Detecting Heel Strike and toe off Events Using Kinematic Methods and LSTM Models
Organizations: School of Artificial Intelligence and Robotics, Hunan University, Changsha, China · School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore · Fourth Hospital of Changsha, Hunan Normal University, Changsha, China · Department of Stomatology, Hunan Provincial People’s Hospital, Hunan Normal University, Changsha, China ·
Abstract
Accurate gait event detection is crucial for gait analysis, rehabilitation, and assistive technology, particularly in exoskeleton control, where precise identification of stance and swing phases is essential. This study evaluated the performance of seven kinematics-based methods and a Long Short-Term Memory (LSTM) model for detecting heel strike and toe-off events across 4363 gait cycles from 588 able-bodied subjects. The results indicated that while the Zeni et al. method achieved the highest accuracy among kinematics-based approaches, other methods exhibited systematic biases or required dataset-specific tuning. The LSTM model performed comparably to Zeni et al., providing a data-driven alternative without systematic bias. These findings highlight the potential of deep learning-based approaches for gait event detection while emphasizing the need for further validation in clinical populations and across diverse gait conditions. Future research will explore the generalizability of these methods in pathological populations, such as individuals with post-stroke conditions and knee osteoarthritis, as well as their robustness across varied gait conditions and data collection settings to enhance their applicability in rehabilitation and exoskeleton control.
Figures & tables
| Method | Markers | HS Detection Features | TO Detection Features |
|---|---|---|---|
| Foot trajectories relative to pelvis (Zeni et al. 2008) | Heel, toe, and pelvis | Local maximum horizontal position of the heel marker relative to the pelvis | Local minimal horizontal position of the toe marker relative to the pelvis |
| High-pass filtered foot markers (Desailly et al. 2009) | Heel and toe | First maximum between high pass filtered (0.5 * gait frequency) horizontal heel marker trajectory | Last minimum between high pass filtered (1.1 * gait frequency) horizontal toe marker trajectory |
| Foot markers velocity extreme (O’Connor et al. 2007) | Heel and toe | Local minimal vertical velocity of the mid-point of heel and toe markers | Local maximum vertical velocity of the mid-point of heel and toe markers |
| Foot markers Velocity Threshold (Ghoussayni et al. 2004) | Heel and toe | Sagittal velocity of heel marker lower than 50 cm/s | Sagittal velocity of toe marker higher than 50 cm/s |
| Foot markers acceleration extreme (Hreljac et al. 2000) | Heel and toe | Local maximum vertical acceleration of the heel marker, with the condition that the derivative of the acceleration (jerk) was zero. | Local maximum horizontal acceleration of the toe marker, also with the constraint that the relevant jerk was zero |
| Foot markers acceleration extreme (Hsue et al. 2007) | Heel and toe | Local minimal horizontal acceleration of heel marker | Local maximum horizontal acceleration of toe marker |
| Method | Heel Strike | Toe off | ||
|---|---|---|---|---|
| Mean (ms) | STD (ms) | Mean (ms) | STD (ms) | |
| Zeni et al. | 4.78 | 9.56 | -8.66 | 6.05 |
| Desailly et al. | 22.17 | 8.66 | 7.45 | 7.62 |
| O’Connor et al. | 46.73 | 7.72 | -16.27 | 8.35 |
| Ghoussayni et al. | 50.49 | 11.28 | -61.96 | 8.67 |
| Hreljac et al. | 32.28 | 31.56 | 4.70 | 8.14 |
Gait Analysis
Gait analysis uses movement and biomechanical signals to characterize walking, recognize people, assess health, and guide rehabilitation or locomotion.
See all 75 papers