| Popis: |
As a major component of riding comfort, surface roughness has a significant impact on peoples' decision to ride bicycles. Riding comfort is most commonly derived from accelerations measured by inertial measurement units (IMUs). However, roughness metrics from different works are not directly comparable as no ‘benchmark data’ exists. This work aims at alleviating this problem by comparing several well-established methods from literature on the same data. Furthermore, to quantify the effect of different sensor systems, for each test run data from both a smartphone and an industrial grade IMU were collected. To compare the derived roughness measurements, the reliability and stability of each sensor-method combination is calculated using non-parametric statistics. The results indicate handlebar mounted smartphones to be sufficient for surface roughness assessment. Furthermore, the selected roughness calculation method has the biggest impact on resulting assessments, above the impacts of both sensor and analyzed segment length. Based on the results, recommendations for surface roughness assessment are provided in the conclusion. |
