01 / Decades of data
The body has always
been part of the mission.
From the early space programmes to the International Space Station, measuring the human body has been central to understanding life in space. Heart activity, breathing, temperature and movement leave a record over time.
Electrodes, temperature probes and blood-pressure sensors.
Watches, headbands and sensor vests track physiology and activity.
Could historical signals help anticipate an individual’s needs?
Historical context: NASA · telemetry and ISS wearables.
02 / Learning from signals
A measurement is a moment.
A signal tells a longer story.
A 1D time series is a sequence of measurements along time. ARIA’s predictive direction starts here: learning patterns across signals, then relating those patterns to changes in health.
Build a record over time.
Combine repeated physiological measurements with their context: sleep, exercise and the individual’s baseline. A time series captures variation that one isolated value cannot describe.
What would make prediction credible?
Access to suitable historical data, reliable health outcomes and testing on people or missions excluded from training. An unusual signal is not automatically a health problem. Earlier warning, false alarms and uncertainty would all need to be evaluated.
03 / Beyond self-report
Some symptoms
never reach the report.
In 2001, an Institute of Medicine review documented pressures that could discourage astronauts from reporting health concerns: fear of losing flight opportunities and a culture that valued stoicism.
That gives ARIA a second motivation: complement what a person reports with changes measured over time. A discrepancy could prompt a conversation or assessment, without treating a sensor as a judge of honesty.
Safe Passage · Institute of Medicine, 2001 ↗
The ARIA proposal
Give the crew another way
to recognise emerging needs.
The hackathon pitch proposed an onboard assistant combining physiological information and mission constraints. Nutrition was the worked example; learning from time series gives the project its broader predictive direction.
Hackathon concept · predictive performance has not been demonstrated.
04 / A route to development
From concept
to a research programme.
The pitch identified organisations spanning space medicine, physiological sensing and habitat engineering. Each could address a different step: access suitable data, test the model, then explore integration.
Potential collaborators and support routes · no partnership, data access or funding secured is claimed.
Start with the data.
MEDES / EveryWear
EveryWear already collects medical and scientific information aboard the ISS; CNES describes recording, preprocessing and synchronising data in flight.
For ARIA: explore a research collaboration and compatibility with an existing collection workflow. Access to historical records would need a separate agreement.
CNES · EveryWear ↗DLR / :envihab
AGBRESA studied 24 participants through 60 days of head-down bed rest, with baseline and recovery measurements: a concrete lead for investigating physiological change.
For ARIA: establish which records and outcomes could support a first study. These are ground-analogue data; validation in astronauts would remain a later step.
AGBRESA · study protocol ↗What access is actually documented?
The paper states that its supporting raw data can be made available by the authors. That does not establish an open download of every sensor stream, or permission to train ARIA on the full study archive.
IBMP / SIRIUS
IBMP’s SIRIUS programme studies crews in prolonged isolation. It provides a research lead for the human context around physiological measurements.
For ARIA: investigate relevant studies and current access conditions. The historical collaboration does not establish an available dataset or a current partnership.
ESA · SIRIUS-21 research ↗Connect the model to real systems.
Equivital + Hexoskin
Equivital’s eq02+ records cardiac, respiratory, temperature and movement signals. Hexoskin’s Astroskin has flown on the ISS and has been selected for adaptation to Gateway.
For ARIA: assess signal quality, data interfaces and a prospective validation setup. Sensor capability does not establish access to previous users’ recordings.
Thales Alenia Space
Its work on Gateway’s HALO pressurised structure makes it a relevant industrial lead for understanding the constraints of crewed habitats.
For ARIA: explore integration requirements once feasibility is established. The pitch identified a potential industrial partner, not an agreed deployment.
Thales Alenia Space · HALO ↗Fund the next piece of evidence.
The first milestone would be a feasibility study: secure an appropriate dataset, define one prediction target and measure performance on participants excluded from training.
- Scope the studyExplore technical guidance and funding orientation through Connect by CNES ↗.
- Support a prototypeAssess an ESA BIC incubation route ↗, subject to company, territorial and programme eligibility.
- Validate before integrationUse the feasibility results to scope a research collaboration, followed by an industrial pilot if the evidence supports it.
Published funding example · ESA BIC Sud France
The programme currently lists a €25,000 grant, alongside business and technical support. Its criteria include a company under five years old, a space connection, eligibility for at least €25,000 in Bpifrance or regional public support, and location within its territory with incubator support.
ARIA’s eligibility and the applicable regional route remain to be established. This is a potential application pathway, not awarded funding. Programme information checked 24 September 2026.
ActInSpace Grenoble 2026
Maxime Raffort · Colin Schotte
The academic foundations