TL;DR
This paper introduces an innovative instrumented seat post capable of measuring all six components of user-induced sitting loads during cycling. This advancement moves beyond previous seat post designs that only captured a limited set of forces, offering a more comprehensive understanding of how cyclists interact with their bikes and potentially leading to better injury prevention and technique optimization.
What Problem Does This Paper Address?
Cycling-related discomfort and injuries, particularly skin issues and soreness, are common among cyclists. While previous research has focused on "context-induced" loads (e.g., from uneven pavement) or solely on pedal forces, the authors argue that "user-induced" sitting loads are crucial for understanding cycling technique, comfort, and injury risk. Existing instrumented seat posts often measure only a subset of these forces (e.g., vertical and anterior) or have limitations in accuracy and signal-to-noise ratio, especially for normal forces. This paper aims to fill this gap by developing and validating a seat post that can measure all six load components (three forces and three torques) to provide a complete picture of the forces a cyclist exerts on the seat.
Methods
The core innovation is a redesigned seat post incorporating a U-shaped component made of S355JR steel. This component is strategically placed to deform elastically under shear forces, enhancing sensitivity to normal loads without compromising the system's dynamics. Strain gauges are then attached to this U-shaped component and the seat post itself.
- Strain Gauge Configuration: A Wheatstone bridge configuration is used with strain gauges placed to measure tensile and compressive stresses. This setup is crucial for amplifying the signal and improving the signal-to-noise ratio.
- Six-Component Measurement: Different configurations of strain gauges are employed:
- Half-bridge configurations are used for sagittal and frontal torques, and for the normal force.
- Full-bridge configurations in a rosette arrangement are used for lateral force, anterior force, and transversal torque. This arrangement is designed to minimize the impact of shifting load application points.
- Calibration: A crucial aspect is the calibration process. The authors developed a calibration matrix that accounts for cross-sensitivity between different strain gauge measurements. This matrix is derived from applying known forces and torques to the seat post in various orientations and positions. Static calibration was deemed sufficient, as the natural frequencies of the instrumented seat post were found to be well above the relevant cycling frequencies.
- Data Processing: Raw sensor data is processed to create interpretable mean load graphs over a pedal cycle. An absolute encoder on the bottom bracket tracks pedal arm angle to synchronize data with the cycling phase. Data is filtered using a low-pass Butterworth filter.
Datasets
The paper does not explicitly mention or introduce new datasets for training machine learning models. Instead, it focuses on the design, calibration, and testing of a physical sensor. The "dataset" used for evaluation consists of measurements taken during a 15-minute relaxed cycling tour by a 60 kg cyclist on a city bike. The data collected includes 689 valid pedal cycles after erroneous data removal.
Key Insights and Ideas
- Comprehensive Load Measurement: The ability to measure all six load components (forces and torques) is a significant step beyond previous seat post sensors, enabling a holistic view of cyclist-bike interaction.
- U-Shaped Component for Sensitivity: The novel U-shaped component is key to improving sensitivity to normal forces, which are notoriously difficult to measure accurately with traditional seat post designs due to their low elastic deformation.
- Static Calibration Sufficiency: The research demonstrates that for the typical frequencies encountered in cycling (below 10 Hz), a static calibration is adequate, simplifying the calibration process compared to dynamic methods required for higher-frequency vibration analysis.
- Cyclist Technique Insights: The resulting load patterns reveal distinct cycles within the pedal stroke, offering insights into how cyclists shift weight, apply force, and potentially how asymmetrical techniques can lead to injuries.
- Commercial Viability: The design is presented as straightforward to construct, calibrate, and implement on standard bicycles, suggesting good commercial potential.
Key Results
- Accuracy: The instrumented seat post achieved a standard error percentage of full scale below 1% for anterior force, normal force, transversal torque, and sagittal torque. The maximal error percentage of full scale was also below 1% for anterior force, lateral force, transversal torque, and sagittal torque. These results are reported to be significantly better than those found in literature for pedal sensors.
- Load Distribution: The mean measured normal force accounted for approximately 65% of the cyclist's weight, with posterior and lateral forces being much smaller (around 2% and 3%, respectively).
- Pedal Cycle Dynamics: The output graphs clearly show two symmetric cycles within the pedal stroke for most loads, reflecting the power and recovery phases of each leg. Differences in force application between the left and right pedal were observed, potentially indicating asymmetrical cycling techniques. For instance, the maximal normal force was higher in the power phase of the left pedal (449 N) compared to the right (418 N).
Limitations and Future Work
- Variability in Recreational Cyclists: The study acknowledges that recreational cyclists exhibit high variability in technique due to kinematic and muscle activity differences. The wide recreational saddle used in testing further contributes to this variability, allowing for a broad range of postures. This variability, while providing rich data, also means that individual results may not be directly generalizable without considering specific cyclist characteristics.
- Future Work: The authors suggest that this sensor principle could be applied to other areas requiring six-degrees-of-freedom load measurement, such as robot arms, motorcycles, and wheelchairs. They also envision the seat post as part of a fully instrumented bicycle system to analyze the influence of external parameters (like pedal assistance, cycling duration, and bicycle type) and user characteristics on cycling technique. Further statistical research using this data could help identify maladaptive cycling techniques, muscle imbalances, and inform optimal bike fitting.
Summary updated