International Research Journal of Engineering and Technology (IRJET) Volume: 12 Issue: 10 | Oct 2025
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
Quantifying Braking Behavior in Cyclists Using Smartphone Sensors: A Comparative Analysis of Novice and Experienced Riders Shrijat Bose1 1Leland High School, 890 Hampswood Way ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Braking skill is central to safe and fast cycling,
braking styles (late/aggressive vs early/conservative) linked to experience [8].
yet it is rarely quantified outside controlled labs. We used a smartphone‑based inertial measurement unit (IMU) and GPS to capture real‑world deceleration during a standardized maneuver performed by multiple riders of known experience. Acceleration time series were smoothed and segmented into four phases—Approach, Braking, Cornering, and Recovery—via thresholding and peak detection. From these, we derived phase durations, peak deceleration, time‑to‑peak, and a post‑braking damping ratio computed from logarithmic decrement. Riders were aligned by (i) first brake onset, (ii) peak decel, and (iii) relative ride time to enable inter‑subject comparison. Group summaries and visualizations (radar/spider plots, scatter plots, correlation heatmaps, and small multiples) reveal that experienced riders tended to brake more decisively (shorter braking duration, earlier time‑to‑peak) and exhibited higher damping ratios (smoother recovery), while novices showed larger, more variable peak decelerations. The pipeline is low‑cost and replicable, enabling classroom‑scale studies and coach feedback. Implications for rider training and safety are discussed.
We propose and validate a field pipeline that (i) segments the braking maneuver into four interpretable phases, (ii) extracts comparable features across riders, and (iii) contrasts novice vs experienced groups. We also introduce a post‑braking damping ratio (from logarithmic decrement) as a candidate stability metric. Contributions. (1) A practical, fully reproducible analysis using only smartphone sensors; (2) a four‑phase segmentation with multi‑anchor alignment; (3) group‑level comparisons and visualizations suitable for coaching dashboards; (4) an annotated dataset template and reporting checklist to support classroom replication.
2. METHODOLOGY 2.1 Participants and Apparatus
This study involved multiple volunteer riders who selfreported their cycling experience level. The cohort was divided into an Experienced group (n = 11 riders, including the author) and a Novice group (n = 9 riders). All riders performed the same standardized braking maneuver under similar environmental conditions. The test scenario was a short straight approach leading into a low-speed turn, requiring the rider to brake before and through the corner. Each rider was instructed on the course and asked to execute the braking as they normally would given the turn.
Key Words: bicycle braking; smartphone IMU; inertial sensors; acceleration analysis; time-to-peak deceleration; damping ratio; logarithmic decrement; phase segmentation; cycling safety; novice vs experienced; field study; coaching analytics
To ensure consistency, all riders used the same bicycle during the trials: a Trek Roscoe 8 hardtail mountain bike (aluminum frame, ~4 years old) weighing approximately 14.2 kg. The bike was equipped with 27.5 × 2.8″ Maxxis tires inflated to 35 PSI (the maximum recommended pressure). Maintaining high tire pressure helped minimize rolling resistance differences and ensured that sluggish tire effects did not confound the results. The seat height was kept fixed at about 38″ (96.5 cm) for all riders – a compromise height that all could use – to eliminate seat position as a variable influencing balance or braking posture. In other words, every participant rode under virtually identical bike setup and surface conditions, isolating rider behavior as the primary source of variation.
1.INTRODUCTION Effective braking is a hallmark of proficient cycling because it governs entry speed, cornering stability, and crash avoidance. Most prior work on bicycle braking has focused on hardware or controlled tests (e.g., brake systems, stopping distances) rather than rider‑specific control strategies in the wild. Recent studies have quantified bicycle stopping performance under different brake configurations and surfaces, reporting mean decelerations on the order of ~2–5 m/s² depending on conditions [1–3]. Complementary strands of research validate consumer‑grade sensors for sports biomechanics, with inertial units enabling high‑frequency motion capture at low cost [4–6], and smartphone GPS offering ~7–13 m median horizontal error in urban settings—adequate for coarse trajectory context but not for fine alignment [7]. Perception–action research further suggests distinct
© 2025, IRJET
|
Impact Factor value: 8.315
2.2 Data Collection A smartphone (placed securely on the bike) running a sensor-logging app recorded tri-axial acceleration and GPS data throughout each run. The accelerometer was the main sensor of interest; we |
ISO 9001:2008 Certified Journal
|
Page 386