MAX PIANTA
← All projectsResume ↓
Undergraduate dissertation · 2025

RowBot: a low-cost rowing power meter

A clip-on sensor system that streams stroke power, stroke rate and blade angles to a phone, designed and validated in Siemens NX for a fraction of the cost of commercial systems.

  • ContextBEng final-year dissertation, Oxford Brookes University
  • RoleSole designer and engineer
  • ToolsSiemens NX, MATLAB
Siemens NXFEAMATLABProduct designSensors & BLE
RowBot sensor module on the oarlock gate, connected by a shielded cable to the communication module

Key numbers

~£224
prototype cost, vs £1,800–£12,000 for commercial systems
2,000 N
peak rowing load the housing was validated against
57×
margin: 3.57 MPa peak stress vs 205 MPa yield
32.7 kHz
first natural frequency, far above the ~0.6 Hz stroke

The problem

Power is the most direct measure of rowing performance, but on-water power meters cost £1,800–£12,000 for a pair of athletes. That puts stroke-by-stroke data out of reach for most clubs, schools and university crews.

Without live feedback, technique errors go uncorrected and training load is hard to track, which raises injury risk. The goal: elite-level insight at a club-level price, mounting to any shell in minutes.

Where the sensor and communication modules mount on a pair shell
Where the sensor and communication modules mount on a pair shell

What I did

  • Sensing: compared load sensors and chose waterproof foil strain gauges on the oarlock pin for force, plus a rotary encoder for blade angle.
  • Design: modelled a two-part system in Siemens NX: a marine-grade stainless steel (AISI 310) sensor housing that slots over the pin, and a 3D-printed Nylon 12 communication module with a BLE microcontroller.
  • Simulation: ran a mesh convergence study, then static, fatigue and modal FEA on the gate, a simplified housing and the optimised housing.
  • Data: wrote a MATLAB script that turns raw strain and encoder data into power, stroke rate and catch/finish angles for every stroke.
Exploded assembly: C-clip, gate, sensor housing top and bottom
Exploded assembly: C-clip, gate, sensor housing top and bottom

Results

  • The optimised housing held the full 2,000 N peak load at just 3.57 MPa von Mises stress, well under the 205 MPa yield strength.
  • Fatigue analysis showed a life far beyond years of daily rowing, and the first natural frequency (32.7 kHz) rules out resonance.
  • The MATLAB pipeline produced clean stroke-by-stroke power, rate and angle data in 2–3 seconds, quick enough for feedback between strokes.
  • A working prototype costs about £224 in parts, with a path to under £300 per unit at scale.
  • Won 3rd Prize at the Oxford Brookes Enterprise Tech Show, out of 500 STEM projects, and was nominated for a National STEM Inspiration Award.
Von Mises stress on the optimised sensor housing under a 2,000 N load (max 3.57 MPa)
Von Mises stress on the optimised sensor housing under a 2,000 N load (max 3.57 MPa)

From data to feedback

Power output across the stroke, plotted against blade angle in MATLAB
Power output across the stroke, plotted against blade angle in MATLAB
App concepts: live power curve, session averages and stroke-by-stroke display
App concepts: live power curve, session averages and stroke-by-stroke display
Communication module: 3D-printed Nylon 12 enclosure (120 × 35 × 75 mm) with status LEDs
Communication module: 3D-printed Nylon 12 enclosure (120 × 35 × 75 mm) with status LEDs

What's next

  • Build a waterproof, field-ready prototype and start on-water testing with athletes and coaches.
  • Move processing from MATLAB to embedded code or Python for true real-time feedback.
  • Combine BLE, GPS and motion sensing on a single chip and redesign the housing for injection moulding.
View the project poster ↗← All projects