Towards Efficient Gait Analysis Using Gyroscope: Preliminary Results and Insights

Wearable sensor technology, combined with smartphone capabilities, has become a powerful and costeffective solution for gait analysis, offering real-time monitoring of human movements. This study introduces a sensor-based gait event detection system using gyroscope data from the publicly available MotionSense dataset to analyze walking and jogging. The dataset, which includes six subjects of varying demographics, provides a broader understanding of gait patterns. Gyroscope data from the X-axis was used to detect key biomechanical events, such as heel strikes, toe-offs, and mid-swings, as these align with the sagittal plane and are critical in understanding walking dynamics. The gait event detection algorithm The detection algorithm, developed in Python, incorporates advanced signal processing techniques like the Savitzky-Golay filter, to reduce noise while preserving key data. By using a threshold-based and sliding window methods, the system identifies gait events accurately and calculates key gait parameters such as cadence, stance time, swing time, and phase distributions. The results show significant differences between walking and jogging, with jogging exhibiting a higher cadence (82.71-90.23 steps/min) and shorter stance times (0.37-0.46 seconds) compared to walking (52.63-60.15 steps/min, stance time 0.61-0.85 seconds). This research demonstrates that wearable sensor technology offers precise and scalable alternatives to expensive, complex traditional gait analysis systems. The findings show the system's potential for clinical monitoring, rehabilitation, sports science, and smart wearable device development.