Step 2: Entering the Calculated Constants
In C++, it is important that the feedforward constants be entered as the correct unit type. For more information on C++ units, see The C++ Units Library.
Now that we have our system constants, it is time to place them in our code. The recommended place for this is the
Constants file of the standard command-based project structure.
Firstly, we must enter the feedforward and feedback gains which we obtained from the identification tool.
Feedforward and feedback gains do not, in general, transfer across robots. Do not use the gains from this tutorial for your own robot.
39 // These are example values only - DO NOT USE THESE FOR YOUR OWN ROBOT! 40 // These characterization values MUST be determined either experimentally or theoretically 41 // for *your* robot's drive. 42 // The Robot Characterization Toolsuite provides a convenient tool for obtaining these 43 // values for your robot. 44 public static final double ksVolts = 0.22; 45 public static final double kvVoltSecondsPerMeter = 1.98; 46 public static final double kaVoltSecondsSquaredPerMeter = 0.2; 47 48 // Example value only - as above, this must be tuned for your drive! 49 public static final double kPDriveVel = 8.5;
47// These are example values only - DO NOT USE THESE FOR YOUR OWN ROBOT! 48// These characterization values MUST be determined either experimentally or 49// theoretically for *your* robot's drive. The Robot Characterization 50// Toolsuite provides a convenient tool for obtaining these values for your 51// robot. 52constexpr auto ks = 0.22_V; 53constexpr auto kv = 1.98 * 1_V * 1_s / 1_m; 54constexpr auto ka = 0.2 * 1_V * 1_s * 1_s / 1_m; 55 56// Example value only - as above, this must be tuned for your drive! 57constexpr double kPDriveVel = 8.5;
Additionally, we must create an instance of the
DifferentialDriveKinematics class, which allows us to use the trackwidth (i.e. horizontal distance between the wheels) of the robot to convert from chassis speeds to wheel speeds. As elsewhere, we keep our units in meters.
29 public static final double kTrackwidthMeters = 0.69; 30 public static final DifferentialDriveKinematics kDriveKinematics = 31 new DifferentialDriveKinematics(kTrackwidthMeters);
38constexpr auto kTrackwidth = 0.69_m; 39extern const frc::DifferentialDriveKinematics kDriveKinematics;
Max Trajectory Velocity/Acceleration
We must also decide on a nominal max acceleration and max velocity for the robot during path-following. The maximum velocity value should be set somewhat below the nominal free-speed of the robot. Due to the later use of the
DifferentialDriveVoltageConstraint, the maximum acceleration value is not extremely crucial.
Max velocity and acceleration, as defined here, are applied only during trajectory generation. They do not limit the
RamseteCommand itself, which may give values to the
DriveSubsystem that can cause the robot to greatly exceed these velocities and accelerations.
57 public static final double kMaxSpeedMetersPerSecond = 3; 58 public static final double kMaxAccelerationMetersPerSecondSquared = 1;
61constexpr auto kMaxSpeed = 3_mps; 62constexpr auto kMaxAcceleration = 1_mps_sq;
Finally, we must include a pair of parameters for the RAMSETE controller. The values shown below should work well for most robots, provided distances have been correctly measured in meters - for more information on tuning these values (if it is required), see Constructing the Ramsete Controller Object.
60 // Reasonable baseline values for a RAMSETE follower in units of meters and seconds 61 public static final double kRamseteB = 2; 62 public static final double kRamseteZeta = 0.7;
64// Reasonable baseline values for a RAMSETE follower in units of meters and 65// seconds 66constexpr auto kRamseteB = 2.0 * 1_rad * 1_rad / (1_m * 1_m); 67constexpr auto kRamseteZeta = 0.7 / 1_rad;