TyrePulse AI is an AI-powered motorsport analytics platform designed to help racing teams understand, predict, and optimize tyre performance and degradation throughout a race.
The platform combines race telemetry, tyre conditions, machine learning, and interactive analytics to estimate tyre wear, predict remaining tyre life, detect abnormal degradation, and provide data-driven insights for race strategy.
In motorsport, tyre performance can determine the outcome of a race.
Tyres continuously change due to:
- 🌡️ Temperature
- 🏎️ Speed
- 🏁 Track conditions
- 🛞 Tyre compound
- ⛽ Fuel load
- 🧑
✈️ Driving style - 🌧️ Weather
- 🔄 Number of laps
- ⚙️ Vehicle setup
A small degradation trend can eventually lead to significant lap-time loss.
TyrePulse AI turns this raw telemetry into actionable intelligence.
Instead of simply showing what happened, the system aims to answer:
"What is happening to the tyre, what will happen next, and when should we react?"
TyrePulse AI focuses on four major areas:
Track tyre performance and degradation in real time.
Estimate future tyre wear and remaining useful tyre life.
Identify unusual degradation patterns and potential tyre-performance issues.
Provide data-driven insights that can support pit-stop and race-strategy decisions.
Machine learning models analyze telemetry and tyre-related parameters to estimate the current and future degradation level.
Example:
Current Tyre Wear: 61%
Estimated Remaining: 14 laps
Predicted Degradation: High
Visualize how tyre performance changes throughout a stint.
The platform can analyze:
- Lap number
- Lap time
- Tyre age
- Tyre compound
- Temperature
- Degradation rate
- Performance drop
This helps identify when a tyre is entering its critical degradation phase.
Tyre temperature is an important indicator of tyre performance.
TyrePulse AI can monitor:
- Surface temperature
- Core temperature
- Temperature trends
- Overheating conditions
- Under-temperature conditions
The system can highlight situations where temperature may negatively affect performance.
Instead of requiring engineers to manually interpret every telemetry signal, TyrePulse AI converts the data into understandable insights.
Example:
⚠️ High Degradation Detected
Tyre degradation has increased by 18%
over the last 5 laps.
Estimated performance loss:
+0.24 seconds/lap
Recommendation:
Consider a pit stop within the next 3–5 laps.
TyrePulse AI can support strategic decision-making by analyzing:
- Current tyre condition
- Remaining tyre life
- Expected degradation
- Lap-time loss
- Tyre compound
- Pit-stop cost
- Race position
The goal is to help teams evaluate potential strategy decisions before committing to them.
The platform provides a centralized dashboard for viewing important race information.
Possible dashboard sections include:
- Current lap
- Race position
- Gap to competitors
- Current stint
- Current tyre compound
- Tyre health
- Wear percentage
- Degradation rate
- Remaining tyre life
- Temperature
- Lap-time trend
- Sector performance
- Pace degradation
- Performance delta
- Recommended pit window
- Tyre-change suggestions
- Risk alerts
- Strategy insights
┌─────────────────────┐
│ Motorsport Data │
│ / Telemetry │
└──────────┬──────────┘
│
▼
┌─────────────────────┐
│ Data Processing │
│ & Feature Engine │
└──────────┬──────────┘
│
▼
┌─────────────────────┐
│ ML Prediction │
│ Engine │
└──────────┬──────────┘
│
┌─────────────┴─────────────┐
▼ ▼
┌─────────────────┐ ┌─────────────────┐
│ Tyre Prediction │ │ Anomaly / Risk │
│ Engine │ │ Detection │
└────────┬────────┘ └────────┬────────┘
│ │
└─────────────┬─────────────┘
▼
┌─────────────────────┐
│ Strategy Intelligence│
└──────────┬──────────┘
│
▼
┌─────────────────────┐
│ Motorsport Dashboard│
└─────────────────────┘
TyrePulse AI follows a typical machine-learning pipeline:
Raw Telemetry
↓
Data Cleaning
↓
Feature Engineering
↓
Feature Selection
↓
Model Training
↓
Model Validation
↓
Prediction
↓
Strategy Intelligence
The prediction system can use features such as:
lap_number
tyre_age
tyre_compound
lap_time
speed
acceleration
braking
track_temperature
air_temperature
tyre_temperature
fuel_load
sector_times
previous_lap_time
degradation_rate
Depending on the available dataset, the system can predict:
Estimated percentage of tyre degradation.
Estimated number of competitive laps remaining.
Expected lap-time loss caused by tyre degradation.
Classification such as:
LOW
MEDIUM
HIGH
CRITICAL
TyrePulse AI can support multiple machine-learning approaches.
Used for predicting continuous values such as:
- Tyre wear %
- Remaining laps
- Expected lap time
- Degradation rate
Potential models:
- Linear Regression
- Random Forest
- Gradient Boosting
- XGBoost
- LightGBM
- Neural Networks
Used for identifying tyre conditions:
Healthy
Moderate Degradation
High Degradation
Critical
Telemetry is inherently sequential, so time-series approaches can be used to understand how tyre performance evolves over a stint.
Potential approaches:
- LSTM
- GRU
- Temporal models
- Rolling-window regression
- Sequence-based neural networks
TyrePulse AI is designed with real-time analysis in mind.
A typical flow is:
Telemetry
↓
Incoming Data
↓
Feature Extraction
↓
ML Prediction
↓
Tyre Health Update
↓
Risk Detection
↓
Strategy Recommendation
↓
Dashboard
This allows the platform to continuously update its understanding of tyre performance during a race.
TyrePulse AI can generate alerts when important conditions are detected.
Example:
⚠️ TYRE DEGRADATION
Degradation rate increased significantly.
Current degradation:
0.082 sec/lap
Previous:
0.051 sec/lap
Risk:
HIGH
Other possible alerts:
- 🔥 Tyre overheating
- ❄️ Tyre under-temperature
- 📉 Rapid degradation
- 🛞 Tyre approaching end-of-life
⚠️ Unexpected performance drop- 🏁 Recommended pit window
One of the platform's key capabilities can be evaluating hypothetical race scenarios.
For example:
Strategy A
Current tyre → Continue 5 laps → Pit
Expected result:
+2.4 sec
Strategy B
Pit immediately → Fresh tyres
Expected result:
+1.1 sec
This allows race engineers to compare potential decisions using predicted tyre behavior.
The dashboard is designed to provide a race engineer with a quick overview of tyre and race conditions.
Example:
┌──────────────────────────────────────────────┐
│ TYREPULSE AI │
├──────────────────────────────────────────────┤
│ LAP 42/70 P3 GAP +2.41s │
├──────────────────────────────────────────────┤
│ TYRE SOFT │
│ TYRE AGE 18 LAPS │
│ TYRE HEALTH 63% │
│ DEGRADATION HIGH │
│ REMAINING ~8 LAPS │
├──────────────────────────────────────────────┤
│ LAP TIME TREND │
│ │
│ 1:32 ────────╲ │
│ 1:33 ╲──────╲ │
│ 1:34 ╲────── │
│ │
├──────────────────────────────────────────────┤
│ 🤖 AI STRATEGY │
│ │
│ Recommended pit window: │
│ LAP 47–50 │
│ │
│ Confidence: 87% │
└──────────────────────────────────────────────┘
The exact stack can evolve with the implementation, but the project is designed around a modern AI/data architecture.
- Python
- FastAPI
- REST APIs
- Python
- Pandas
- NumPy
- Scikit-learn
- XGBoost / Gradient Boosting
- PyTorch or TensorFlow
- React
- TypeScript
- Vite
- Interactive data visualizations
- Plotly
- Recharts
- Custom dashboard components
Potential options:
- PostgreSQL
- TimescaleDB
- SQLite for development
- Docker
- Git
- GitHub
TyrePulse-AI/
│
├── backend/
│ ├── app/
│ │ ├── api/
│ │ ├── models/
│ │ ├── services/
│ │ ├── ml/
│ │ └── main.py
│ │
│ ├── requirements.txt
│ └── Dockerfile
│
├── frontend/
│ ├── src/
│ │ ├── components/
│ │ ├── pages/
│ │ ├── charts/
│ │ ├── services/
│ │ └── App.tsx
│ │
│ ├── package.json
│ └── Dockerfile
│
├── data/
│ ├── raw/
│ ├── processed/
│ └── samples/
│
├── notebooks/
│ ├── exploration.ipynb
│ ├── feature_engineering.ipynb
│ └── model_training.ipynb
│
├── models/
│ └── trained_models/
│
├── docker-compose.yml
├── README.md
└── .gitignore
git clone https://github.com/your-username/tyrepulse-ai.git
cd tyrepulse-aiCreate a virtual environment:
python -m venv .venvActivate it on Windows:
.venv\Scripts\activateLinux/macOS:
source .venv/bin/activateInstall dependencies:
pip install -r backend/requirements.txtRun the backend:
uvicorn backend.app.main:app --reloadBackend:
http://localhost:8000
Navigate to the frontend:
cd frontendInstall dependencies:
npm installStart development server:
npm run devThe frontend will typically be available at:
http://localhost:5173
Build and start the application:
docker compose up --buildStop the containers:
docker compose downPOST /api/tyre/predictExample request:
{
"lap": 42,
"tyre_age": 18,
"compound": "soft",
"lap_time": 92.41,
"tyre_temperature": 104.5,
"track_temperature": 42.1
}Example response:
{
"wear_percentage": 63.4,
"remaining_laps": 8,
"degradation_rate": 0.082,
"risk_level": "HIGH",
"confidence": 0.87
}Motorsport decisions can be highly sensitive to data quality.
TyrePulse AI should therefore consider:
- Missing telemetry
- Sensor noise
- Outliers
- Data drift
- Model confidence
- Track-specific behavior
- Different tyre compounds
- Changing weather conditions
Predictions should be presented as decision-support intelligence, not as guaranteed race outcomes.
- Project architecture
- Telemetry ingestion
- Data cleaning
- Feature engineering
- Exploratory data analysis
- Baseline tyre-wear model
- Degradation prediction
- Remaining-life prediction
- Model evaluation
- Confidence scoring
- Anomaly detection
- Tyre health scoring
- Pit-window prediction
- Strategy simulation
- AI-generated race insights
- Live telemetry dashboard
- Tyre degradation graphs
- Driver performance analysis
- Strategy comparison
- Alert system
- Multi-driver comparison
- Weather-aware predictions
- Track-specific models
- Online model adaptation
- Predictive race strategy
- Digital race engineer
TyrePulse AI can eventually be applied to:
- 🏎️ Formula racing
- 🏁 Endurance racing
- 🏍️ Motorcycle racing
- 🚗 GT racing
- 🏆 Karting
- 🧪 Motorsport research
- 🎮 Racing simulations
The long-term vision of TyrePulse AI is to evolve from a tyre-monitoring system into an AI race engineer.
The system should continuously understand:
What is happening?
↓
Why is it happening?
↓
What will happen next?
↓
What should the team do?
Ultimately:
TyrePulse AI turns tyre telemetry into race-winning intelligence.
Contributions are welcome.
- Fork the repository
- Create a feature branch
git checkout -b feature/new-feature- Commit your changes
git commit -m "Add new feature"- Push the branch
git push origin feature/new-feature- Open a Pull Request
This project is currently intended for educational, research, portfolio, and experimental motorsport analytics purposes.
Add an appropriate open-source license when the project's licensing terms are finalized.
Lidiya
Aspiring Data Scientist & ML Enthusiast
Interested in:
- Machine Learning
- Data Science
- AI Engineering
- Data Analytics
- Motorsport Intelligence
- Full-Stack Development
If you find TyrePulse AI interesting, consider giving the repository a ⭐ on GitHub.
Sense the tyre. Predict the degradation. Optimize the race.
From telemetry → to prediction → to strategy.