AlphaBit - OpenML

Auto Aiming Turret - Getting Started

Auto Aim Starter

Auto aim in Decode is centered in ArtifactControl. The safest way to adopt it is to build in layers: geometry first, then turret/flywheel, then burst automation.

Step 1

Verify Manual Shooter First

Before auto aim, ensure manual turret and flywheel controls work correctly.

Calibrate servo centers and bounds for LeftTurret, RightTurret, AngleTurret.

Step 2

Enable Distance-Based Models

Decode uses polynomial models and clamp limits:

angleTurretPosition = (0.0000207725 * (basketDistance * basketDistance))
                    - (0.00755001 * basketDistance)
                    + 0.865169;
angleTurretPosition = clamp(angleTurretPosition, 0.25, 0.75);

flyWheelPower = ((-3.15936e-7) * d * d * d)
              + (0.000074273 * d * d)
              - (0.00230794 * d)
              + 0.606381;
flyWheelPower = clamp(flyWheelPower, 0.6, 0.87);

Step 3

Protect Servos With Deadzones

Use horizontal and vertical deadzones to prevent jitter and wear.

Only update servos when delta is meaningful.

Step 4

Add Safety Gates for Launch

Use area checks (example: areaOfThrowing()).

Require stationary or controlled state for auto-shoot actions.

Keep immediate manual override on controller.

Auto aim without strong safety gates is risky in scrimmage and match conditions.

Step 5

Integrate With Burst Logic

Connect aim output to burst sequence only after individual subsystems are stable.

Validate with repeated cycle tests, not only single-shot tests.

Modes

IMU only: most deterministic fallback.

Webcam only: useful for vision experiments and corrections.

IMU + Webcam fusion: preferred when both sources are reliable.

Continue with IMU Only implementation first, then compare against webcam-based modes.

Support -> Discord

Choose Language / Alege Limba

AlphaBit AI Assistant