Skip to main content
Have a personal or library account? Click to login
Evaluating pedaling strategy using a principal component analysis-based analysis of crank force vectors collected during outdoor cycling Cover

Evaluating pedaling strategy using a principal component analysis-based analysis of crank force vectors collected during outdoor cycling

Open Access
|Sep 2026

Figures & Tables

Table 1

Characteristics of the study participants; FTP/BW for professional and nonprofessional participants, and ride frequency for nonprofessional participants

25% quartileMedian75% quartile
Professional: FTP/BW (W/kg)4.88∼5.07∼5.22
Non-Professional: FTP/BW (W/kg)2.75∼3.57∼4.15
Non-Professional: Ride frequency (rides/week)2.03∼3.29∼5.52
Figure 1

Overview of the analysis.

Figure 2

Slope calculation. (a) Example of instantaneous pedal force variation across crank angles with increasing mean pedal force. (b) Linear relationship between instantaneous and mean pedal forces at each crank angle (selected angles).

Figure 3

Pedal load of extracted data (W).

Figure 4

Root mean square error of linear regression by crank angle.

Table 2

Results of principal component analysis

Eigen valueCumulative contribution rateStandard deviation
PC13.96541.241.9913
PC22.24764.611.4990
PC31.22477.311.1051
PC40.58483.390.7644
Figure 5

Baseline and principal component waveforms; baseline waveform (left axis) and PC1 and PC2 waveforms (right axis). (a) Tangential component. (b) Radial component.

Figure 6

Changes in pedal force vector waveforms by principal component values; effects of PC1 (horizontal axis) and PC2 (vertical axis) at mean loads of 100, 200, and 300 W. (a) Tangential force. (b) Radial force.

Figure 7

Changes in pedal force vector waveforms by principal component values; vector representation at 30° intervals under a mean load of 200 W.

Figure 8

Scatterplot of principal components 1 and 2; professional participants (○) and nonprofessional participants (●), with × indicating the five pedaling types.

Table 3

Results of statistical tests for differences in principal component values (* p < 0.001)

Professional participantsNonprofessional participantsMedianVariability
25% quartileMedian75% quartile25% quartileMedian75% quartile p p
PC10.761.302.31−1.59−0.271.29<0.0001*0.0007*
PC2−0.59−0.260.35−1.13−0.321.160.4238<0.0001*
DOI: https://doi.org/10.37190/abb/220441 | Journal eISSN: 2450-6303 | Journal ISSN: 1509-409X
Language: English
Page range: 48 - 61
Submitted on: Jan 8, 2026
Accepted on: Apr 8, 2026
Published on: Sep 21, 2026
Published by: Wroclaw University of Science and Technology
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2026 Masahiro Fukuda, Ryo Koyama, Takashi Inomata, Tomoki Kitawaki, published by Wroclaw University of Science and Technology
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.