Three different teams worked to design, build, and test three different ways to launch the ball. They built a catapult, a linear launcher, and a flywheel shooter.
Catapult Launcher
A catapult was designed and built today using a system similar to other catapults we've seen for this year and 2008. This quick prototype was not strong enough to launch balls far enough and the team needs to improve the design tomorrow perhaps by making it more stable and utilizing another tension device such as surgical tubing.
Linear "Trident" Launcher
Flywheel Shooter
The flywheel shooter was completed today, and students and mentors began testing with it. Of the three shooter prototypes, the flywheel had the best shots and most consistency. Additionally, the team could test the effects of ball pressure on the consistency and accuracy of shots for the first time in the season. By the end of the night the flywheel shooter made about 1/3 to 1/2 of its shots.
Please enable Javascript to watch this video
Programming
The programmers met today and decided on what they wanted to work on for the next few weeks.
It was decided that we would like to spend time building debugging tools for the robot. The main thing that needs to be built is a way to live graph motor and sensor states, much like the PID tuner we had in 2012.
The programmers also decided to rework how autonomous mode will be executed so that control flow can be used in autonomous rather than just a straight sequence of events.
Tom and Stephen started a new Java project for the 2014 robot and got basic CheesyDrive working on the 2013 practice drivebase. This code will use a new subsystem architecture and will use very little code provided by WPI.
Field Assembly
The team completed the construction and mounting of one full set of high goals. The goals have a ramp designed to return the balls back onto the field so that they do not fall onto drivers. We will use these wooden goals while we wait for the official set to come in from FIRST.
Drive Base Design
Today, the team began modeling parts for the gearboxes and chassis. We plan to design a drive base similar to last year but without a PTO on the gearbox. Tonight, we designed the gearbox output shafts, intermediate shafts and wheel shafts which will be sent out to sponsor Modern Machine Company tomorrow.
Action Items
Continue gearbox design
Continue testing and improving flywheel prototype.
Continue testing puncher prototype, try to increase power.
Rebuild catapult prototype, must be stronger and more stable.
Programmers continue improving drive code and autonomous tuning with Tom.
Improve catcher prototype with Dan.
Possibly begin machining parts for the drive base.
Today a group of students attended the 6:30 to 9:30 local San Jose kickoff at Morris Dailey Auditorium. A few others watched the NASA Live Stream from home. Subsequently, at 10, the team met for the team's kickoff event and discussed the plan for the coming six-week build season. The team then broke up into groups and discussed strategies in terms of necessary robot qualities, priority tasks, and possible scores. The parents and mentors simultaneously discussed logistics and the build plan in more detail until the student leaders returned to present their ideas to the entire group.
Game Analysis
After lunch, about 70 students congregated in the Bellarmine Robotics Lab to examine and better understand the team rules. In particular, they discussed the definitions of assist and possession and clarified the rules for such actions and others on the playing field. For the first two and a half hours, the group went through the relevant sections of the game manual and encouraged group discussion to find possible holes and list them to check with FIRST later. The group noticed some deviations in rules from previous years: the field is two feet narrower at 25 feet, and the match is 30 seconds longer. The autonomous period is five seconds shorter at 10 seconds, and the tele-operated period is 35 seconds longer at 2:20. Andrew Torrance noted the possible issues with battery life with the 25% longer match time. In addition to discussing rules, the team also formulated questions concerning the regulations to ask FIRST:
Can teams do assists in autonomous?
Can teams hold balls in autonomous to do an assist later?
Can a team throw over the truss again to try and get a catch, say, if an alliance earns a successful truss score, but does not successfully catch. Can the alliance try for a second chance? In other words, while the group recognizes that only one truss score counts per cycle, do the truss score and the associated catch score have to be earned on the first attempt?
Do teams get points (ex. throwing and catching in the end game) immediately, or do the teams only get them once the cycle completes? This is related to the question of assisting in autonomous. Possible answer: Maybe not because autonomous shots are not cycles.
Teams need clarification on which goalie zone to which the initial position regulations refer. In which goalie zone do teams start if they do not want to start in the white zone? Can human players help in the middle of assists?
Can human players catch truss scores to make them legal?
Teams need a better definition of possession for robots in general. For example, for a trough-like robot, is holding the ball as it rolls considered possession? What time period differentiates harmless "deflection" from "possession" of the opponent game object (which earns a severe technical foul penalty)
After a short break, the team's new mentor Jared set up a Google Spreadsheet that evaluated the effectiveness of certain "assist" strategies, with combinations of 1-3 assists, a truss "hurdle," and the truss "catch." He took student input for the point values corresponding to each action, the expected time taken for each action, and the probability of the action being carried out successfuly. He then calculated out a rough model of the weighted point values expected from strategies in the tele-operated mode of the match. The group was then able to analyze the efficiency of certain strategies. In the process of discussing strategies, the team also created two other strategy charts and lists (pictured).
Further Analysis and Prototyping
After dinner, approximately 20 students worked with the mentors to list some possible designs and to plan out the next week's prototyping schedule. The group prioritized the prototyping of possible modules, specifically shooter designs and catching mechanisms, with design input from students and advice from our experienced mentors. They also discussed what field elements were most necessary for prototyping for the next few weeks and listed them as well. Finally, they discussed the mathematics and physics of shooting a 25-inch, compressible, bulky, and irregular projectile into a 3-foot goal. For example, mentors noted that we might as well design a robot to shoot from farther away from the goal, since the shallower angle of entry would give greater allowance for error. Mentor Paul Ventimiglia also determined that giving the large exercise ball some backspin helped in precision, though it also required more energy to apply both linear force and rotational torque.
Tonight, we mainly focused on the building a preliminary goal with which to prototype before the field arrives and on prototyping and testing the effectiveness of a flywheel shooter and a slingshot shooter. Paul worked primarily with Godwin Vincent, James Holden, Jeremy Tanlimco, Vidur Maheshwari, Miggy Francisco, and Joncarlo Avila to design, sketch, and begin building a prototype for testing the effectiveness of a flywheel on such a large game object. They determined the best wheels with which to work based on their ability to grip the surface of the unusual game object. The prototype design was finalized and they plan to use four large wheels on the same axle hooked up to curved plywood "rims" to optimize the compression on the ball and somewhat minimize the size of the already bulky prototype. EJ Sabathia and Noah Marcel worked on the slingshot design, taping a 2013 Ultimate Ascent Frisbee to surgical tubing and hooking up the system to an alliance wall. The ball consistently bounced about 1/3 of the way down the court, given about 4 or 5 feet of backward stretch.
Lab closing time: 11:00
Action Items
Translation: Future Work Opportunities for Members
Help Paul with the Flywheel Shooter design
Help EJ with Slingshot Shooter design
Help Nick Mercadante machine metal for prototypes
Ask mentors to begin and lead prototyping projects for other designs
