The hospital bed, once a simple support for patient admission, has now become a key element for continuous monitoring and patient care, reducing the need for movement and increasing the effectiveness of treatment. In this innovative direction, the company Malvestio S.p.A. has launched a project to integrate non-invasive detection systems into its beds, capable of automatically identifying the patient’s posture and condition. This approach provides a competitive advantage over existing devices on the market, ensuring faster monitoring, less staff intervention, and the possibility of remote supervision. During the internship at the host company, a comparative evaluation was carried out between two prototype solutions for detecting the posture and condition of a lying patient. The systems were subjected to experimental tests using metrics such as accuracy, reliability, and robustness, and adopting processing techniques based on convolutional neural networks. The analysis highlighted strengths and weaknesses, providing insights for potential improvements. Finally, a feasibility study was conducted to assess industrial integration, taking into account the company’s production technologies and the regulatory constraints of the medical sector. The results confirm the technical feasibility of the analyzed solutions and outline a development path aimed at improving performance and reliability, with a view to future application in real clinical settings.
Il letto ospedaliero, da semplice supporto per il ricovero, è oggi un elemento chiave per il monitoraggio continuo e l’assistenza al paziente, riducendo la necessità di spostamenti e aumentando l’efficacia delle cure. In questa direzione innovativa, l’azienda Malvestio S.p.A. ha avviato un progetto per integrare nei propri letti sistemi di rilevamento non invasivi, capaci di identificare automaticamente postura e condizioni del paziente. Questo approccio offre un vantaggio competitivo rispetto ai dispositivi presenti sul mercato, garantendo un monitoraggio più rapido, minore intervento del personale e possibilità di supervisione a distanza. Durante il tirocinio presso l’azienda ospitante, è stata effettuata una valutazione comparativa di due soluzioni prototipali per il rilevamento della postura e delle condizioni del paziente disteso. I sistemi sono stati sottoposti a test sperimentali, utilizzando metriche quali precisione, affidabilità e robustezza, e adottando tecniche di elaborazione basate su reti neurali convoluzionali. L’analisi ha permesso di evidenziarne punti di forza e criticità, fornendo indicazioni per possibili miglioramenti. Infine, è stato realizzato uno studio di fattibilità per valutarne l’integrazione industriale, considerando le tecnologie produttive dell’azienda e i vincoli normativi del settore medicale. I risultati ottenuti confermano la fattibilità tecnica delle soluzioni analizzate e delineano un percorso di sviluppo volto a migliorarne prestazioni e affidabilità, in vista di una futura applicazione in contesti clinici reali.
Valutazione, confronto ed integrazione di soluzioni innovative per l’identificazione della postura del paziente sul letto ospedaliero
PALMISANO, ENRICO
2024/2025
Abstract
The hospital bed, once a simple support for patient admission, has now become a key element for continuous monitoring and patient care, reducing the need for movement and increasing the effectiveness of treatment. In this innovative direction, the company Malvestio S.p.A. has launched a project to integrate non-invasive detection systems into its beds, capable of automatically identifying the patient’s posture and condition. This approach provides a competitive advantage over existing devices on the market, ensuring faster monitoring, less staff intervention, and the possibility of remote supervision. During the internship at the host company, a comparative evaluation was carried out between two prototype solutions for detecting the posture and condition of a lying patient. The systems were subjected to experimental tests using metrics such as accuracy, reliability, and robustness, and adopting processing techniques based on convolutional neural networks. The analysis highlighted strengths and weaknesses, providing insights for potential improvements. Finally, a feasibility study was conducted to assess industrial integration, taking into account the company’s production technologies and the regulatory constraints of the medical sector. The results confirm the technical feasibility of the analyzed solutions and outline a development path aimed at improving performance and reliability, with a view to future application in real clinical settings.| File | Dimensione | Formato | |
|---|---|---|---|
|
Palmisano_Enrico.pdf
accesso aperto
Dimensione
6.57 MB
Formato
Adobe PDF
|
6.57 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/94385