Robotic Localization Challenge

After having been hard working on this Lego robot for the past 2 months, we finally get to this last challenge.

Again, this is based on Monte Carlo Localization and Place Recognition. The implementation is based on RobotC. There are several challenges:

• The robot will be placed randomly by the examiner in one of the five marked way-point position. The orientation of the robot is also randomly picked.
• The robot has to work out where it is placed and then navigate as quickly as possible to each of the five points and navigates back to the start point.
• Design of the robot is on our own choice. The choice of methods will not be judged, but performance of the robot counts.
• Kits/Sensors allowed is the what it comes with the standard education pack. Thus, only one sonar sensor, three motors and two light, two touch sensors are allowed.

The video below was how our robot performed for the first try. The robot first correctly figured out the current way-point position and its direction using sonar sensor, then started its navigation to other points.

Notice that each point has a precision measure, if the centre of the robot moves preciously to the centre of the points (marked red), we get 3 points. Otherwise, lower precision will result 2, 1 or 0 marks.