pywayne-calibration-magnetometer-calibration

name	pywayne-calibration-magnetometer-calibration
description	Sensor calibration toolkit with magnetometer calibration using close-form method. Use when calibrating IMU sensors (accelerometer, gyroscope, magnetometer) to compute soft-iron matrix and hard-iron offset for magnetometer correction. Requires sensor fusion with VQF for orientation estimation.

Pywayne Calibration

pywayne.calibration.MagnetometerCalibrator provides magnetometer calibration using sensor data (accelerometer, gyroscope, magnetometer).

Quick Start

from pywayne.calibration import MagnetometerCalibrator
import numpy as np

# Sensor data: ts (N,), acc (N,3), gyro (N,3), mag (N,3)
calibrator = MagnetometerCalibrator(method='close_form')
Sm, h = calibrator.process(ts, acc, gyro, mag)

# Sm: Soft-iron matrix (3x3)
# h: Hard-iron offset vector (3,)

Input Data Format

Sensor data must be numpy arrays with matching sample counts:

ts   # (N,)     - Timestamps (seconds)
acc  # (N, 3)   - Accelerometer [ax, ay, az]
gyro # (N, 3)   - Gyroscope [gx, gy, gz]
mag  # (N, 3)   - Magnetometer [mx, my, mz]

Data requirements:

• Sensor data should cover various orientations for effective calibration
• Minimum data points required (exact number depends on calibration stability)
• Arrays must be C-contiguous (auto-converted internally)

Calibration Parameters

process() returns:

Parameter	Shape	Description
Sm	(3, 3)	Soft-iron matrix
h	(3,)	Hard-iron offset vector

Usage in Application

Apply calibration to raw magnetometer readings:

# Calibrated reading
m_calibrated = Sm @ (m_raw - h)

Temporal Calibration

Temporal calibration module exists but is reserved for future expansion. Currently no functionality is implemented.

Important Notes

• Dependencies: Requires vqf (VQF quaternion filter) and qmt (quaternion math) modules
• Method: Currently only supports close_form method
• Orientation: Uses VQF for sensor fusion and orientation estimation during calibration
• Performance: process() uses VQF/qmt batch operations and solves the final accumulated calibration matrix once
• Output: Prints calibration parameters during processing