[apalrd]

We have a few safety features, both in software and hardware. Most of these aren't unique to this year.

Software:

-We have incrementally (over the past few years) adding 'ok to actuate' (OKTA) signals in addition to the command signals for actuators. This year we've added them for all solenoids. Part of the output layer (in code) takes a boolean position for the solenoid and a boolean OKTA for the solenoid, and then the output layer takes care of dual vs single coil, safe states during OKTA, and polarity inversion.

-We also added OKTA signals for several control loops this year. A negative setpoint (since all of our setpoints are in the positive region) will trigger the control loop to hold the integrator in reset and hold the output to 0.

-We have a boolean signal in code called 'reinit' which is triggered every time we exit disabled mode, and it forces all the code into a 'safe' state. All of the state machines to go a 'default' or 'manual' state which sets all of the actuators to 'not OKTA'. The user must command another state for the actuators to go OKTA (usually all of them commanded by that system go OKTA at once, which is acceptable to us).
--For this year, we used several dual-coil solenoids vs single-coil solenoids for safety reasons as we did not want actuators moving when we disabled the robot. We also used a single-coil solenoid for similar reasons.

Hardware:

-We are planning on building a metal bar with features which latches between parts of the robot, preventing it from firing. This would be removed when the robot is set on the field

-We will release the stored energy between competitions and overnight.