Instrumentation & Measurement Magazine 26-4 - 37

Efficient Sitting Posture
Recognition for Wheelchair Users:
An Unsupervised Data-Driven
Framework
K. Ramakrishna Kini, Fouzi Harrou, Muddu Madakyaru, Farid Kadri, and Ying Sun
A
utomatic and reliable detection of a person's posture
when sitting in a wheelchair is necessary to
prevent major health issues. This study introduces
an unsupervised anomaly detection and isolation approach
to automatically recognize unbalanced sitting posture in a
wheelchair using data from pressure sensors embedded in
the wheelchair. Importantly, the advantages of independent
component analysis (ICA) will be integrated with those of a
Kantorovich Distance (KD)-driven anomaly detector by developing
an ICA-driven KD methodology that can handle
non-Gaussianity in the data and ameliorates the quality of
anomaly detection. Due to pressure data displaying a nonGaussian
behavior, this work adopts ICA, which is well suited
to handle this type of data. At the same time, the KD scheme
is an effective anomaly detection indicator to evaluate the ICA
residuals. Furthermore, the contribution plot strategy, which
does not need a priori knowledge of anomalies, is employed
for discriminating the type of the detected abnormal posture if
it is caused due to higher pressure on the right side, on the left
side, or higher forward pressure. The ICA-KD approach only
employs normal events data to train the detection model, making
them more attractive for identifying a person's posture in
practice. The overall detection system provides a promising
performance with an F1-score around 99.41%, outperforming
some commonly used monitoring methods.
Sitting Posture Monitoring for
Wheelchair Users
The number of wheelchair users increases year by year due
to different factors, including traffic accidents, falls, and violence.
According to the WHO World Report on Disability and
the Wheelchair Foundation, approximately 1.85% of people
worldwide need a wheelchair. For instance, there are 3.3 million
wheelchair users in the US, and the number is increasing
with an expected 2 million new wheelchair users every year.
In Europe, the number of wheelchair users is around 5 million
people who represent 1% of the population [1].
Wheelchairs are designed to improve users' quality of life
by facilitating mobility, social interaction, and occupation.
June 2023
However, unbalanced sitting posture in a wheelchair can
negatively affect the health condition of wheelchair users. It
is worth noting that even healthy people sitting in front of a
computer all day can be affected by sitting posture. Specifically,
incorrect posture in a wheelchair causes chronic pain,
sclerosis, kyphosis, skin and respiratory problems, loss of
brain skills, and physical health problems, such as muscle rigidity,
fatigue, and muscle pain [2]. Alternatively, adequate
sitting in the wheelchair can decrease pain intensity and the
possibility of ulcers formation [3]. Thus, accurately detecting
unbalanced sitting posture in a wheelchair provides relevant
information to the wheelchair user to avoid improper
postures.
In recent years, increased attention to sitting posture recognition
resulted in the development of various technologies in
terms of the adopted sensors, including wearable sensors, visual
sensors, and non-intrusive pressure sensors [2], [4]. In [5],
wearable optical fiber sensors were adopted to monitor the
user's sitting posture. However, the use of wearable sensors
by wheelchair users is not comfortable and is very intrusive.
Other researchers utilize information from images and videos
collected via visual sensors to recognize the sitting posture
[6]. However, the vision-based solution can seriously be impacted
by the lighting level. In addition, this solution could be
limited by privacy concerns for people being monitored. Alternatively,
non-intrusive approaches are generally based on
data gathered from pressure sensors embedded in the wheelchair.
Such approaches do not require human intervention and
no wearing of sensors, making them very suitable for practical
application [7].
Accurate recognition of sitting posture is vital to enhancing
wheelchair users' comfort and health monitoring.
Various non-intrusive methods have been introduced in the
literature for sitting posture recognition. For instance, in
[8], a supervised approach is employed that uses principal
component analysis (PCA) as a feature extractor and the knearest
neighbors (KNN) classifier for posture recognition
in conventional chairs. This approach reached a classification
accuracy of 75%. This is because PCA is suitable only for
IEEE Instrumentation & Measurement Magazine
1094-6969/23/$25.00©2023IEEE
37
Instrumentation & Measurement Magazine 26-4

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 26-4
