dual-Boot is very useful.
I was booted into WinXP for most of the IGVC competition (see below post) for all the robot vision work. The code was written in C++ using Visual Studio 2003 with support utilities written in MATLAB. The MacBook handled some harsh development conditions in stride, even in the WinXP environment. However there are some problems with dual-boot...

2. WinXP is not quite right on the MacBook...
Where should I start? First, Apple hasn't written a trackpad driver for the MacBook under WinXP yet. So, the trackpad is jerky and unresponsive. Wonderful! I'm sure this will get fixed in time, but for now, it is pretty annoying.

Also, the network connection seems to be flaky under Windows. When connecting to a network gateway via one of its wired connections, sometimes after waking from stand-by in Windows, I couldn't reconnect to any of the machines on the network until I not only rebooted, but rebooted into MacOS X and then back to Windows! Argh...

To get the keyboard and mouse working as expected, I highly recommend Input Remapper to get the brightness, eject, and other keys working, as well as a right-click option.

Otherwise, Windows seems to be pretty happy. Of course, the built-in camera doesn't work, but again, that's just a drivers issue.

3. Glossy screen is OK
I had my doubts when I first heard that the MacBook had a glossy (read: reflective) screen as opposed to the matte-type LCDs used on all other models. The screen is pretty nice, actually. There are only a few circumstances where I see distracting reflections in the screen, and those can usually be fixed by adjusting my seating position or changing the angle of the screen. I can also see the screen pretty well outside (probably better than on a matte screen) so that is cool too.

Overall I'm happy with my purchase and would recommend it to anyone who wants a lightweight, powerful, versitile and inexpensive laptop. Of course, if you are getting one, make sure you try out MacSaber. Thank me later.

Intellegent Ground Vehicle Competition


Last weekend, I went to Michigan with the rest of the IGVC Team from Case with our robot, Roberto. We had high hopes of bringing it all together after a month's worth of long days (and nights) trying to get all the robot's systems integrated for our first showing at the competition. We made it so far as to qualify our robot for competition, and we also won $500 for implementing their wireless robot control protocol over UDP. Unfortunately, the USB cable connected to our hub dangled into the chassis of the robot and touched the +24V terminal of the lead-acid battery, which toasted our NI DAQ's, serial-to-USB adapters, digital compass, and the hub itself. We stayed up all night rebuilding, got it to the point where we completed the wireless command challenge, but a subsequent electrical failure killed our replacement DAQ and ended our bid to compete.

Going into next year, we have a much better idea of what challenges we must overcome, as well as a lot of the work already being done. As of Sunday morning, all sensor systems were functional and data was being integrated and logged - we were just nailing down final bugs... with a year to rebuild, we will definitely have a strong showing next year.

I worked with Scott McMichael on the vision system for the robot. I was primarially responsible for integrating vision with the rest of the robot, making sure the vision data was being used properly for mapping and planning, and providing as much functionality as possible given our vision data. Also, since Scott couldn't make it to the competition, I was the resident vision "expert" in case anything went wrong (of course things went wrong!). Vision was up and running well for the competition, but we never got to run the full challenge due to the electrical problems we encountered. Hopefully this summer, we will be able to test robot vision more thoroughly.