The robot should be able to: (i) detect the position of the target, (ii) formulate a plan for hitting the ball to the target, and (iii) have the ability to hit the ball to the target. This should work by having a user interface with a graphical program. When a target is visible to the robot, the interface should allow you to send a “target” command. The robot will then calculate the batting plan and move to the initial position for that plan. When at the initial position, the interface should then allow the user to send a “fire” command. The robot will then complete the batting plan and hit the ball to the target.

 

In (ii) we planned to simulate our mathematical model of the double pendulum to optimize the parameters for batting the ball to the target height. However we were unable to account for all the dynamics of our system. Our theoretical results differed enough from our experimental values that we went with a different approach. Instead of simulating the dynamics at runtime and finding the batting parameters, we precomputed parameters that hit the ball to varying heights. We then tested those parameters and put their experimental values of target heights into a lookup table. With this, during planning stage we simply find the experiment that hit the ball closest to the desired target height and use its parameters for the batting plan.