IMU-based gait analysis in lower limb prosthesis users: Comparison of step demarcation algorithms

Area of Science:

• Biomechanics
• Rehabilitation Engineering
• Wearable Technology

Background:

• Inertial Measurement Unit (IMU)-based gait analysis is validated in healthy individuals.
• Limited data exists on IMU algorithm performance in clinical populations with gait deviations, like lower limb prosthesis users (LLPUs).

Purpose of the Study:

• To compare the efficacy of three distinct IMU-based algorithms for step demarcation in LLPUs.
• To evaluate algorithm performance in transtibial (TTA), transfemoral (TFA) amputees, and healthy controls (HC).

Main Methods:

• A single IMU sensor was used on 17 TTAs, 16 TFAs, and 14 HCs during overground walking.
• Step demarcation was assessed using fore-aft acceleration via peak detection, zero-crossing, or peak preceding zero-crossing.
• Variability in acceleration waveforms and step duration was quantified for each algorithm.

Main Results:

• The zero-crossing algorithm demonstrated superior performance in 65% of TTAs, 81% of TFAs, and 71% of HCs.
• This superiority was indicated by lower standard deviations in acceleration and more consistent step duration distribution.
• Qualitative assessment showed more reliable step demarcation with the zero-crossing method.

Conclusions:

• Algorithm choice significantly impacts gait spatiotemporal metric interpretation in LLPUs.
• The fore-aft acceleration zero-crossing is a more dependable feature for step demarcation in LLPU gait analysis.
• This finding enhances the reliability of IMU-based gait analysis for clinical applications in amputee populations.