The vision
Imagine a pit wall that sees not just tyre temperatures, but whether the driver is still in their window.
Imagine a young driver who learns to come back from a mistake in one lap instead of five.
Imagine racing that takes the driver's mental load as seriously as tyre wear.
Imagine pilots and operators who find their best state faster, and hold it longer.
That's what we're researching. Now.
01The missing channel
Every car is measured to the millisecond. The driver isn't.
Choose a driver state and watch the same car, on the same track, find or lose time. As on a timing screen, each sector turns purple, green, yellow or red when the car completes it. The tyres stay in their window. The driver doesn't.
Keep scrolling to pan across all channels, or swipe →
Sector times · vs. target
| S1 | S2 | S3 | Lap | |
|---|---|---|---|---|
| Time | ||||
| Δ target |
Tyres · °C
Braking points
Heart rate · lap avg
Medium to high, as expected under load.
From one fast lap to the state behind it
The scene shows one lap per driver state. The three steps below show what we want to find out from it: whether one state is behind the fastest laps, how long a driver can hold it, and how quickly they find their way back.
One pattern we keep seeing in our fastest laps
In our recordings so far, the fastest laps came with LF and HF power rising and falling together. We call this pattern fully engaged: it may reflect good stress combined with a strong ability to regulate. Whether it holds across drivers is hypothesis H3.
How long does the driver stay there?
If the state is real, the lap time is its result. What a team wants to know is how much of a stint the driver spends in it, and what pushes them out: a mistake, traffic, a long stint.
How fast do they get back in?
Every driver drops out. How fast they return may say more than how far they drop, and the way back may be trainable in the simulator.
That is what Season Zero measures, per driver and next to the lap times:
- 01Time in state
- 02What pushes a driver out
- 03Time to return
02From heartbeat to frequency bands
From every heartbeat to continuous autonomic dynamics.
Each R-peak marks a beat. The time between beats changes constantly. Different physiological control mechanisms leave characteristic rhythms in that beat-to-beat variation. CADA follows two of these frequency bands continuously while the session is running.
Keep scrolling to follow the signal, or swipe →
03During the run
Every mistake has a prologue.
Our hypothesis: the frequency bands show a loss of concentration before the lap time does. You rarely notice it yourself when stress or fatigue wear your concentration down; flow and a tired autopilot can feel the same from the inside. In this example the driver tires over one and a half laps: HF power climbs, LF power drops, and the two bands stop moving together. Only then does the lock-up show in the lap time. Seeing the prologue early is what we are researching.
Keep scrolling to follow the stint, or swipe →
04At the pit wall
Decide with the driver in the data.
Two frequency bands go in. A model learned from each driver's own sessions turns them into one state line with a personal window around it. The pit wall sees when someone slowly drifts out.
Personal window
Learned from the driver's own sessions. Individual, not a universal threshold.
Pit wall note · Lap 44
Driver B has been drifting out of the personal window for three laps.
Options: radio check-in, adjust the stint, driver change at the next stop.
Chestra informs. The team decides.
Keep scrolling to see the whole screen, or swipe →
05In the simulator
Analyse and practise recovery after stress triggers.
Every trigger causes a stress response. What matters is how fast the driver comes back. Repeat the situation in the simulator, see the response, and practise the way back into the window, session by session.
Keep scrolling to see the chart, or swipe →
06Why this signal
There is more in a heartbeat than a number.
The heart is steered by several control loops at once. Some act within seconds and follow breathing, others act over tens of seconds and follow blood-pressure regulation. Each leaves its own rhythm in the beat-to-beat signal.
Most tools use HRV to give one readiness number, calculated from the intervals between beats over a time window. That simple method cannot see how LF and HF interact, and it does not work live. CADA resolves both frequency bands live and shows how they interact from moment to moment. We believe that interplay holds information about regulation under load that no single metric shows.
How CADA differs from classic HRV analysis
Classic frequency-domain HRV takes a time window, typically several minutes long, and summarises it into one value per band. CADA is designed to follow changes in LF and HF power continuously over time, while the session is running.
No proprietary hardware: CADA works with the beat-to-beat signal of a consumer ECG chest strap.
Indicative assessment based on published research and typical devices, as of 2026. Individual products may have solved specific limitations, for example newer EEG systems with better motion handling. Three dots = strongest. N/A = does not apply, the sample is analysed in a lab. CADA's sensor signal is robust to motion, sweat and heat; breathing and physical load influence the bands, which is why we model them. Background: artifacts in electrodermal activity, artifacts in wearable EEG, respiration and HF heart rate variability.
07The data
A new kind of training data for AI.
AI has advanced fastest where large, well-labelled datasets exist. Continuous autonomic data from people performing under real load is still rare. CADA produces this signal, and Season Zero synchronises it with lap times, braking, mistakes, traffic and driver feedback. We bring the method, thirty years of autonomic research and proprietary data. Season Zero puts today's AI to work on it and trains the first models.
- 01Heartbeatconsumer ECG chest strap
- 02CADALF and HF power, live
- 03Synchronisedwith driving telemetry
- 04Labelled datasetmistakes, traffic, feedback
- 05Personal modelswindows, prologues, recovery
08The research
The signal is there. Now we decode it.
Our research so far has given us six hypotheses. Now they need to be validated with a larger sample and across the full range of driver skill, from rookies to professionals. Each one can be tested in a sim racing lab within weeks.
A personal window
Every driver performs best in an individual range of raised but regulated activation.
Recovery beats peaks
How fast a driver returns after stress says more than how high the spike goes.
Co-activation shows engagement
When LF and HF power move together, the driver is engaged and flexible.
Mistakes have a prologue
Patterns shift minutes before a series of errors, long enough to act on.
Recovery is trainable
Repeated triggers and breathing work shorten the way back into the window.
Ready to drive today?
The state before a session shapes the session.
How we test
Pre-registered protocol, errors labelled automatically from sim telemetry, every session with informed consent.
Why this research is worth backing
Sim racing lab, consumer chest strap, full telemetry. Weeks, not years.
Recovery training for academies and sim racers, from one confirmed hypothesis.
A dataset nobody has, models trained on it, and a platform for teams, sim brands and operators.
09Faster and safer
One physiology. Two questions: performance and wellbeing.
The bands that find a tenth may also show mental load building over a long stint, how well a driver recovers between sessions, or when a young driver needs support. Better racing and healthier racing point the same way.
WHAT WE WILL NOT DO
- No stress score out of a black box.
- No diagnosis. Chestra is a research and training technology, not a medical device.
- No warnings that distract a driver mid-race. Training and awareness first.
- No data without explicit, informed consent.
10Season Zero
Back the first season of human telemetry.
Six months to turn a working signal into the first models of driver state. Built with sim racing brands, teams and their drivers, who invest, bring data, or both.
Funding partners
Invest in Season Zero
Sim racing brands, teams and individuals who fund the programme.
You get first access to findings and models, a say in what we build, and partnership terms on the platform.
Data partners
Bring drivers and sessions
Teams, academies and hardware makers with drivers and rigs.
You get driver-specific insight, co-branded research, and early integration with your hardware or software.
Drivers
Race with Chestra
Sim racers from our partners' communities.
You get your autonomic response live on screen, free during Season Zero.
Sim racing is the laboratory, not the limit.
- Sim racingNow · Season Zero
- MotorsportNext
- AviationLater
- Mission-critical operatorsLater
The core platform (autonomic telemetry, context data and personal models) is built to transfer wherever people operate under sustained load.
11Science & team
Thirty years of autonomic research, now on track.
Chestra is a brand of FINTEC GmbH, a spin-off of the University of Trier's Center for Neuropsychological Research. The research continues at the non-profit Center for Neuroscientific Research (ZNF gGmbH).
Dr. Arne Wittling
Scientist & Co-FounderManaging Director of FINTEC GmbH and the non-profit ZNF in Trier. Develops bioalgorithms for autonomic regulation; co-inventor of a patented HRV analysis method.
Aaron Siebierski
Lab Rat & Co-FounderTurns scientific methods into products. Background in digital strategy for pharmaceutical R&D and advisory work in HealthTech and human performance.
Granted European patent
Method and device for determining heart rate variability: from heartbeats, for example from a chest strap, to LF and HF power, with breathing-aware correction of irregular beats.
CADA · patent pending
Continuous Autonomic Dynamics Analysis.
- Wittling et al. (2010) Complex demodulation of cardiorespiratory activity and cardiovagal recovery after exertion.
- Jarczok, …, Thayer (2013) Artifact correction in R-R interval analysis: man vs. machine.
- Wekenborg, …, Kirschbaum (2019) Longitudinal association of reduced vagal tone with burnout.
- Lueken, Schweiger & Wittling (2005) Autonomic and cortisol responses to a cognitive activation task.
Every team measures the car to the millisecond. Be the first to measure the driver.
Thank you, your message is on its way. We usually reply within two working days.
Your message could not be sent right now. Copy it and send it to ams@chestra.eu, or open it in your email app.
PREFER EMAIL?
ams@chestra.eu
What happens next:
- A 30-minute call about your drivers, rigs and goals.
- A short pilot outline for Season Zero.