[TIPA] 2026 TIPA Global Cooperative R&D Project (pre-planning stage) — multimodal biosignal-based pain prediction and clinical decision support system

Overview

Selected for the 2026 Global Collaborative R&D Program organized by the Korea Technology and Information Promotion Agency for SMEs (TIPA), with PainCube, StellarCube's pain prediction platform. The project is supported by up to KRW 1.5 billion in government funding over three years.

Consortium and roles

StellarCube formed a consortium with Johns Hopkins University (JHU) in the United States and Korea University Anam Hospital. Professor SungMin Cho serves as principal investigator at Johns Hopkins University and Professor Hee Jung Kim at Korea University Anam Hospital. StellarCube is responsible for AI algorithm development, clinical data analysis and platform enhancement, while the two hospitals handle clinical data collection, research protocol design and validation of the prediction model's performance.

Technology

AI jointly analyses patient biosignals and clinical data - ECG, PPG, HRV and SpO2 - to quantify pain and predict the likelihood of pain occurring in advance. The model works on high-resolution signals (ECG at 240 Hz or above, PPG and ABP at 120 Hz or above) together with respiratory signals, and is trained using federated learning so that data from the two countries is never pooled in one place.

Targets

The algorithm is being refined towards predicting the likelihood of pain in ICU patients 30 to 60 minutes in advance with AUROC of 0.90 or higher. Prediction performance and real-world applicability will then be validated in stages against clinical data from actual intensive care settings, with global SaMD (Software as a Medical Device) approval and entry into overseas medical markets to follow.