Motor sim control loops

[catr1xLiu]

Hardware Control Loops Simulation for MapleMotorSim

[oh-yes-0-fps]

Do the PID controllers themselves have to be in the config. I feel like that's greatly expanding the scope of what a config is.

[catr1xLiu]

technically ur right, but it saves maplemotorsim from beeing like 600lines because of all the "with" methods

so, yeah, I guess it's no harm

[oh-yes-0-fps]

Does this have to be in wpilib units? it may make it easier moving all stuff not exposed to the user back to doubles.

[catr1xLiu]

I also thought that the wpilib units syntax isn't great for computation and therefore shouldn't be used for internal variables
but so far it's been working better than what I expect, I guess I'll give it a try in maple motor sim

[oh-yes-0-fps]

The control request taking the config and not the motor itself greatly limits what it can do, I could see the argument for limiting miss-use by restricting what they have access to but the motors don't really have any scenario breaking methods.

[catr1xLiu]

ur right on that

the only concern I have is future contributors will directly call to MapleMotorSim.requestOutput(new VoltageOut(...)) in the updateSignal() method to pass the applied volts

I guess I should warn people, in the javadocs of ControlRequest inerface, to not do that?

[oh-yes-0-fps]

If we intend for users to use the motor sims themselves I believe it would be beneficial if we handle the polling, higher hz running on the PID loops makes them respond closer to real life aswell.

[catr1xLiu]

NP, I'll have all maple motor sims updated in simulated arena subtick

[catr1xLiu]

hold on, I thought they should be updated in SimRobot, where the current drawn are also sent to the battery simulation

[oh-yes-0-fps]

I still stand by the fact the motor sim shouldn't do FeedForward for you, it could cause the motor to behave differently than a real motor without a good explanation why if the user didn't configure feed forward on the real motor. Adding this code to the javadoc of withFeedForward could be a good alternative to show the user what they could do easily.

Another note is that the MOI, gearing and motor are enough to figure out kA aswell :)

[catr1xLiu]

NP! I'll move this code to the javadocs of SimMotorConfigs class instead.

[catr1xLiu]

Not good at math but I think it's

kA = nomical voltage / (stall torque * gear ratio / MOI)

is that correct?

[catr1xLiu]

end up having this

    /**
     * <h2>Configures a default set of feedforward gains.</h2>
     *
     * <p>The feedforward gains (ks, kv, and ka) will be calculated using the data of the{@link DCMotor} model.</p>
     *
     * @return this instance for method changing
     * */
    public SimMotorConfigs withDefaultFeedForward() {
        return this.withFeedForward(
                this.calculateVoltage(calculateCurrent(friction), RadiansPerSecond.zero()),
                VoltsPerRadianPerSecond.ofNative(motor.nominalVoltageVolts / motor.freeSpeedRadPerSec),
                VoltsPerRadianPerSecondSquared.ofNative(
                        motor.nominalVoltageVolts / (motor.stallTorqueNewtonMeters / this.loadMOI.in(KilogramSquareMeters))
                ),
                Seconds.of(SimulatedArena.getSimulationDt()));
    }