=Paper=
{{Paper
|id=Vol-260/paper-16
|storemode=property
|title=Context based physiological signal analysis in a ubiquitous VR environment
|pdfUrl=https://ceur-ws.org/Vol-260/paper16.pdf
|volume=Vol-260
|dblpUrl=https://dblp.org/rec/conf/isuvr/ChoiW07
}}
==Context based physiological signal analysis in a ubiquitous VR environment==
https://ceur-ws.org/Vol-260/paper16.pdf
International Symposium on Ubiquitous VR 2007 1
Context based physiological signal analysis in a
ubiquitous VR environment
Ahyoung Choi, and Woontack Woo
precisely and reliably. Finally we derive the stress level
Abstract— With the advent of smart sensing technology, concerning about user characteristics such as sensitive or
environmental context and user context is widely deployed by the non-sensitive against the stimulus, pessimist or optimist, and
context awareness middleware developers and application healthy or non-healthy and the environmental condition such as
developers. This environment leverages to utilize the abundant
resources of information. In this paper, we present the principles
outdoor temperature, weather, sensing time and noise.
of context based physiological signal analysis in a ubiquitous VR
environment. Ultimate goal of this work is building the principles II. RELATED WORKS
of physiological signal analysis with the knowledge of unknown In previous work, there are some researches about context
control variables such as environmental condition and user
fusion and sensory data fusion, or health information fusion.
characteristics.
Subrata Das et al indicated the life status estimation with the
Index Terms— physiological signal analysis, context awareness, Bayesian network with history node [1]. It combines the heart
and personalization related factors such as certainty, previous measurements and
current measurements. However, it is not concerned about
sensing condition and noise, but physiological signal analysis
I. INTRODUCTION itself. In the field of wearable sensing technology, they also
have been concerned about tiny and light-weight wearable
W EARABLE or mobile physiological sensors have been
commercially sold in common with the development of a
next generation healthcare applications. One of the most
sensors and its integration. Armband type wearable sensor was
released from Bodymedia Company [2]. The sensor measure
serious problems in wearable physiological sensing system is multiple signals, which are galvanic skin response, skin-near
low reliability because of low SNR ratio in physiological signal. body temperature, two axis accelerometer and heat flux
There are some limitations when we analyze the physiological through the wireless network. They concerned motion artifacts
signal with wearable or mobile sensors due to the sensing in physiological signal analysis. However, motion artifact
noises and changes of sensing condition. In addition, it is hard filtering is just for the help of heart related signal analysis and
to obtain the same results obtained from distinct products with the system works after collecting the data in a offline manner.
respect to the same measurements. To solve this problem, we ubiMon project [3] also indicates the sensory information
are able to be aid from heterogeneous sensors and services in integration with multiple wearable physiological sensing nodes.
the ubiquitous VR environments. There are huge amount of However, it just collects the multiple contexts.
sensors and services which allow us to understand the current
situation and context. With this regards, we conclude that III. PROPOSED MNETHOD
sensor fusion techniques are extremely essential part in Context based physiological analysis with multiple context
analyzing physiological signal. information is enabling to make personalization and adaptation
In this paper, we describe the way of analyzing the to the user and the environment. In order to build the context
physiological signal indicator with multiple contexts. This based analysis, we model the relationship among the
analyzer has a role to integrate multiple physiological signals environmental context and physiological signal from the
for reducing measurement error which occurs in wearable previous literature and exploratory experiments for checking
computing environment and to combine user-centric context the relationship as a first step. The relationship is modeled with
into physiological states for reliable analysis of physiological the Bayesian belief network as Fig. 1 illustrated. The
signal. Knowing the previous condition before measuring the conditional probability table (CPT) is set with the previous
literature and the experiments with the process analysis or CNX
physiological signal gives us clues to understand the user status
analysis. As a results, we found that the injected stimulus, body
constitution, outdoor temperature, gender, clinical history, age
This work was supported by the IT R&D program of MIC/IITA. cause the increase of hear rate, and the radical changes of skin
[2005-S-069-02, Development of Technologies for Unifying and Fusing response and body temperature.
Context]
All is with GIST U-VR Lab, Gwangju, Korea (e-mail: {achoi,
wwoo}@gist.ac.kr).
International Symposium on Ubiquitous VR 2007 2
IV. EXPERIMENTAL ANALYSIS
In order to get the proper parameter for building relationship
model and decision making parameters, we did preliminary
experiments. In this experiment, we utilize the wrist type
physiological sensor which includes a 3-axis accelerometer
sensor and a PPG sensor and commercial sensing equipment-
BIOPAC. This experiment consisted of two phase. At first, we
should understand the human itself. With this expectation, we
observed the physiological signal from distinct user with same
measurement devise. Secondly, we verified the fact that the
environmental condition such as temperature or humidity levels
which may cause the certain stressful or uncomfortable
Fig. 1 Conditional probability table condition is used for setting the basic offset. Some items were
control variables when the experiments were conducted and
Secondly, we analyze the stress level based on the decision others were added by the theory referenced in this works. Table
making process. The stress levels and contextual information 1 and Table 2 illustrate the measurements from different user
are combined with the context inference module. After the and different sensing condition. From these experiments, we
relationship is set, we select some of the states which have observed that the different user and conditions are influenced
strong relationship from Bayesian belief network model. In the on the different situation.
decision making step, we determine the type of process and the
input factor. There are two kinds of decision making process. Table 1 Measurements of different users (User A, B, C)
First one is decision in final stage. It means that all feature PPG GSR
variables in the sensing step is obtained and the features pp min max stddev Med. mean slope
directly combines with weighted sum, which is illustrated in A 1.415 -0.331 0.952 0.330 -0.033 5.743 0.000
1.014 -0.534 0.683 0.232 -0.189 5.557 -0.003
Fig. 3.
B 3.179 -0.462 2.000 0.605 -0.026 14.202 0.008
Body 1.495 -1.179 0.986 0.390 -0.161 10.752 -0.010
Traits C 3.311 -0.835 1.740 0.685 0.011 4.566 -0.006
Character
1.848 -1.801 1.013 0.423 -0.053 2.473 -0.013
σ i
Temp Weather Making a
Stress level Table 2 Measurements in distinct conditions (PPG sensor)
Pre-stress Conditions pp Mean stddev Median
PPG
level Basic 1.546 12.242 0.111 -0.099
Noisy 2.829 14.298 1.155 -1.113
Fig. 2 Decision making process A Humid and hot 1.009 15.221 2.000 0.008
Second processing type is multiple threshold based decision, V. DISCUSSIONS
as depicted in Fig. 4. Classified the traits and weather condition
In this paper, we indicate the context based physiological
is inputted into decision threshold making step and these are
signal analysis method. After checking all the control
combined with each other. In this process, we can express the
parameters which we concern about, we conclude that the
inter-relationship such as body character and temperature if it is
higher temperature is derived from the anger of the subject with
said that higher temperature has more influence on the people
higher stress level. Therefore, we conjectured that we need at
who is sensitive against the stimulus. In this step, basic idea is
least two axes for supporting reliable analysis in order to allow
derived from the previous literature about sasang category in
the reliable analysis. As a future work, we have a plan to
oriental medicine which depicted the person into four types
investigate whether the selected sensing parameters and
such as tae-yang, tae-um, so-yang and so-eum [4]. This theory
decision parameters is appropriate.
complies with the personal distinction from the body shape,
characteristics, skin type and color and walking patterns. And REFERENCES
each person has different response in terms of the outdoor
[1] Das, S. K., Introne, J., Lawless, D., Hoyt, R. and Muza, S, “Probabilistic
condition, stress and so on. Unit Life Status Estimation (PULSE),” Proceedings of the 7th
International Conference on Information Fusion, pp. 959-966, 2004
[2] BodyMedia. Healthwear armband, bodybugg.
Body Traits σ i http://www.bodymedia.com
Character classification
N
β i = ∑ a × σ i +b × ωi [3] Kristof Van Laerhoven et al, “Medical Healthcare Monitoring with
i =1
Wearable and Implantable Sensors,” UbiHealth ,2004
[4] Chae H et al, “Alternative way to individualized medicine: psychological
Temp Weather ωi and physical traits of Sasang typology,” Journal of Alternative and
classification
Complementary Medicine, vol 9, 2003
Fig. 3 Decision making process B