School of Engineering and Materials Science
Research Student Awards
PhD Thesis: Methods for Seated Posture Recognition
Author: CIAMPONE, Sandra
Supervisor(s): Peter Dabnichki
With changes in business practice and life style, increasingly people travel spending prolonged periods of time sitting on airplanes, trains, or cars. Hence sitting comfort has become a critical issue due to fitness and health implications. Focusing on airplanes and based on the idea that a ’responsive environment’ (environment that adapts to independently assessed passengers conditions and adjust itself to them) will provide high level of comfort, we investigated ways of performing a continuos real time physiological monitoring of the passengers. A smart seat is hence the interface structure to allow passengers to be monitored, together with external sensors, such as environmental temperature sensors, humidity sensors, etc. (those last ones are not subjects of the present work). The smart seat is the physiological monitoring system that analyses in real time individual passengers’ status, such as their body temperature, heart rate, posture, activity, etc. Seated posture recognition is the subject of this study. The aim of this study is to evaluate the relevant postures assumed on airplanes’ seats and to develop a robust method for their recognition. Moreover we wanted to develop a commercial prototype of a non-invasive passengers monitoring system on airplanes (i.e. a smart seat capable to detect different aspects of the passenger’s conditions like its body temperature, its activity, etc.). Few studies have been conducted in the past concerning seated posture recognition; all of them involved a high number of sensors for an accurate result. The novelty of this study is the overall simplicity of the system and the commercial feasibility; it is indeed based on the following requirements: the number of sensors involved must be as small as possible so to not interfere with other physiological monitoring modules integrated on the seat (such as temperature evaluation, hearth rate analysis, etc.), and to keep the weight and the cost relatively low. 4 5 The recognition was performed by using a combination of the following devices, depending on the analysis performed: pressure mats mounted on the seat, force plates, and eventually cameras for image analysis. After determining the postures to be classified we created a ’Posture Database’ containing information about the relevant parameters for each of those. Two different approaches to posture recognition were investigated: an analytical method which reconstructs the passenger’s posture by using a small number of inputs (solely form force plates data); and a pattern recognition method that makes guesses about the passenger posture by accessing to the database of classified postures. Both the approaches showed successful results. In this pioneering study we created a Posture Database, not found in literature, and developed methods for seated posture recognition that rely on the use of a limited number of sensor. Furtherly, a prototype of a smart seat capable to recognise the user’s posture was developed within the FP6 project “SEAT: Smart tEchnologies for stress free Air Travel” and it can lead to many exciting applications such as responsive environment and personal comfort settings in vehicles, automatic control of airbag deployment forces, ergonomics of furniture design, and biometric authentication for computer security.