Securing Biomechanical Data Quality: A Comprehensive Evaluation of On-Board Accelerometers for Shock and Vibration Analysis

Abstract

(1) On-board accelerometers are increasingly employed in real-world biomechanics to monitor vibrations and shocks. This study assesses the accuracy, repeatability, and variability of three commercially available inertial measurement units (IMUs)-Xsens, Blue Trident, and Shimmer 3-in measuring vibration and shock parameters relevant to human motion analysis. (2) A controlled laboratory setup utilizing an electrodynamic shaker was employed to generate sine waves at varying frequencies and amplitudes, as well as shock profiles with defined peak accelerations and durations. (3) The results showed that Blue Trident demonstrated the highest accuracy in shock amplitude and timing, with relative errors below 6%, while Xsens provided stable measurements for low-frequency vibrations. In contrast, Shimmer 3 exhibited considerable variability in signal quality. (4) These findings offer critical insights into sensor selection based on specific application needs, ensuring optimal accuracy and reliability in dynamic measurement environments. This study lays the groundwork for improved IMU application in biomechanical research and practical deployments. Future research should continue to investigate sensor performance, particularly in angular motion contexts, to further enhance motion analysis capabilities.

Keywords: accelerometer; metrology; shock; vibration.

Figure 1

Closed-loop command of the electrodynamic shaker, studied and reference sensor.

Figure 2

Tested signals and quantities of interest ($A_s, f_s, A_i, t_i$, respectively, represent sine amplitude, sine frequency, impact amplitude, and shock duration) for sines and shocks.

Figure 3

Amplitude standard deviation normalized to the input amplitude for the same trials and the same sensors: 6 Xsens, 5 Blue Tridents, 3 Shimmer3 (specifications in Table 1), considering frequency/amplitude pairs.

Figure 4

Amplitude percentage of error of each sensor relative to the reference sensor, sine pairs [Xsens, Blue Trident, Shimmer3].

Figure 5

Percentage of relative error, for 16 g of maximum amplitude for the Xsens and Shimmer3 sensors and 35 g of maximum amplitude for the Blue Trident sensors, with 7 ms and 13 ms of shock time tests, sorted accordingly. The red dashed line indicates a 5% relative error threshold, commonly used as a benchmark for acceptable measurement accuracy.