Instrumentation & Measurement Magazine 24-6 - 80

Fig. 1. Layers of mobility related cognition.
indicative of cognitive decline, there have been many clinical
tools developed to assess cognitive well being directly.
Examples of clinical tools used to assess cognition include
the Montreal Cognitive Assessment (MoCA), Trail Making
Test A and B, the Mini-Mental State Examination and the Repeatable
Battery for the Assessment of Neuropsychological
Status (RBANS) [10]-[13]. These clinical tools are created by
healthcare professionals and are often administered via visual
and written observation. They are therefore limited to
what can be observed by the human eye. Ambient sensing,
in conjunction with signal and image processing allows for
the collection of data from patients in an unobtrusive manner
capable of capturing changes in balance and mobility that
cannot be observed visually.
Ambient sensing is not a new area of research; sensor
technologies have become smaller, more sensitive and more
sophisticated [14]. Pervasive sensing systems embedded in a
subject's environment can provide continuous monitoring of
mobility and activities of daily living (ADLs) [14]. This paper
examines current ambient sensing systems, including conjunctive
signal and image processing techniques used to measure
and monitor physical functioning; the structure for which is illustrated
in Fig. 1. This figure illustrates the first two sections
of this paper and depicts two mobility layers that are divided
based on the general level of required cognition. The first layer
is considered the basic step in cognitive function: movement,
requiring a low level of cognition. Building upon movement
is the activity layer, where movements combine to become action,
a set of multiple actions can be considered activity and
ADLs are activities that have been shown to be representative
of higher cognitive functioning required to coordinate ADL
movements [4], [15].
Ambient Sensing for Movement
The relationship between mobility and cognition is not
straightforward [16]. Here, two main distinctions are made
between the first layer of mobility, movement, and the second
layer of mobility, action. This distinction being the level of
cognition required to perform a movement or an action where
movement requires less cognitive load than action does. The
term 'less' is relative; movement is the result of a series of mental
processes that can be occurring whether movement results
or not [17]. When movement does result, it involves feedback
80
that allows for the perception of change and subsequent motion
planning [17]. Movement is therefore considered a higher
order function, despite requiring less cognitive load than
action.
Typical human mobility is bipedal and primarily defined
by the ability to walk [7]. Normal gait requires several co-functioning
neural systems that include initiation and maintenance
of steps, balance and maneuvering through environmental
factors [17]. If any of these neural systems is damaged, gait is
affected [5]. Impaired high order function has been associated
with decreased walking speed, increased stride time variability
and increased incidence of falls [5], [18], [19]. Gait analysis
is therefore an important measure in clinical practice as it provides
information about a subject's overall health [20].
Technologies used to capture gait measurements can be
wearable or non-wearable [21]. Ambient sensors are nonwearable
devices that are embedded in one's environment
and can capture information regarding a subject's everyday
activities. In the context of gait, there are two primary types
of sensors used: pressure based or image based [21]. Pressure
based sensors can be embedded in the floor to capture the force
exerted by the subject's feet while they walk. Image based systems
can be visible light cameras or thermal infrared cameras
and are designed capture the way a subject moves [21]. These
primary ambient sensor types are explored in the context of assessing
movement in addition to ambient sensor types that are
not pressure or image based.
Pressure-based Sensing
Pressure-based sensors have been comprised of varying technologies,
but generally are of four main types: capacitive,
resistive, piezoelectric and piezoresistive sensors [22]. Sensor
choice is often a subjective, multifactorial decision, some
general considerations being that resistive sensors have high
resolution but are highly subject to temperature and humidity,
where capacitive and piezoelectric sensors are reliable
and easily calibrated but often expensive [22]. Some examples
of pressure based sensor technologies include the use of
transducers in load cells, polypropylene with air pockets in
electromechanical film sensors, resistive ink in piezoresistive
sensors and light deformation in fiber optic sensors [23]-[27].
While the specific technologies vary, all sensors directly or
indirectly measure the degree to which a subject is exerting
pressure on a surface. The main differences in sensor technologies
are the range of pressure, linearity and sensitivity and
the sensor types are chosen to reflect application needs. In the
measurement of gait, pressure sensors are widely used and
can be placed on the floor to quantify pressure patterns underneath
the foot. Most pressure sensors do not typically detect
vertical force components (they detect the summed force vector)
and as a result cannot determine shear force that can lead
to tissue stress [22]. Despite the inability to determine shear
force, pressure sensors are widely used in determining under
foot pressure patterns in addition to gait velocity, step length
and step width [22]. Several clinical studies have used pressure
sensors to characterize gait impairments in subjects with
IEEE Instrumentation & Measurement Magazine
September 2021
Instrumentation & Measurement Magazine 24-6

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 24-6
