Thursday, June 9, 2011

Society of Physics students presentation

We attended the Society of Physics Students (SPS) regional conference for zone 18 and brought Ferdinand with us. We presented a poster along with some other Mt. SAC students. Here is the .jpg of our poster.
The conference was fun. The CEO from the SETI institute gave a talk about their new radio telescope array, and detailed their summer internship program. Other meetings discussed methods for maintaining a healthy student club, and the was a physics magic show.

Wednesday, June 8, 2011

CORE comes back with Silver

After a grueling competition, our Robomagellan platform took second place. The motor controller burnt out on our second to last run, and we were unable to identify the problem before the competition finished. The gold medal went to UCLA's GLaDOS project, which has taken the gold at robogames for the past 3 years. It is unfortunate to lose, but to lose to a worthy opponent brings knowledge instead of shame.

Above: from left to right: Brian Walker, Andy Gabriel, Gabriel Chou.

Here is the platform in its moment of glory. It strikes me that most robots have names, because calling it the platform repeatedly is unpalatable. I suggest that Mt. SAC's entry be referred to as Ferdinand, in honor of the explorer for whom the contest is named.

Next year we will be coming back with more robust hardware, and more advanced navigation software, so GLaDOS might have its work cut out for it.

Monday, June 6, 2011

Video from Robogames 2011

Here is a video clip of the platform in action at Robogames 2011.


video
The system was mostly dead reckoned, taking compass readings from point to point, and using odometry to determine the distance traveled. This system generated consistent behavior, which we then tuned on successive trial. As a method for navigating a field, it has its limitations.

In order to improve on this system, a more advanced method for keeping track of position needs to be implemented. Our early attempts at tracking the robot graphically might work if the visual elements were removed and the data was handled in the dark. The graphical interface was excellent for debugging purposes, but the accumulation of lag in the robot's position may be due to the increased processing requirements.