IMU

Functions for processing inertial measurement unit (IMU) data, including orientation kinematics, tilt estimation, and step detection.

Kinematics

biomechzoo.imu.tilt_algorithm.tilt_algorithm_data(data, ch_vert, ch_medlat, ch_antpost, plot_or_not=None)[source]

Apply the tilt correction algorithm to the vertical, mediolateral, and anteroposterior channels of zoo data.

Parameters:

  • data (dict) – Zoo file data dictionary.

  • ch_vert (str) – Name of the vertical acceleration channel.

  • ch_medlat (str) – Name of the mediolateral acceleration channel.

  • ch_antpost (str) – Name of the anteroposterior acceleration channel.

  • plot_or_not (bool, optional) – Unused. Reserved for future plotting support.

Returns:

data (dict) – Zoo file data dictionary with the tilt-corrected channels (<channel>_tilt_corr) added.

Parameters:

  • data (Dict)

  • ch_vert (str)

  • ch_medlat (str)

  • ch_antpost (str)

  • plot_or_not (bool | None)

Return type:

Dict

biomechzoo.imu.tilt_algorithm.tilt_algorithm_line(avert, amedlat, aantpost)[source]

Account for gravity and improper tilt alignment of a tri-axial trunk accelerometer.

Step 1: Extract raw measured (mean) accelerations Step 2: Calculate tilt angles Step 3: Calculate horizontal dynamic accelerations vectors Step 4: Calculate estimated provisional vertical vector Step 5: Calculate vertical dynamic vector step 6.1: Calculate the contribution of static components step 6.2 Transpose static component matrices step 7: Remove the static components from the templates of pre and post

Parameters:

  • avert (array_like) – Data predominantly in vertical direction. Expressed in g’s.

  • amedlat (array_like) – Data predominantly in medio-lateral direction. Expressed in g’s.

  • aantpost (array_like) – Data predominantly in anterior-posterior direction. Expressed in g’s.

Returns:

  • df_corrected (pandas.DataFrame) – The tilt corrected and gravity subtracted vertical, medio-lateral, and anterior-posterior acceleration signals (n x 3).

  • avert2 (ndarray) – The tilt corrected acceleration data in vertical direction.

  • amedlat2 (ndarray) – The tilt corrected acceleration data in medio-lateral direction.

  • aantpost2 (ndarray) – The tilt corrected acceleration data in anterior-posterior direction.

Parameters:
Return type:

Tuple[DataFrame, ndarray, ndarray, ndarray]

Notes

  • If average acceleration is above 5m/s^2, the signal will be corrected.

Step Detection

biomechzoo.imu.step_detection.imu_kielmat(vertical_acceleration, fsamp)[source]

Detect foot-strike and foot-off events using KielMAT.

Parameters:

  • vertical_acceleration (ndarray) – One-dimensional vertical acceleration signal in m/s/s.

  • fsamp (float) – Sampling frequency in Hz.

Returns:

  • fs (ndarray) – Foot-strike frame indices.

  • fo (ndarray) – Foot-off frame indices.

Raises:

ValueError – If the acceleration signal is not one-dimensional, contains non-finite values, or if the sampling frequency is invalid.

Parameters:
Return type:

Tuple[ndarray[Any, dtype[int64]], ndarray[Any, dtype[int64]]]

biomechzoo.imu.step_detection.imu_mcgrath(ch_line, fsamp, min_stance_t, is_filtered=False)[source]

Detect gait events using the method of McGrath et al. (2012).

The first minimum after a local maximum midswing peak is taken as initial contact (heel strike); the first valid minimum before a midswing peak is taken as terminal contact (toe off). Reference: https://doi.org/10.1007/s12283-012-0093-8

Parameters:

  • ch_line (array_like) – Vertical acceleration signal.

  • fsamp (float) – Sampling frequency in Hz.

  • min_stance_t (float) – Minimum stance time, in milliseconds, used to validate detected steps.

  • is_filtered (bool, optional) – If True, ch_line is assumed to already be filtered and no additional low-pass filtering is applied. Default is False.

Returns:

  • IC (ndarray) – Indices of detected initial contact (heel strike) events.

  • TC (ndarray) – Indices of detected terminal contact (toe off) events.

Parameters:
Return type:

Tuple[ndarray, ndarray]

biomechzoo.imu.step_detection.crash_catch(min_stance_samples, IC, TC)[source]

Ensure initial and terminal contact index arrays are the same length, truncating any extra detections.

Parameters:

  • min_stance_samples (int) – Unused. Reserved for future stance-time validation.

  • IC (list of int) – Indices of detected initial contact events.

  • TC (list of int) – Indices of detected terminal contact events.

Returns:

  • IC (ndarray) – Initial contact indices, truncated to match TC length.

  • TC (ndarray) – Terminal contact indices, truncated to match IC length.

Parameters:
Return type:

Tuple[ndarray, ndarray]