Robot Sumo 2006
Robot Sumo basically involves creating a robot capable of staying on a white table encompassed by a black ring (usually via the use of a light sensor), and pushing one's opponents off the table. Although it is possible to create a fairly competitive robot with a very simple program, the task can be made more complex by including rules such as no repetitive behavior. Although not too much math is involved, it is excellent for building reasoning skills (for example, to keep one's robot from falling off the table, it should turn when it detects black). This movie is of the final competition from the first year I worked with ASRA in 2006.

Robot Sumo 2007
Here is another sumo match from 2007 using the new NXT system.

Maze
One task we've given students is to build and program a robot to solve a maze. While there is a small degree of design required, the bulk of the challenge involves basic programming. It is a good task for helping students to develop basic programming skills. In the future I'd like to build a covered maze so that students do not know what it looks like. This would force them to write a different type of program-one in which a robot must make intelligent decisions.

Climbing Robots
Another cool task is to have students build robots capable of climbing steep inclines. Prior to the NXT's powerful motors, such a challenge would not have been worthwhile. However, the robots these students made not only climbed a steep wall, but followed a line in the process. Hence, the robot had to be well built and have a semi-decent program.

Racing Robots
A fun activity to help break up the monotony of an 8 hour day of building and programming robots is to have races. Almost no programming is involved, but students are forced to redesign the robots from the ground up for speed. Most eventually take advantage of some sort of gearing as well as all three motors. Initially, a straight shot race is probably best so that students can focus on fine tuning the gearing. However, a second race which incorporates a turn requires a program that does more than simply direct the robot to drive forward.

Harvesters
This was a fun activity in which robots were built to collect small rubber balls. I was very impressed by the creativity of some of the students at this point.

Tug of War
This is another excellent task that does not require a whole lot of programming (but students have lost by programming their robots to drive the wrong direction). An understanding of some basic physics and its application to the design of one's robot makes a significant difference in this activity. Early on a few students discovered the importance of a lever arm, and after working out the kinks, soundly defeated their rivals repeatedly.

Final Project
For the final project, students were put into teams of 3 and basically told to transfer as many of the balls on their side of the table to the other side of the table. It was a challenge to get some of the kids to work together, but in the end every student's robot contributed to their team's goal. It was really cool to see robots coordinate their activities to accomplish a semi-complicated objective.