Inertial sensors for assessment of back movement in horses during locomotion over ground.

The research article involves an examination of using inertial sensors for evaluating back movement in horses during ground locomotion. Its aim is to assess the accuracy and consistency of the measurements these sensors produce and to compare the results with data from optical motion capture.

Objectives and Methodology

• The study had two primary objectives. The first was to assess the accuracy and consistency of measurements obtained by inertial sensors for evaluating back movement in horses. The second was a comparison of these sensor readings against measurements collected through optical motion capture.
• For achieving these objectives, six non-lame horses were made to trot while their movements were recorded through both inertial sensors and optical motion capture.
• These recordings were made at five different points on the horses’ back: T6, T10, T13, L1, and S3. The inertial sensor data, thus obtained, was processed using previously published methods.
• The researchers also examined the symmetry of dorsoventral displacement based on an energy ratio, a measure of symmetry derived from Fourier methods.
• Furthermore, to compare the mocap data and that of the sensor, limits of agreement were calculated and visualised.
• Finally, the consistency of the sensor measurements was assessed through Pearson correlation coefficients and linear regression to investigate any effects of speed on movement symmetry.

Results

• The study found that dorsoventral and mediolateral sensor displacement was within ± 4-5mm (±2 standard deviations, 9-28% of movement amplitude) and energy ratio was within ±0.03 of the mocap data. These results indicate that the measurements obtained from the inertial sensors fell within acceptable accuracy ranges, and they showed good levels of consistency for back movement assessment.
• The strains and trials showed high levels of correlation (0.86-1.0) for each horse and each sensor. Variability of symmetry was lowest for T13, followed by T10, T6, L1, and S3.
• The data also showed no significant impact of speed on movement symmetry at the T6, T10, and T13 points.

Conclusion

• The study concluded that the measurements of back movement obtained from inertial sensors in the horses had acceptable accuracy and showed good consistency.
• However, due to the small mediolateral movement amplitude, changes of <25% in mediolateral amplitude, which are unlikely to be detected by visual assessment, might not be recognized.
• The research suggests that newer sensor generations with improved sensitivity and robustness promise good potential for practical use in field situations.