Undae Science is built on deep sleep expertise with the ambition to transform clinical sleep management and drive meaningful change in population health. Sleep disturbances pose a growing societal burden, costing over 10 billion CHF annually in Switzerland. Current diagnostics rely on intrusive, single-night hospital exams that disrupt natural sleep, often leading to delayed or inaccurate diagnoses. This contributes to the chronicisation of sleep disorders and related conditions. Clinicians agree on the urgent need for an innovation that matches the precision of clinical gold standards while overcoming their limitations—needs unmet by current solutions. Our innovation, Noctae, reimagines the sleep medical station. It’s a user-friendly, contactless overmattress device that enables clinical-grade, multi-night sleep monitoring at home. Noctae is the result of years of cutting-edge research, combining a unique configuration of multiple embedded sensor types with a proprietary AI-powered SaaS platform that transforms raw mechanical signals into clinically validated sleep metrics, delivering accuracy on par with hospital gold standards, but in the comfort of home.

Undae Science SA was founded in November 2024, but the proof-of-concept (PoC) phase of the project “SleepSensor” began in April 2021. During this phase, five functional prototypes were developed and tested on nearly 100 participants, including both healthy volunteers and patients, using the clinical gold standard (polysomnography-PSG) for direct comparison. A machine learning-based classifier, trained on features extracted from our sensor data, achieved clinically suitable accuracy, demonstrating the feasibility of non-intrusive, AI-powered sleep monitoring.

This three-years PoC was conducted at the University of Lausanne (UNIL) with the support of both the InnoTrek grant (CHF 100 K) and Innosuisse (Project No. 101.451 IP-LS, total funding: CHF 393K), without implementation partners. Following its successful completion, all R&D activities were transferred to Undae Science SA upon its creation.

The company is now entering the next development phase, supported again by Innosuisse (Project No. 122.997 IP-LS, total funding: CHF 415 K), in collaboration with EPFL Lausanne, Universitätsspital Basel, and HEIG-VD Yverdon as accademic partners. This stage focuses on optimizing hardware signal quality to better capture patient-specific features and advancing our AI models to address inter-patient variability—critical steps toward delivering clinical-grade performance. The current Innosuisse project (122.997 IP-LS) runs until October 2026 but the final hardware prototype for industrialization is expected to be available by the end of 2025.