org.firstinspires.ftc.teamcode.util

Class RegressionUtil

java.lang.Object

org.firstinspires.ftc.teamcode.util.RegressionUtil

public class RegressionUtil
extends java.lang.Object

Various regression utilities.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	RegressionUtil.AccelResult Feedforward parameter estimates from the ramp regression and additional summary statistics
static class	RegressionUtil.RampResult Feedforward parameter estimates from the ramp regression and additional summary statistics

Constructor Summary

Constructors

Constructor and Description
RegressionUtil()

Method Summary

Modifier and Type	Method and Description
static RegressionUtil.AccelResult	fitAccelData(java.util.List<java.lang.Double> timeSamples, java.util.List<java.lang.Double> positionSamples, java.util.List<java.lang.Double> powerSamples, RegressionUtil.RampResult rampResult, java.io.File file) Run regression to compute acceleration feedforward.
static RegressionUtil.RampResult	fitRampData(java.util.List<java.lang.Double> timeSamples, java.util.List<java.lang.Double> positionSamples, java.util.List<java.lang.Double> powerSamples, boolean fitStatic, java.io.File file) Run regression to compute velocity and static feedforward from ramp test data.
private static java.util.List<java.lang.Double>	numericalDerivative(java.util.List<java.lang.Double> x, java.util.List<java.lang.Double> y) Numerically compute dy/dx from the given x and y values.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RegressionUtil

public RegressionUtil()

Method Detail

numericalDerivative

private static java.util.List<java.lang.Double> numericalDerivative(java.util.List<java.lang.Double> x,
                                                                    java.util.List<java.lang.Double> y)

Numerically compute dy/dx from the given x and y values. The returned list is padded to match the length of the original sequences.

Parameters:

x - x-values

y - y-values

Returns:

derivative values

fitRampData

public static RegressionUtil.RampResult fitRampData(java.util.List<java.lang.Double> timeSamples,
                                                    java.util.List<java.lang.Double> positionSamples,
                                                    java.util.List<java.lang.Double> powerSamples,
                                                    boolean fitStatic,
                                                    @Nullable
                                                    java.io.File file)

Run regression to compute velocity and static feedforward from ramp test data. Here's the general procedure for gathering the requisite data: 1. Slowly ramp the motor power/voltage and record encoder values along the way. 2. Run a linear regression on the encoder velocity vs. motor power plot to obtain a slope (kV) and an optional intercept (kStatic).

Parameters:

timeSamples - time samples

positionSamples - position samples

powerSamples - power samples

fitStatic - fit kStatic

file - log file

fitAccelData

public static RegressionUtil.AccelResult fitAccelData(java.util.List<java.lang.Double> timeSamples,
                                                      java.util.List<java.lang.Double> positionSamples,
                                                      java.util.List<java.lang.Double> powerSamples,
                                                      RegressionUtil.RampResult rampResult,
                                                      @Nullable
                                                      java.io.File file)

Run regression to compute acceleration feedforward.

Parameters:

timeSamples - time samples

positionSamples - position samples

powerSamples - power samples

rampResult - ramp result

file - log file