Sunday, February 3, 2008

Peg Leg: Our 6.270 Robot!


MIT’s annual 6.270 Lego robotic competition (2007-2008) was a blast for our team. (We were team 29, Arjun, Amber, and Emily). Over 3.5 weeks during our Independent Activities Period (IAP), we built a robot using Legos and a variety of other electronics. Our final robot was named “Peg Leg” because of the peg that supports his front.

The goal of the competition was to carry balls to different goals on the course. Two robots competed at once to get a certain score (15-22). Each goal was either worth 4, 2, or 1 points.

For more information about 6.270 please visit the competition webpage (http://web.mit.edu/6.270/www/contestants/)

Available Equipment:

  • Bump sensor (senses touch) [Peg Leg used one to detect the wall for scoring]
  • Distance sensor (can tell how far away an object is, has a limited range of a few feet)
  • Gyroscope (detects change in angle, can be used to drive straight or turn accurately)
  • LEDs and photo-sensors (detects light so the robot can follow lines or detect changes in color))
  • IR sensors (detects heat, but can be used to follow lines and detect changes in color, generally more reliable than light sensors if a change in lighting occurs) [Peg Leg used this to line follow]
  • DC motors (to help move the robot) [All robots need these if they want to move! We had one for each of our two wheels]
  • Servos (similar to DC motors but they only rotate ninety degrees unless modified) [Peg Leg uses a servo to open up his ball mechanism]
  • Shaft encoders (to drive straight)

We rebuilt our robot many times because we were never satisfied with the stability of the legos and gear train. In the end, we ended up with a very sturdy, dense, and small robot.

We started out with a complex robot plan, but as we went along the plan changed and became simpler. Below are some of the things we tried:

We had a “Practice Robot” which was not stable, only used to practice building and to pass the first few assignments, we then rebuilt it a ton later

Peg Leg’s first gear train was a 75:1 ratio, but we later changed to 125:1 so the motors would have more torque

Peg Leg’s first line following was with 3 LEDs and light sensors on the front. We later changed to 5 IR sensors, 2 on the left/right side, and three in the middle. Peg Leg was able to react much more quickly with more sensors. We used IR sensors because they are supposedly more reliable and do not vary as much with small changes in lighting. Here's a video of our first line follow (3 light sensors):


In the video below, Arjun discusses our robot's state right before the mock competition (5 days before the real competition).

We had lots of problems getting the line following to work. The sensors must be at an angle towards the LEDs or IR to detect the beam coming off of the surface. We also noticed that different surfaces and colors changed the values a ton. The robot worked best on surfaces that were not shiny and that were black and white. All our hard work paid off though, because Peg Leg ended up line following like a pro!

We tried wall following with bump sensors, which worked okay, but we decided to stick with line following and we got rid of all the bump sensors except for the front one.

In the video below, Emily discusses how our robot is doing. This is a day or so before impounding our robot.


At the beginning of the competition, the robot is placed in a certain direction. He is placed on a square with a varying amount of white and black, so it is easy to use LED or IR sensors to determine his orientation. Once that is done, we had Peg Leg follow the line towards the other end of the board.

Getting Peg Leg to turn corners was not too difficult, but we had to really think what conditions we would need so he would turn correctly. We ended up having him stop when the left or right front sensor turned black. He then moved forward until the same sensors turned white. He then turned the direction of the end sensor that originally turned black until hitting the line again. When he hit the second corner, we told him to turn slightly until he hit the goal (we used a bump sensor in the front to detect this). We used a servo to move the gate holding the balls.

Here is a video of Peg Leg in his almost final state (still the first ball mechanism) line following and scoring!

Here's a picture of Peg Leg with our first ball mechanism!


I will try and find a video of the robot scoring with his final ball mechanism, but in the meantime, here is a video that shows our final ball mechanism in action. It lets out a specific number of balls at a time, instead of dropping them all in at once.



Here we took a picture of the clock the night before impounding. We worked late :-).















Overall, we had a lot of fun with 6.270, and we are proud of our robot.