Saturday, February 6, 2016

Robotics Challenge 3: Strategy

We have 6 weeks to prepare for the FIRST robotics competition. Our primary goals are to be able to traverse obstacles and shoot balls at tower windows to score. During week three, our drive-train still isn't done, so I analyzed the game rules. Every good gamer does this. I helped the kids piece together how the game is played, what is a penalty, and what earns points. Next, I designed a strategy map for building a team play book and communicating to allied teams during a competition (left). This shows the entire field, including towers, obstacles, and balls.

The simplest inferences were about firing positions. There are a number of positions on the field where other robots can't bump into you or block you. Therefore, this areas act as sniper nests. (right)

Moreover, the three tower goals have only about a forty degree range of access, Any attempt at a goal beyond this range will likely hit the goal support structure and bounce off. Still, the zones overlap (purple on map below). A sniper in this zone has choice of targets. As you can see, entrance two puts us firmly in this magic zone.

Many of my observations dealt with robot roles on the team (attacker, defender, or sapper) and how they can score/prevent the most points. I did an analysis of the minimum possible score for a win and the estimated time for each operation. Initially, our team designed for the low goal (2 points each) because aiming for the high goal (5 points) seemed too hard. Then I pointed out two key facts. First, a high goal shooter with a 25% accuracy will out score a low goal shooter with a 90% accuracy--due to the difference in points as well as the high amount of time needed to get a fresh ball rather than your own rebound. 

I built an excel spreadsheet that showed that our prototype wooden launcher without certain features could peg the high goal from the white line marked on the map above.  Once we cross the defenses, this line is only 7 inches from the front of our shooter arm! The official launcher will shoot faster. Indeed, low goal attempts had to be from closer due to the shallow angle of attack (4-5 degrees) required until you get within 5 feet. Encouraged, the team shifted strategy.
On a whim, I looked for ways to score even if being blocked by an enemy robot. The result was the bank-shot chart. While not feasible without a launcher speed I've dubbed "hammer of the gods" range, it would be a real crowd pleaser. 

My last challenge for the team while the robot was still being assembled was to teach computer vision to the kids so they the robot could recognize the goals on our camera and allow automated scoring,

No comments:

Post a Comment