Motorcycle track days are thrilling, but improving your lap times and technique often requires expensive professional telemetry equipment. To bridge this gap, I built PulseCraft, a system also known as the Rider Analytics Platform for Track Optimization (RAPTOR). This project leverages IoT, data science, and modern web development to provide accessible, high-quality analytics for amateur riders.
PulseCraft Dashboard
The Hardware: Collecting the Data
At the heart of PulseCraft is a custom hardware module designed around the ESP32 microcontroller. We needed to capture a comprehensive picture of the motorcycle's dynamics without breaking the bank.
The sensor array includes:
- MPU-6050 IMU: To track lean angles, lateral acceleration, and pitch (braking/acceleration forces).
- NEO-6M GPS: Essential for tracking speed, mapping the racing line, and lap timing.
- Bluetooth OBD Adapter: Connects directly to the bike's diagnostic port to pull engine data like RPM and throttle percentage.
- MicroSD Logging: Ensures high-frequency data is safely stored locally during the intense vibrations of track riding.
Processing the Telemetry
Raw data is just noise until you process it. Once a track session is complete, the ESP32 uploads the CSV/JSON logs via Wi-Fi to a Python-based backend (built with Flask/FastAPI).
Here, tools like Pandas and NumPy go to work. We process the timestamped sensor streams to compute crucial performance metrics for each lap:
- Control Metrics: Throttle Smoothness Index and Braking Jerk.
- Stability Metrics: Lean Angle Variance and Lateral Acceleration RMS.
- Efficiency Metrics: Speed-to-Lean Ratio and Speed Loss During Braking.
Machine Learning: Beyond Simple Charts
Instead of just showing the rider a bunch of graphs, PulseCraft uses Machine Learning (Scikit-learn) to provide actionable insights. We focused on unsupervised learning approaches:
- Clustering (K-Means/DBSCAN): Groups similar riding styles to help riders understand their dominant habits.
- Anomaly Detection (Isolation Forest): Automatically flags unstable braking zones or erratic cornering events, acting as a virtual riding coach.
Lean Angle and Braking Analytics
The Frontend: A Motorsport-Inspired Dashboard
The final piece of the puzzle is the web dashboard where the rider interacts with their data. The design philosophy was simple: Motorsport-inspired, dark-themed, and data-dense.
Using modern web technologies, the dashboard features interactive charts (Plotly.js/Recharts) that allow riders to overlay different laps. You can hover over a specific corner on the track map and instantly see your lean angle, speed, and braking heatmaps compared to your best lap.
Conclusion
PulseCraft represents a perfect intersection of embedded systems, data engineering, and frontend design. It takes raw physical forces—gravity, acceleration, and friction—and turns them into pixels on a screen that help a rider shave seconds off their lap time.
Whether it's for academic evaluation or prototyping for real-world deployment, RAPTOR proves that professional-grade telemetry doesn't have to be limited to factory racing teams.