Title: Ergonomic Analysis of Integrated Bed Measurements: Towards Smart Sleep Systems (Ergonomische analyse van geïntegreerde bedmetingen: op weg naar slimme slaapsystemen)
Other Titles: Ergonomic Analysis of Integrated Bed Measurements: Towards Smart Sleep Systems
Authors: Verhaert, Vincent
Issue Date: 12-Dec-2011
Abstract: Ergonomic aspects of sleep are gaining importance in order to develop individually optimized sleep environments promoting sleep initiation and maintenance. The research domain of sleep ergonomics is relatively young and requires bridging the expertise of sleep physiologists, psychologists, ergonomists and engineers; a challenge in itself. This PhD dissertation focuses on one of the main environmental components the sleeping body interacts with, namely the sleep system (i.e. the combination of mattress, bed base and pillow). The presented work contributes to the state-of-the-art in bed design by integrating control algorithms that allow bed properties to be actively changed during sleep in order to continuously optimize body support. Furthermore, the developed tools were validated during overnight experiments in a dedicated sleep laboratory.First, the effect of spinal alignment on objective and subjective sleep parameters was studied. Sleep systems with adjustable comfort zones were used in order to alter spinal alignment independently from other confounding variables. During overnight studies two well-defined experimental conditions were compared: a sagging support and a personalized support. Since spinal alignment is dependent on the adopted posture, the amount of time spent in each sleep posture was considered as an influencing factor during the results analysis. Secondly, an unobtrusive technique was introduced for the assessment of motor patterns during sleep. An algorithm was developed to automatically detect body movements and sleep postures based on integrated mattress indentation measurements. Validation was performed during an overnight study by means of combined polysomnographic and video recordings.Thirdly, the problem of evaluating spine shape without disturbing sleep was dealt with. Therefore, a generic surface model representing human body shape was developed. Based on specific anthropometric information (derived from measured body contours), personalized human models could be generated. The combination of these surface models with an inner skeleton model allowed simulating distinct sleep postures. Furthermore, an algorithm was implemented to automatically fit the person-specific model in the measured mattress indentation in order to estimate spine shape. Validation of the estimated spine shapes was performed by means of simultaneous 3-D scans of the back surface in lateral positions.Fourthly, the previous two developments were integrated in a feedback control system aimed at optimizing spinal alignment by means of controlling the mechanical properties of eight comfort zones continuously throughout the night. The performance of the developed control system was tested both in laboratory conditions and during overnight sleep experiments. Finally, the effect of such an actively controlled sleep system on sleep was studied. Therefore, the actively controlled sleep system was compared to a standard (static) sleep system. Polysomnographic measurements were performed and questionnaires were completed to derive both objective and subjective outcome measures.The results presented in this work indicated that spinal alignment indeed affects sleep. Primarily, it was shown that habitual sleep postures determine to what extent subjects experience a negative effect while sleeping on a sagging sleep system. More specifically, subjects preferring prone and lateral sleep posturesdemonstrated impaired sleep on a sagging support whereas no such effect was noted for supine sleepers.Furthermore, the results proved the feasibility of integrated mattress indentation measurements to detect body movements and adopted sleep postures unobtrusively. Movement detection based on mattress indentation measurements showed to be more sensitive than movement detection based on standardactigraphic recordings. In addition, a high accuracy was achieved for the automatic classification of indentation images into supine, left lateral, right lateral and prone positions based on five image features.In order to evaluate spinal alignment, knowledge of mattress indentation alone was not sufficient. However, combined with personalized body shape models, proper spine shape estimation was shown to be feasible. The actively controlled sleep system resulted in significant improvements in terms of spinal alignment compared to a static sleep system. In addition, the ability to change mattress characteristics dynamically was successfully used to detect intermediate postures (between lateral and prone), providing a proof of concept for the added value of dynamic classification algorithms. Results of overnight experiments demonstrated the stability of the control when subjects were not bound to prescribed postures and showed a positive effect on subjectively perceived sleep.The multidisciplinary work presented in this thesis, was merely a first step towards creating a smart sleep environment. Several tools were developed and validated for ergonomic evaluation and control during sleep. Although a clear focus was put on the sleep system, a similar approach can be followed for othercomponents of the sleep environment.
Publication status: published
KU Leuven publication type: TH
Appears in Collections:Biomechanics Section
Spin-off & Innovation

Files in This Item:
File Status SizeFormat
PhD Vincent Verhaert.pdf Published 9635KbAdobe PDFView/Open


All items in Lirias are protected by copyright, with all rights reserved.