Skip to main content
SysId is a tool for performing system identification on robot mechanisms. It helps you characterize your robot’s drivetrain, arms, elevators, and other mechanisms by collecting data and calculating feedforward and feedback gains.

​
Overview

System identification is the process of determining mathematical models for your robot’s physical systems. SysId automates data collection and analysis to generate accurate motor characterization values that can be used in your robot code for better control.

​
What is System Identification?

System identification determines the relationship between voltage applied to motors and the resulting motion. The key values calculated are:
  • kS (Static Gain): Voltage needed to overcome static friction
  • kV (Velocity Gain): Voltage needed per unit of velocity
  • kA (Acceleration Gain): Voltage needed per unit of acceleration
  • kG (Gravity Gain): Voltage needed to overcome gravity (for arms/elevators)
These values are used in feedforward controllers to improve motion control accuracy.

​
Running SysId

1

Navigate to the WPILib repository

Open a terminal and navigate to the root of the WPILib repository:
cd /path/to/allwpilib
2

Run SysId using Gradle

Execute the SysId application using the Gradle wrapper:
./gradlew sysid:run
On Windows:
gradlew.bat sysid:run
3

Create a robot project

In SysId, create a new project or use the built-in robot code generator to create a test project for your mechanism.
4

Deploy to your robot

Deploy the generated code to your robot using your normal deployment process.
5

Run tests

Back in SysId, connect to your robot and run the following test types:
  • Quasistatic Tests: Slowly ramp voltage to measure steady-state behavior
  • Dynamic Tests: Apply a step voltage to measure transient response
Run tests in both forward and backward directions.
6

Analyze data

After collecting data, SysId will automatically analyze it and calculate the system parameters.

​
Supported Mechanism Types

SysId supports characterization for:
  • Drivetrain: Linear and angular characterization
  • Arm: Mechanisms that rotate against gravity
  • Elevator: Mechanisms that translate against gravity
  • Simple: Mechanisms without gravitational effects (flywheels, rollers, etc.)

​
Data Collection

​
NetworkTables Entries

SysId communicates with your robot program via NetworkTables:
EntryTypeDescription
/SmartDashboard/SysIdTelemetrystringTelemetry data sent from robot after test completes
/SmartDashboard/SysIdVoltageCommanddoubleVoltage command or ramp rate
/SmartDashboard/SysIdTestTypestring”Quasistatic” or “Dynamic”
/SmartDashboard/SysIdRotateboolFor drivetrain rotation tests

​
Telemetry Format

Data is collected with the following format: Non-Drivetrain: timestamp, voltage, position, velocity Drivetrain: timestamp, l voltage, r voltage, l position, r position, l velocity, r velocity, angle, angular rate

​
Supported Units

  • Meters
  • Feet
  • Inches
  • Radians
  • Rotations
  • Degrees

​
Troubleshooting

If SysId fails to generate plots or encounters errors:
1

Check the data logs

SysId saves data in .wpilog format. Use AdvantageScope to view these files for troubleshooting.
2

Verify NetworkTables connection

Ensure your robot is connected and publishing to the correct NetworkTables entries.
3

Check data quality

Ensure your robot has enough space to safely run the tests and that encoders are properly configured.

​
Using Results

After characterization, SysId provides feedforward gains that can be used in your robot code: 
// Example: Using SysId results in Java
SimpleMotorFeedforward feedforward = new SimpleMotorFeedforward(
    kS, // Static gain from SysId
    kV, // Velocity gain from SysId
    kA  // Acceleration gain from SysId
);

// Example: Using SysId results in C++
frc::SimpleMotorFeedforward feedforward{kS, kV, kA};

​
Building SysId

To build SysId from source: 
./gradlew sysid:build

​
System Requirements

  • Java Development Kit (JDK) 17 or later
  • C++ Compiler:
    • Linux: GCC 11 or greater
    • Windows: Visual Studio 2022 with C++ support
    • macOS: Xcode 14 or later command-line tools

Build docs developers (and LLMs) love

Get started for freeTalk to us