AlphaBit - OpenML

Odometry Pods and Localization Strategy

Localization Core

Odometry is the foundation of autonomous consistency. This page explains the two main FTC layouts and how to calibrate them using your Decode codebase.

Layouts

Two pod odometry layout — Two pods + IMU: most common FTC architecture.

Three pod odometry layout — Three pods: better redundancy, more mechanics to maintain.

Compare

Type	Good For	Main Cost
2-pod + IMU	Beginner and intermediate teams, simpler maintenance, lighter design	Heading depends heavily on IMU quality and calibration
3-pod	Teams needing stronger independent heading estimation from odometry	More hardware complexity and higher mechanical tuning overhead

Step 1

Configure the Real Localizer Files

Regional RR localizer: RoadRunner/drive/opmode/TwoWheelTrackingLocalizer.java

IMU helper: drive/Skeletal_Structures/GyroscopeBHIMU.java

National Pedro reference: .../pedroPathing/Constants.java

Step 2

Set Correct Hardware Names and Signs

Mismatch here will produce fake drift and impossible trajectory tuning.

parallelEncoder = new Encoder(hardwareMap.get(DcMotorEx.class, "Back_Left"));
perpendicularEncoder = new Encoder(hardwareMap.get(DcMotorEx.class, "Back_Right"));

// Check reversals based on your mount direction
// parallelEncoder.setDirection(...);
// perpendicularEncoder.setDirection(...);

Step 3

Calibrate in This Order

Encoder direction and unit conversion first.

IMU heading offset and orientation second.

Path follower and marker timing last.

If you tune follower gains before localization is stable, every later result becomes unreliable.

Step 4

Run Standard Tests

Straight 2m forward/back endpoint check.

Square path endpoint + heading error check.

Repeated quick stop/start to catch pod slip issues.

Full-cycle test with shooter vibration to detect mount flex.

Step 5

Use Vision as an Assist, Not a Crutch

AprilTag pose refresh can reduce long-run drift.

Only apply pose correction under safe conditions (stationary/valid tags).

Your base odometry should still work even with camera blocked.

Continue to Road Runner 0.5.6 Implementation to connect localization into full autonomous cycle paths.

Support -> Discord

Choose Language / Alege Limba

AlphaBit AI Assistant