A Novel Algorithm based on Ankle-Mounted IMU Configuration for Computing Spatiotemporal Parameters

Gait analysis plays a crucial role in understanding human locomotion, with applications in biomechanics, rehabilitation, and clinical diagnostics. Optical motion capture systems represent the gold standard in the field of gait analysis providing accurate information about human movement although they have the limit to be expensive, time-consuming and limited to the laboratory. Recently, wearable sensors, based mainly on inertial measurement units (IMUs), have emerged as a promising alternative to optoelectronic systems, enabling real-world gait monitoring and in a simpler, faster and cheaper way. However, variations in sensor placement and in the type of inertial signal used to detect gait events can impact on the accuracy of gait parameter estimation. This study presents a novel algorithm for computing spatiotemporal gait parameters using angular velocity signals around the mediolateral axis acquired from two IMUs paced on the ankles. The algorithm was validated comparing its measures with the ones computed using a commercial wearable inertial system based on foot-mounted IMUs. Thirty healthy participants performed a 7-meter walkway protocol wearing IMUs on the lower limbs according to the specification of both methods. Gait events, including toe-off, mid-swing, and heel-strike, were identified, and five spatiotemporal parameters (cadence, gait cycle time, stance phase, swing phase, and double support phase) were calculated. The validation study was carried out using two-tailed paired test and Bland-Altman analysis. The results showed the presence of constant systematic errors in estimating the following parameters: cadence, stance phase, swing phase, and double support phase; while gait cycle time showed a good agreement between the two methods. However, a constant systematic error can be simply corrected by zeroing of the bias through a subtraction operation. In conclusion, the proposed algorithm can be considered a valid tool for gait assessment even if further refinements are needed to remove the constant systematic errors that occur in the spatiotemporal parameters estimation. Future development, considering all the components of both acceleration and angular velocity signals acquired from the two IMUs placed on the ankles, might improve the reliability of the proposed algorithm to compute spatiotemporal parameters.