Detekce míry diskomfortu v paliativní péči

Abstract

This master's thesis focuses on the design of a fuzzy expert system for the contactless detection of discomfort in palliative care patients. The MATLAB-based model processes objective data from a smart sensory bed pad (heart and respiratory rate, restlessness) and fuses it with manually entered clinical parameters (pain, medication). The system was initially verified on a synthetic dataset and subsequently tested on real clinical data. Results demonstrated that in ambiguous situations, the model exhibits high sensitivity and prefers to report potential discomfort, thereby safely preventing the oversight of patient deterioration. The thesis confirms the suitability of fuzzy logic for processing biological data; however, further expansion of the knowledge base is necessary for full clinical deployment in order to more accurately classify borderline conditions and minimize false alarms.