Gyroscopes - Software

Note

This section covers gyros in software. For a hardware guide to gyros, see Gyroscopes - Hardware.

A gyroscope, or “gyro,” is an angular rate sensor typically used in robotics to measure and/or stabilize robot headings. WPILib natively provides specific support for the ADXRS450 gyro available in the kit of parts, as well as more general support for a wider variety of analog gyros through the AnalogGyro class. Most common 3rd party gyros inherit from the Gyro interface making them easily usable too!

The Gyro interface

All natively-supported gyro objects in WPILib implement the Gyro interface (Java, C++). This interface provides methods for getting the current angular rate and heading, zeroing the current heading, and calibrating the gyro.

Note

It is crucial that the robot remain stationary while calibrating a gyro.

ADIS16448

The ADIS16448 uses the ADIS16448_IMU class and implements the Gyro interface. See the ADIS16448 documentation for additional information and examples.

// ADIS16448 plugged into the MXP port
ADIS16448_IMU gyro = new ADIS16448_IMU();

ADIS16470

The ADIS16470 uses the ADIS16470_IMU class and implements the Gyro interface. See the ADIS16470 documentation for additional information and examples.

// ADIS16470 plugged into the MXP port
ADIS16470_IMU gyro = new ADIS16470_IMU();

ADXRS450_Gyro

The ADXRS450_Gyro class (Java, C++) provides support for the Analog Devices ADXRS450 gyro available in the kit of parts, which connects over the SPI bus.

Note

ADXRS450 Gyro accumulation is handled through special circuitry in the FPGA; accordingly only a single instance of ADXRS450_Gyro may be used.

// Creates an ADXRS450_Gyro object on the onboard SPI port
ADXRS450_Gyro gyro = new ADXRS450_Gyro();

AnalogGyro

The AnalogGyro class (Java, C++) provides support for any single-axis gyro with an analog output.

Note

Gyro accumulation is handled through special circuitry in the FPGA; accordingly, AnalogGyro`s may only be used on analog ports 0 and 1.

// Creates an AnalogGyro object on port 0
AnalogGyro gyro = new AnalogGyro(0);

Pigeon

The Pigeon should use the WPI_PigeonIMU class that implements Gyro. The Pigeon can either be connected with CAN or by data cable to a TalonSRX. The Pigeon IMU User’s Guide contains full details on using the Pigeon.

WPI_PigeonIMU gyro = new WPI_PigeonIMU(0); // Pigeon is on CAN Bus with device ID 0
// OR (choose one or the other based on your connection)
TalonSRX talon = new TalonSRX(0); // TalonSRX is on CAN Bus with device ID 0
WPI_PigeonIMU gyro = new WPI_PigeonIMU(talon); // Pigeon uses the talon created above

Using gyros in code

Note

As gyros measure rate rather than position, position is inferred by integrating (adding up) the rate signal to get the total change in angle. Thus, gyro angle measurements are always relative to some arbitrary zero angle (determined by the angle of the gyro when either the robot was turned on or a zeroing method was called), and are also subject to accumulated errors (called “drift”) that increase in magnitude the longer the gyro is used. The amount of drift varies with the type of gyro.

Gyros are extremely useful in FRC for both measuring and controlling robot heading. Since FRC matches are generally short, total gyro drift over the course of an FRC match tends to be manageably small (on the order of a couple of degrees for a good-quality gyro). Moreover, not all useful gyro applications require the absolute heading measurement to remain accurate over the course of the entire match.

Displaying the robot heading on the dashboard

Shuffleboard includes a widget for displaying heading data from a Gyro in the form of a compass. This can be helpful for viewing the robot heading when sight lines to the robot are obscured:

// Use gyro declaration from above here

public void robotInit() {
    // Places a compass indicator for the gyro heading on the dashboard
    Shuffleboard.getTab("Example tab").add(gyro);
}

Stabilizing heading while driving

A very common use for a gyro is to stabilize robot heading while driving, so that the robot drives straight. This is especially important for holonomic drives such as mecanum and swerve, but is extremely useful for tank drives as well.

This is typically achieved by closing a PID controller on either the turn rate or the heading, and piping the output of the loop to one’s turning control (for a tank drive, this would be a speed differential between the two sides of the drive).

Warning

Like with all control loops, users should be careful to ensure that the sensor direction and the turning direction are consistent. If they are not, the loop will be unstable and the robot will turn wildly.

Example: Tank drive stabilization using turn rate

The following example shows how to stabilize heading using a simple P loop closed on the turn rate. Since a robot that is not turning should have a turn rate of zero, the setpoint for the loop is implicitly zero, making this method very simple.

// Use gyro declaration from above here

// The gain for a simple P loop
double kP = 1;

// Initialize motor controllers and drive
Spark left1 = new Spark(0);
Spark left2 = new Spark(1);

Spark right1 = new Spark(2);
Spark right2 = new Spark(3);

MotorControllerGroup leftMotors = new MotorControllerGroup(left1, left2);
MotorControllerGroup rightMotors = new MotorControllerGroup(right1, right2);

DifferentialDrive drive = new DifferentialDrive(leftMotors, rightMotors);

@Override
public void autonomousPeriodic() {
    // Setpoint is implicitly 0, since we don't want the heading to change
    double error = -gyro.getRate();

    // Drives forward continuously at half speed, using the gyro to stabilize the heading
    drive.tankDrive(.5 + kP * error, .5 - kP * error);
}

More-advanced implementations can use a more-complicated control loop. When closing the loop on the turn rate for heading stabilization, PI loops are particularly effective.

Example: Tank drive stabilization using heading

The following example shows how to stabilize heading using a simple P loop closed on the heading. Unlike in the turn rate example, we will need to set the setpoint to the current heading before starting motion, making this method slightly more-complicated.

// Use gyro declaration from above here

// The gain for a simple P loop
double kP = 1;

// The heading of the robot when starting the motion
double heading;

// Initialize motor controllers and drive
Spark left1 = new Spark(0);
Spark left2 = new Spark(1);

Spark right1 = new Spark(2);
Spark right2 = new Spark(3);

MotorControllerGroup leftMotors = new MotorControllerGroup(left1, left2);
MotorControllerGroup rightMotors = new MotorControllerGroup(right1, right2);

DifferentialDrive drive = new DifferentialDrive(leftMotors, rightMotors);

@Override
public void autonomousInit() {
    // Set setpoint to current heading at start of auto
    heading = gyro.getAngle();
}

@Override
public void autonomousPeriodic() {
    double error = heading - gyro.getAngle();

    // Drives forward continuously at half speed, using the gyro to stabilize the heading
    drive.tankDrive(.5 + kP * error, .5 - kP * error);
}

More-advanced implementations can use a more-complicated control loop. When closing the loop on the heading for heading stabilization, PD loops are particularly effective.

Turning to a set heading

Another common and highly-useful application for a gyro is turning a robot to face a specified direction. This can be a component of an autonomous driving routine, or can be used during teleoperated control to help align a robot with field elements.

Much like with heading stabilization, this is often accomplished with a PID loop - unlike with stabilization, however, the loop can only be closed on the heading. The following example code will turn the robot to face 90 degrees with a simple P loop:

// Use gyro declaration from above here

// The gain for a simple P loop
double kP = 1;

// Initialize motor controllers and drive
Spark left1 = new Spark(0);
Spark left2 = new Spark(1);

Spark right1 = new Spark(2);
Spark right2 = new Spark(3);

MotorControllerGroup leftMotors = new MotorControllerGroup(left1, left2);
MotorControllerGroup rightMotors = new MotorControllerGroup(right1, right2);

DifferentialDrive drive = new DifferentialDrive(leftMotors, rightMotors);

@Override
public void autonomousPeriodic() {
    // Find the heading error; setpoint is 90
    double error = 90 - gyro.getAngle();

    // Turns the robot to face the desired direction
    drive.tankDrive(kP * error, kP * error);
}

As before, more-advanced implementations can use more-complicated control loops.

Note

Turn-to-angle loops can be tricky to tune correctly due to static friction in the drivetrain, especially if a simple P loop is used. There are a number of ways to account for this; one of the most common/effective is to add a “minimum output” to the output of the control loop. Another effective strategy is to cascade to well-tuned velocity controllers on each side of the drive.