"APEX" - Adaptive Performance Engine for eXcellence

APEX turns raw race telemetry into real-time. Physics-driven AI shows where time is lost, how to fix it, and predicts performance changes. F1-level analytics for every driver, any track.

Comment

Inspiration

Motorsport generates huge amounts of telemetry — speed, throttle, braking, GPS paths, gears, and acceleration data.
Teams know this data is powerful, but extracting insights manually is slow and difficult.

We wanted to create a tool that transforms raw Toyota GR Cup telemetry into clear, real-time racing intelligence.
This led to the creation of APEX – Adaptive Performance Engine for Excellence.

What it does

APEX is an end-to-end racing analytics engine that takes GR Cup telemetry and produces:

  • Real-time lap analysis
  • Racing line reconstruction with GPS smoothing
  • Braking zone and throttle efficiency mapping
  • Lap time prediction using ML models
  • Sector-by-sector performance scoring
  • Driver behavior and consistency insights
  • A clean, interactive React dashboard to explore results

Upload telemetry → APEX analyzes it instantly.

How we built it

Frontend:

  • React + TypeScript for a fast, modular dashboard
  • Visualizations for racing lines, telemetry graphs, and sector breakdowns

Backend:

  • FastAPI powering low-latency inference and data processing
  • 15+ structured API endpoints
  • Custom pipelines for cleaning, parsing, and analyzing telemetry

Machine Learning & Analysis:

  • Algorithms for GPS smoothing, racing line reconstruction, braking/throttle detection
  • Lap time prediction models trained on all 7 Toyota GR Cup tracks
  • Driver consistency scoring and corner-entry/exit performance metrics

Architecture Highlights:

  • Data ingestion → preprocessing → ML inference → analysis → visualization
  • Modular design so each component can be expanded independently

Challenges we ran into

  • Cleaning and smoothing noisy GPS data from multiple tracks
  • Designing ML models that generalize across all 7 tracks
  • Creating fast inference without requiring GPU hardware
  • Managing millions of telemetry points efficiently
  • Building a polished full-stack system under hackathon time limits

Accomplishments that we're proud of

  • Complete full-stack racing intelligence platform built from scratch
  • Accurate racing line visualizations for multiple tracks
  • Real-time inference with sub-100ms response
  • Clear, driver-friendly insights that actually help improve lap times
  • Handling large telemetry datasets smoothly and reliably

What we learned

  • The complexities of real racing telemetry and motorsport analytics
  • How to design robust preprocessing pipelines for noisy signals
  • Building fast and scalable APIs for data-heavy workloads
  • Frontend techniques for visualizing racing data efficiently
  • Balancing ML accuracy with real-time performance

What's next for "APEX" - Adaptive Performance Engine for Excellence

  • Real-time live telemetry streaming during races
  • Personalized driver profiles and habit analytics
  • Reinforcement-learning based driving suggestions
  • Mobile version for engineers and pit crews
  • Support for additional racing series beyond GR Cup

APEX is built to push the boundaries of racing intelligence — making drivers faster, engineers smarter, and data more powerful than ever.

Share this project:

Updates