Web Posted on: August 24, 1998
Remote Monitoring System to Measure Indoors Mobility and Transfer of the Elderly
M. Chan
INSERM CJF 94-06, 37 allées J. Guesdes, 31073 Toulouse, France, tel: 33 5 61 14 59 57
LAAS-CNRS, 7 avenue du Colonel Roche, 31077 Toulouse, France, tel: 33 5 61 33 69 51
H. Bocquet
INSERM CJF 94-06, 37 allées J. Guesdes, 31073 Toulouse, France, tel: 33 5 61 14 59 57
E. Campo
LAAS-CNRS, 7 avenue du Colonel Roche, 31077 Toulouse, France, tel: 33 5 61 33 69 51
3 IUT B, 1 place G. Brassens, 31730 Blagnac, France, tel: 33 5 62 74 75 59
J. Pous
INSERM CJF 94-06, 37 allées J. Guesdes, 31073 Toulouse, France, tel: 33 5 61 14 59 57
email: chan@laas.fr and campo@laas.fr
Abstract
The paper presents a multisensor system devoted to measuring the indoors mobility and transfer of the elderly. A prototype has been built in the laboratory and its components are shown. Some preliminary experiments and results are presented.
| Top |
1. Introduction
In the case of the elderly, dependency is defined as the recourse to a technological or human aid for the accomplishment of the activities of daily living. This dependency is secondary to the ageing process or chronic diseases. Its measurement relies on scales, grids or models. Human observers or caregivers are asked various series of questions on different disorders. The measures are so designed to provide a one dimensional aggregation of the different types of disorders into an overall and synthetic score based on the observers'answers. The occurrence or the frequency of the disorder is measured more than the severity and the duration of the disorder itself. Owing to the nature of observation by humans, subjectivity may come into play according to how the disorder is lived through by the carer or according to the latter's personality. And there may equally exist a risk of lack of adequate empirical data on the nature (Matteson et al.) and progress of these disorders with time (Wagner et al.); some periods of time are more difficult to observe, during the night for example. The duration of the disorder is often ignored as it is difficult to follow, an episode of behavioural disturbance may occur only once a week but lasts for 10 hours at a time (O'Leary et al.).
Therefore, we propose the development of a remote monitoring system capable of operating 24 hours a day, 7 days a week designed for caregivers and researchers to automatically and objectively quantify the functional ability of the elderly in terms of mobility and transfer within the premises. The person is not instrumented his intimacy is respected, and his daily life is not disturbed. His active participation is not required in the handling of the instrumentation, in the progress of the experiment or in the system routine use. It is part of a global system developed and evaluated in a multidisciplinary project to assist caregivers in reducing their burden in the care of the demented or non demented elderly at home or in long stay unit by warning them in case of abnormal events.
| Top |
2. Methods
2.1 Participants
The people enrolled in this study all volunteered. They simulated motion and transfer in a lab room used as a replica of an old person's flat.
2.2 Room and instrumentation
A lab room has been used to simulate the living quarters of an elderly person in a long stay unit including a main room used as a bedroom with one bed and a bathroom. A remote monitoring system also has been installed in this lab room (Chan et al., 1995 Chan et al., 1996) including a set of sensors, a communication network linking the sensors to a computer outside the room for follows up.
Thus any movement including displacement, going to bed, getting up, etc., in the lab room is monitored by the passive infrared (IR) sensors and by a pressure sensor located in the bed. The human body emits some electromagnetic radiation known as thermal radiation in the IR frequency range. It can be highlighted by transforming the carried power into electrical or thermal energy. It is then possible to detect its presence through use of the IR sensor (Turck 1988). The IR sensors selected exhibit a short response time and an information retention of about 0.5 s. Their detection sensitivity is 0.1 m/s. For a total floor surface about 20 sq.m with a main room and a bathroom, nine passive IR sensors are secured to the ceiling covering eight non-lapping adjacent areas for the main room and one zone for the bathroom as shown in Fig. 1. Any person entering or leaving the room is detected by an active IR sensor including a transmitter and receiver built in near the door. The pressure sensor in the bed is made of a synthetic foam with an internal resistance varying according to the pressure exerted. By tuning the threshold, it is possible to avoid the detection of small weights. The analog information is then transformed into binary data (bed occupied or not). The bed position or the passage of the person from one area to another in the room or bathroom triggers a change in the status of the bed pressure sensor or of the IR sensors. Information is directly transmitted to the computer via the communication network. An industrial type bus (RS485 link) is used from the digital input modules (Fig. 2.). These modules dialog on the bus at throughputs that can go up to 38400 bits/s. A gateway is used as a link to the computer through a serial port. The computer makes it possible, through C++ language software, to handle the computations of the person's mobility.
2.3 Mobility programming
The scenario selected is that of a mobile person in his living quarters simulated in the lab. The sensors fitted out in the room and the C++ language software allow to know whether the person
- entered the room and at what time;
- occupies the bed, in which case, at what time did he go to bed, get up, how long did he stay in bed
- is mobile in a narrow zone covered by one IR passive sensor to detect movements (case of a fall where the person is unable to get up again but can still move his arms and talk, or of a wandering in a small area)
- moves in the main room or bathroom, the system records the motion pattern, path followed and duration, and computes the mean speed rate
- left his room and at what time.
The information collected can be accessed:
- in real time for monitoring
- in differed time for the research needs and for the handling of data by the staff.
With respect to the scenarios simulated by volunteers, reducing the duration of experiment from some hours to fifteen minutes, the aim is to verify the feasibility of the system.
The following displacements were made : in a straight line with a precise goal, pacing or random motion without any precise goal, any dubious immobility in areas other than the bed, physical mobility within a narrow area, immobility on the floor (the person can move his arms however for example). All these events are recorded on PC. Following processing of data, a diagram or pattern of the displacements (the path covered, pattern, the duration, etc.,...) is obtained with its salient features as depicted in Fig. 3. showing the displacements in the room. This could correspond to the displacements of a normal person or of someone suffering from motor behavioural disorders. Fig. 3a. corresponds to a direct displacement while Fig. 3b. shows the pacing motion between three zones. The third displacement (Fig. 3c.) is purely random.
We have simulated the occupation of the room by several scenarios of 15 minutes including a bed occupation, some immobility for one minute in any area, bathroom occupation, displacement in the main room and bathroom. Some pieces of software written in C++ language and the results displayed on screen give for each monitoring period of 15 minutes:
- the number of times the person goes to bed, gets up
- the duration of each going to bed and getting up
- the start and the end of each event
- the path followed.
Within the framework of displacements in a straight line or pacing or circular or random wanderings involving several zones, the path length can be estimated with an error rate about 10 percent as shown in Fig. 4. On the other hand, the displacements that involve one small zone only cannot be estimated in terms of distance. The system can only show the number of changes of the state of the IR sensor.
| Top |
3. Concluding remarks and future prospects
This system which is operational in a laboratory setting is currently being installed on a real site in a long stay unit for the elderly in a hospital setting. The different components of the system are being developed. Displacements are recorded by taking into account the number of passive IR sensors used to monitor a person's mobility in the room. Nine sensors are now used for a surface of about 20 sq. m. Studies are being continued to improve the data acquisition system in terms of accuracy, reliability and sensitivity.
| Top |
Acknowledgments
The authors gratefully acknowledge the support of the Regional Council of Midi-Pyrénées (Research and Technology Transfers) through grants allotted to the project.
| Top |
References
Chan, M., Bocquet, H., Campo, E., Val, T., Estève, D. and Pous, J. (1996). Multisensor system and artificial intelligence in housing for the elderly and the disabled. Second International Gerontechnology Conference, Helsinki Finland, 15-17 October.
Chan, M., Hariton, C., Ringeard, P. and Campo, E. (1995). Smart house automation system for the elderly and the disabled. IEEE International Conference on Systems, Man and Cybernetics, Vancouver Canada, 22-25 October, 1586-1589.
O'Leary, P.A., Haley, W.E. and Paul, P.B.(1993). Behavioral assessment in Alzheimer's disease: Use of a 24-hr log. Psychology and Aging, 8(2), 139-143.
Matteson, M.A. and Linton, A. (1996). Wandering behaviors in institutionalized persons with dementia. Journal of Gerontological Nursing. 22(9), 39-46.
Turck, J. (1988). Les bases de la détection infrarouge. L'Onde Electrique. 68(2), 36-39.
Wagner, A.W., Teri, L. and Orr-Rainey N. (1995). Behavior problems of residents with dementia in special care units. Alzheimer Disease and Associated Disorders, 9(3), 121-127.
| Top | |TIDE 98 Papers |