Descripton of Cosmo
This year, the Mavericks are participating in the Interstellar Accuracy Challenge, the Power Port Challenge, and the Hyperdrive Challenge.
Cosmo has a swerve drivetrain, which we found performed better than the mecanum we designed because it was faster and more accurate. We knew we wouldn't need a climber or an intake, and we only needed to hold three balls at a time, so it has a small frame perimeter which helps to make it more maneuverable in the driving challenges. We use 2910’s Mark-2 swerve modules, which each have 2 neos driving them. To code the swerve, we adapted WPILib’s example swerve project and swapped it for our tank drive subsystem code. Keeping the same interface allows us to easily change from mecanum to tank to swerve whenever we need to.
Our flywheel shooter subsystem (similar to our shooter from last year) uses our previous year's tech with an angled sheet of polycarbonate feeding power cells into a set of wheels, which then bring the cells up to our flywheels to be launched. We considered adding the ability to angle the back of the shooter, but this ended up being redundant and overcomplicated for the task at hand. Our shooter has a 4 inch solid urethane roller flywheel, directly driven by 2 falcons at a 1:1 ratio. The flywheel is heavy so that it doesn’t lose too much speed after shooting a ball, allowing us to shoot faster. This is desirable for the power port challenge. On the shooter, we use a PIF controller to achieve very precise accuracy. We decided to use a PIF controller instead of a PID, to account for the slowing down of the flywheel that we knew would occur between balls being fired. The shooter is mounted on a turret, but we decided to lock the turret for two reasons. First, shooting on the fly was not allowed in these skill competitions, and second, we found that the swerve drive gave us the flexibility to shoot from many positions without needing a turret.
To calculate how far we were from the powerport, we used the limelight. We used the distance calculated using limelight data and a unique function and compared it to the actual distance the limelight was from the target, using a tape measure. This contextualized the data for us, and also allowed us to understand, for example, how far an autonomous program would need to drive backwards to get to the ball loading port between shooting cycles.
Do students build the robots by themselves? No. We have skilled community mentors who introduce robotics engineering to students. Students work side by side with mentors to learn all aspects of work on the team: robotics, machining, software, business, communications, etc. The mentors serve as role models and walk students through projects, slowly allowing them to take on more and more responsibility as they show competence and understanding. As well, this year our team had two specific training sessions of four weekends each (open to all Kingston teams) of Computer Aided Drawing and Robotics Instrumentation.
Cosmo has a swerve drivetrain, which we found performed better than the mecanum we designed because it was faster and more accurate. We knew we wouldn't need a climber or an intake, and we only needed to hold three balls at a time, so it has a small frame perimeter which helps to make it more maneuverable in the driving challenges. We use 2910’s Mark-2 swerve modules, which each have 2 neos driving them. To code the swerve, we adapted WPILib’s example swerve project and swapped it for our tank drive subsystem code. Keeping the same interface allows us to easily change from mecanum to tank to swerve whenever we need to.
Our flywheel shooter subsystem (similar to our shooter from last year) uses our previous year's tech with an angled sheet of polycarbonate feeding power cells into a set of wheels, which then bring the cells up to our flywheels to be launched. We considered adding the ability to angle the back of the shooter, but this ended up being redundant and overcomplicated for the task at hand. Our shooter has a 4 inch solid urethane roller flywheel, directly driven by 2 falcons at a 1:1 ratio. The flywheel is heavy so that it doesn’t lose too much speed after shooting a ball, allowing us to shoot faster. This is desirable for the power port challenge. On the shooter, we use a PIF controller to achieve very precise accuracy. We decided to use a PIF controller instead of a PID, to account for the slowing down of the flywheel that we knew would occur between balls being fired. The shooter is mounted on a turret, but we decided to lock the turret for two reasons. First, shooting on the fly was not allowed in these skill competitions, and second, we found that the swerve drive gave us the flexibility to shoot from many positions without needing a turret.
To calculate how far we were from the powerport, we used the limelight. We used the distance calculated using limelight data and a unique function and compared it to the actual distance the limelight was from the target, using a tape measure. This contextualized the data for us, and also allowed us to understand, for example, how far an autonomous program would need to drive backwards to get to the ball loading port between shooting cycles.
Do students build the robots by themselves? No. We have skilled community mentors who introduce robotics engineering to students. Students work side by side with mentors to learn all aspects of work on the team: robotics, machining, software, business, communications, etc. The mentors serve as role models and walk students through projects, slowly allowing them to take on more and more responsibility as they show competence and understanding. As well, this year our team had two specific training sessions of four weekends each (open to all Kingston teams) of Computer Aided Drawing and Robotics Instrumentation.