The Engineering Thunderbots team is back from Graz, Austria after competing in the 2009 Robocup Competition. The world-renowned event attracted thousands of participants from over 40 different countries, as well as lots of talent.  The goal?  To compete in the small sized league, where they had autonomous robots (pictured above) playing soccer against robots created by other schools!

Since 1997, RoboCup has been the frontier of research in Robotics and Artificial Intelligence, providing many different competitions. For example, two of the main ones would be Robocup Rescue, where robots are given tasks that simulate real-life scenarios of search and rescue, as well as Robocup Soccer; the first competition in which Robocup was created upon, its vision is to “develop a team of completely autonomous, humanoid robots which can beat the human football world champions by 2050” (to see a video reel with highlights from Robocup 2009, click here)

This year, Thunderbots participated in the small size league competition, a contest which pits two teams of five robots against each other. The amount of work that goes into creating these robots, having them actually able to play Soccer autonomously, is no small task. It involves a high level of programming, electrical and mechanical work.

The electronics (the majority of which consists of the wireless communication, the actuators and the kicking system drivers) sit on a single main board which is connected directly to the microcontroller board. The control algorithm relies on the feedback loop created using the optical encoders, each having a resolution of 360 counts per rotation. Each encoder is connected to an up/down counter chip which records the wheel speed as measured by the encoders and provides it to the microcontroller when enabled. The DC motors are controlled using two dual motor drivers, with each motor driver providing up to 10 Amps at 14 volts through a Pulse Width Modulated (PWM) signal.

Programming these robots to properly play soccer was a long and involved process, requiring many different parts to achieve this lofty goal. Each robot is controlled by an external computer, loaded with AI created by each team. Data is collected using cameras positioned above the field of play, which is then interpreted by the AI. It then sends wireless signals to the robots to provide split second responses.

To test the AI, Thunderbots created a simulator which does a simple 2D physics simulation to approximate what the robots might do in reality, allowing them to test different strategies as well as the AI, without actually using physical robots. The visualizer module was a very important component to testing each robot. As Byron Knoll, the lead AI programmer, so aptly puts it, “The visualizer can be run either during a simulation or during a real game. It is useful during a real game because it allows us to monitor what the AI “thinks” is happening (but which isn’t always happening in reality)”.

The mechanical workings of each robot is very complex, and can vary from team to team; the Thunderbots robots are equipped with 4 omnidirectional wheels, positioned concentrically around the robot’s center of mass. This configuration allows for rotation of the robot while simultaneously translating. In a 2D (top view) perspective, it has 3 degrees of freedom.

To fulfill its other main functions such as passing, dribbling and kicking, the robots require a kicker and a dribbler mechanism; the dribbler mechanism consists of a soft and grippy roller powered by a 2 watt motor. When the robot comes into contact with the ball, the roller grips the ball and applies a backward spin to the ball as it slips on the felt. This causes the ball to constantly roll towards the robot, obtaining possession of the ball. The kicker mechanism involves a kicker head mounted on a solenoid, which is controlled via a kicker circuit and the microcontroller. Using PWM, they are able to control the strength of each kick.

The team had spent the weeks before the trip working day and night to make the final adjustments and preparations to have competition ready robots.  It was all worth it though for the team members as they were days away from carrying out the dream that Bahador Moosavi had when he revived Thunderbots 3 years ago (for an article featuring the creation of the team, click here).

In the very beginning, the team consisted of a small group of engineering physics students all from the same class, but over the years has grown to encompass engineers from almost all disciplines and from all years.

“It doesn’t matter what year you come from, or what department, the people we look for are the ones that tinker around with stuff at home, the ones really interested in this kind of stuff” says Johnathan Fraser, the control leader of the group.

Being their first venture into the small sized league, and having to play with teams that had more grad students than undergrad, it is surprising to note how well they did, even beating Robopet, a team from Brazil which had shown considerable promise during the competition. The team finished within the top 16, a definite acknowledgment of the extraordinary amount of time and effort the team gave to participate in this event.

“The best bit about Robocup though, is that, although it is a competition, it stresses learning, sharing and teamwork, rather than heavily guarded secrecy in order to win” says Amanda Li, the administrative captain for the team. “We made friends with many teams from across the world, shared ideas and traded advice”.

The Thunderbots team is currently planning on going to Robocup 2010, Singapore, next year’s competition. For more info on Thunderbots and how to join, check out their website at www.ubcrobocup.com.

This entry was posted on Friday, July 24th, 2009 at 5:27 am and is filed under Student team. Both comments and pings are currently closed.