A Personalized Rehabilitation System based on Wireless Motion

Capture Sensors

Pedro Macedo

1,2

, José A. Afonso

and Ricardo Simoes

2,3,4

Department of Industrial Electronics, University of Minho, Guimarães, Portugal

Institute for Polymers and Composites IPC/I3N, University of Minho, Guimarães, Portugal

Life and Health Sciences Research Institute (ICVS/3Bs), University of Minho, Braga, Portugal

School of Technology, Polytechnic Institute of Cávado and Ave, Barcelos, Portugal

Keywords: Motion Capture, Physiotherapy, Wireless Sensor Networks, Wearable Sensors.

Abstract: We live in an aging society, an issue that will be exacerbated in the coming decades, due to low birth rates

and increasing life expectancy. With the decline in physical and cognitive functions with age, it is of the

utmost importance to maintain regular physical activity, in order to preserve an individual’s mobility, motor

capabilities and coordination. Within this context, this paper describes the development of a wireless sensor

network and its application in a human motion capture system based on wearable inertial and magnetic

sensors. The goal is to enable, through continuous real-time monitoring, the creation of a personalized

home-based rehabilitation system for the elderly population and/or injured people. Within this system, the

user can benefit from an assisted mode, in which their movements can be compared to a reference motion

model of the same movements, resulting in visual feedback alerts given by the application. This motion

model can be created previously, in a ‘learning phase’, under supervision of a caregiver.

1 INTRODUCTION

Physiotherapy is a therapy which aims to help

recover movement and restore normal body

function, when a person is affected by illness, injury

or disability. Musculoskeletal, neuromuscular,

cardiovascular and respiratory are some of the

physical problems physiotherapists can help treat.

Both physical and cognitive functions decline

with aging. Compensatory and preventive

management can control the effects of physiological

system decline. Generally people reach their peak of

performance abilities and health at ages between

adolescence and 30 years. After this period, it is

known that functional capacity declines throughout

the person’s lifespan, depending on genetics,

lifestyle and overall health (Nitz and Hourigan,

2004).

Physiotherapy plays an important role in this

aging society. Well trained health care professionals

in this area can assist the patient in the rehabilitation

process and improve his quality of life. It is

important to maintain constant physical activity, in

order to preserve movement ability, physical

adaptability, mobility and coordination of the

individual, by his own means or with assistance,

whether at home or in an assisted living facility.

Movements can be categorized within three levels:

active movement, which describes the movement

that a can be initiated and controlled independently;

assisted active, where in order to achieve a complete

movement some initial and control movement effort

has to be taken from the resident, with the caregiver

assistance; and passive movement, where all the

movement is performed by the caregiver. Among the

many benefits of doing exercise, we can cite the

improvement of strength, increased range of motion,

improved flexibility, improved functional mobility,

increased self-esteem, improved posture, improved

gait, improved efficiency of movement and

improved quality of life (Nitz and Hourigan, 2004).

Taking this into consideration, this project aims

to create a platform which enables and supports the

practice of physical activity; and/or serve as a

support platform for home-based physical

rehabilitation, providing feedbacks concerning the

correctness of the executed movement.

In the case of rehabilitation, the physiatrist

and/or the physiotherapist will assess the nature of

the patient’s problems, set goals based on the extent

220

Macedo P., Afonso J. and Simões R..

A Personalized Rehabilitation System based on Wireless Motion Capture Sensors.

DOI: 10.5220/0005238202200228

In Proceedings of the 4th International Conference on Sensor Networks (SENSORNETS-2015), pages 220-228

ISBN: 978-989-758-086-4

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

of those problems, provide an adequate treatment

and continuously evaluate the patient’s progress.

Physiatry, or rehabilitation medicine, has as one of

its main aspects the recognition of functional deficits

caused by injury or illness. The identification of

these deficits is of utmost importance, in order to

best implement a proper treatment program to

restore the patient’s performance (Braddom, 2006).

A superior treatment can only be taken if the

physiatrist is fully aware of the patient’s history and

difficulties he may have conducting the prescribed

exercises. According to the author (Braddom, 2006),

there are seven levels of disabilities. The level 1 of

the scale is the most critical case, where total

assistance is needed and the patient expends less

than 25% of the effort. The scale from 1 to 5

encompasses a dependent disability, meaning that

the patient requires another person to either

supervise or assist in the performed activity; whereas

in scales 6 to 7 assistance in not necessary to

perform the activity.

Normal activities of daily living (ADL) like

bathing and showering, dressing, eating and

functional mobility, may suffer due to any existing

physical problem. The combination of these

problems, adding to the effort that some activities,

like standing up, sitting down or picking up things,

exert in our body joints, leads, in extreme cases, to

the use of a prosthetic device to preserve the

person’s quality of life. Total hip replacement and

rehabilitation is a perfect example of the interaction

that should exist between the patient and his

caregiver. After the surgery, in order to achieve a

satisfactory level of functionality and independence

in the patient’s daily life, it is important to continue

with the prescribed daily exercises. Patients with hip

replacement have limited motion amplitude, so

movements have to be restricted during the first

three months after the surgery. Flexion above 90º

and flexion combined with abduction movements

should be avoided, to prevent the prosthesis

displacement (Brander and Stulberg, 2006) (O’Leary

et al., 2011).

The continuous growth of the ratio of elderly

population compared to the total population in

developed societies causes a real and possible

problematic demographic change (Linz and Stula,

2012) (Stula, 2012). In order to prevent this

structural demographic change, new services and

products or ambient assisted living (AAL) must be

created. The goal of such products is to provide

better life conditions for the older generation in their

environment, by increasing their self-confidence,

autonomy and mobility (Sun et al., 2009)

(Fuchsberger, 2008) (Kleinberger et al., 2007).

This paper aims to present the current status of a

wireless posture monitoring system, based on

wearable inertial and magnetic sensors, applied to

rehabilitation. The purpose is to create a

personalized home-based exercise assisted mode for

the elderly people. The captured movements are

evaluated and compared to a prescribed exercise,

with the purpose of assist the user to correctly

perform the proposed movements.

This paper is organized as follows. The next

section presents the related work. An overview of

the implemented system is presented in Section 3.

Section 4 introduces and explains the methods used

to create a rehabilitation system, provided with

visual feedback to the user, while section 5 presents

the conclusions and discusses future work.

2 RELATED WORK

The authors in (Cavallo et al., 2009) describe the

implementation of a pervasive intelligent system for

rehabilitation, composed of a ZigBee network, with

a coordinator and actuator nodes able to identify and

control the patient’s activities, and send warnings to

a caregiver.

Within the same context, with the goal to deliver

health care services to the community, a

rehabilitation service operating at a person’s home

over a telecommunications network was developed

in (Hamel et al., 2008). According to the authors,

most types of telerehabilitation services fall into two

categories: clinical assessment of the patient’s

functional abilities in his environment; and clinical

therapy. The developed system was based in

cameras, local and remote computers, with dedicated

(and user-friendly) modular software interfaces for

videoconference connections, and sensors

(accelerometer and gyroscope).

The authors in (Lu et al., 2013) developed a

system to enable a set of balance rehabilitation

exercises for patients with spinocerebellar ataxia

(SCA), for continuous and safe practice at home,

through information and communication technology.

The system integrates physiological monitoring and

feedback coaching with a telecare center that enables

real-time one-to-multiple personal exercises

monitored by care managers.

Unlike the last two studies, which are based on

the combination of data from several cameras, to

obtain a 3D location of the patient’s body and limbs,

this paper aims to represent a 3D model of a

patient’s body segments and the development of a

APersonalizedRehabilitationSystembasedonWirelessMotionCaptureSensors

221

system based on a wireless sensor network (WSN)

and wearable inertial and magnetic sensors. When

compared to a camera approach, this method has the

advantage of higher flexibility and mobility, as it can

be used in uncontrolled environments, without

lighting or line-of-sight concerns (Aminian and

Najafi, 2004).

Physiotherapy can take advantage of the

monitoring of human body movements (body

kinematics) in areas such as health care (to treat

patients) or sports (to support the athlete recovery or

improve his performance). However when creating a

WSN to monitor human body motion, several

factors must be taken into account in order to assure

a reliable system. These factors include: energy

efficiency, since normally the sensor nodes are

energy constrained; sensor node fixation in the body;

high amount of data generated per sensor node

(unlike typical WSNs); and the impact of the human

body on the wireless signal propagation. Some of the

technical challenges faced at the creation of a WSN

for rehabilitation purposes are described by the

author in (Hadjidj et al., 2012).

The performance of inertial, magnetic and

gyroscope sensors, when applied to body kinematics

measurement, was underlined by several authors. A

study that compares the anatomical joint angles

obtained by an inertial measurement unit (IMU) to

those calculated from position data of an optical

tracking device is presented in (Bergmann et

al., 2009). The two measurement methods were

evaluated by calculating the root mean square error

(RMSE) and by calculating a two-tailed Pearson

product-moment correlation coefficient between the

two signals. Studies show a strong correlations,

range 0.93 to 0.99, between the two signals, as well

an average RMSE of 4 degrees over the joint angle.

It is concluded that IMUs offers a good alternative

system for measuring anatomical joint angles, by

providing an opportunity to perform accurate

measurements in complex real-life environments

without using constrained measurement device

(markers system). In (Lin and Kulic, 2012), the

authors developed a kinematics system to estimate

the human leg posture and velocity. The posture was

captured through wearable sensors (accelerometer

and gyroscope) during the performance of typical

physiotherapy and training exercises. An extended

Kalman filter is applied to estimate joint angles

during an arbitrary three dimensional motion. This

type of system enables applications such as

monitoring during knee and hip rehabilitation.

The two last mentioned studies can merely serve

as an alternative tool for the traditional gait analysis

system based on high-speed cameras. On the other

hand the system presented here, which also serves as

a gait analysis, offers in addition the ability to

ascertain the proper body posture and good

movement execution by the user in real-time.

3 SYSTEM OVERVIEW

The developed wireless posture monitoring

comprises multiple sensor nodes, each one attached

to a monitored body segment, and a base station that

sends the collected data to a personal computer (PC).

The collected information consists in inertial and

magnetic readings. The processing of the

information in the PC enables the calculation, in

real-time, of the 3D orientation of the module,

expressed by the pitch, roll and yaw angles. Figure 1

represents the structure of the developed system.

Figure 1: System components.

In this system, the CC2530, from Texas

Instruments (Texas Instruments, 2009), allows the

wireless communication between the base station

and the sensor nodes. The CC2530 is a true system-

on-chip (SoC) solution for IEEE 802.15.4

applications (IEEE Std 802.15.4, 2006) which

integrates an 8051 based microcontroller and an

802.15.4 transceiver working in the license-free

2.4 GHz frequency band.

The control of the communication through the

wireless medium is achieved through the Enhanced

Low Power Real Time (eLPRT) MAC protocol

(Afonso et al., 2011), which was designed to

optimize the quality of service (QoS) provisioning

and the bandwidth utilization efficiency.

3.1 Node Architecture

The base station and sensors nodes have the same

architectural principle, as illustrated in Figure 2, but

differentiating on the communication between the

microcontroller and the sensors, which is absent on

the base station.

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

222

Figure 2: Base station and sensor architecture.

The Universal Synchronous Asynchronous

Receiver Transmitter (USART) and Radio

peripherals are controlled by the CC2530 with the

assistance of the Direct Memory Access (DMA)

subsystem. Some of the station base responsibilities

are the association of new sensor nodes into the

network, the allocation of time slots on the eLPRT

superframe for the nodes transmission, and the

network synchronization with the aid of periodic

beacons.

The information regarding the number of

readings of the sensors per superframe and the

transmission period is set by the user’s application.

In the sensor nodes, data messages containing the

readings of the 3-axis inertial, magnetic and

gyroscope sensors, temperature and battery level are

aggregated into a single message, according to the

number of readings per superframe. The sensor

information is collected through the serial peripheral

interface (SPI) bus and sent to the base station once

per superframe using the eLPRT protocol.

3.2 Sensor Node Prototype

Within the sensor prototype (Macedo et al., 2014),

found, the MPU-6000 Motion Processing Unit from

InvenSense, which has an embedded 3-axis MEMS

accelerometer, a 3-axis gyroscope (readings not used

at the moment) and a digital motion processor

(DMP) hardware accelerator engine with an

auxiliary I

C port that interfaces to a third party

digital sensor, such as a magnetometer. The

Honeywell 3-axis Digital Compass IC HMC5883L

is used in this prototype. More detailed information

about the sensors is described in (Macedo et al.,

2014).

3.3 PC Software

The system component with the greatest relevance to

the user is the PC application. This application

serves as a mediator between the users and the base

station; it can send commands and receive

information sent by the sensor nodes attached to the

body segments. Upon reception of the data from the

sensors, it calculates the angles of rotation and

presents the movement of the user’s body in real-

time on a 3D model of the human body.

One sensor node per monitored segment is

necessary to measure the movement, with three

degrees of freedom (DOF), expressed by the Euler

angles pitch, roll and yaw. The overall system was

designed to be as intuitive and easy to handle by the

user as possible. When compared to the previously

presented application (Macedo et al., 2014), this one

presents new features. The software provides means

to create learning files, to later compare movements;

and rehabilitation sessions files, where user

information and motion angles are stored. It also

introduces an assisted mode, where feedback

regarding the correctness of the movements is given

to the user. These session files can later be employed

to evaluate the user’s progress. A toolbar with the

most used buttons was also included, as well as

LEDs indicating that a learning session or

rehabilitation is being recorded. Figure 3 presents

the main window of the developed software.

Figure 3: Application interface.

4 SYSTEM EVALUATION

Advances have been made regarding the application

described in (Macedo et al., 2014). At this point, the

developed real-time 3D Java application, not only

evaluates the motion capture capabilities of the

system, but can also provide movement feedback.

The goal is to develop a home-based rehabilitation

system that, through constant monitoring of the

movement, is capable of interacting in real-time with

APersonalizedRehabilitationSystembasedonWirelessMotionCaptureSensors

223

the user. The purpose of this interaction is to assist

the user to perform the prescribed exercises

correctly. Poorly executed exercises can delay the

rehabilitation process, or even cause more damage.

Thus, the created application can serve as a

personalized exercise reference to a prescribed and

firstly assisted session, or even generic movement

body amplitudes; or merely serve to assess the

patient’s progress through the registration of

rehabilitation log sessions (if desired, both can be

performed simultaneously). This personalized

exercise reference system has three stages. The first

stage consists of a learning mode (section 4.1),

where it is showed to the patient how to perform the

exercise according to its specifications (special

needs he may have). This stage can be replaced for a

generic amplitude movement (correct patient

positioning and plane of motion for the moved body

joints), as the ones seen in (Braddom, 2006).

Amplitude movements, such as shoulder flexion and

extension, shoulder abduction, shoulder internal and

external rotation, elbow flexion and more. The

second stage (section 4.2) is characterized by the

personification of each patient movement into an

individual model. Finally, in the third stage (section

4.3), the patient can perform the prescribed exercises

at home, through the guidance of the assisted mode.

The learning mode and personalized model will

be further discussed in the sections 4.1 and 4.2

respectively. Section 4.3 will focus on the assisted

mode and a more detailed explanation of how this

assisted mode is achieved is described in section

4.3.1. Finally section 4.4 deals with the

rehabilitation sessions.

4.1 Learning Mode

Within this phase the patient will be taught, by

medical specialist, the correct movement that should

be replicated at home during the assisted mode. This

method aims to create a text file, which will serve,

later on, to learn and create a motion model of a

correctly performed exercise.

Each type of performed exercise creates a unique

motion model, depending on the type of exercise and

subject. Hence, every recorded movement is saved

on a text file, containing the angles between each

module present in the network and the planes of

motion of the global axis system and/or the

reference sensor node.

Within the learning file, a header (common to all

files) is created containing the learning session

number, transmission period (superframe duration),

number of samples per period, date of session and a

record of the angles saved to the file (the angle name

given below depend on the specific movement, in

this case arm):

#(angle chestTranversePlane

#chestCoronalPlane chestSagitalPlane

#armTranversePlane armCoronalPlane

#armSagitalPlane forearmTransversePlane

forearmCoronalPlane forearmSagitalPlane

#armOrientation forearmOrientation)

The angle mentioned in the header file it is the

angle made between two segments of the body.

Taking as an example the arm, the angle made

between the arm and the forearm. Chest angles with

the motion plane are set with the global axis system;

arm and forearm angles are set with chest as

reference.

armOrientation and

forearmOrientation are direction angles,

calculated with the chest reference. Further along

(Section 4.3.1), will explained with more detail how

these angles are calculated. Later these angles will

serve as a means to create the motion model to an

exercise. The process of creating a new learning file

is initiated by the user and/or the physiotherapist.

Also, this process is protected by a password in

order to prevent recording false movements. Every

new reading (correspondent angle calculation) from

the sensors represents a new line in the text file.

4.2 Personalized Motion Model

The goal of this phase is to obtain, for the individual

subject and for each exercise, a personalized model

of one motion cycle, which will serve as reference to

the assisted exercise phase (section 4.3). After

opening the file, a motion model is immediately

created. This motion model is not more than the

processing of the angles recorded into a text file

during the learning phase.

At this moment the application only deals with

uniform movements, that is, simple movements, as

the ones seen in (Braddom, 2006) like: shoulder

flexion and extension; shoulder abduction; elbow

flexion; hip flexion, knee flexion; hip abduction; hip

flexion, knee extension; knee flexion; hip internal

and external rotation. The current motion model uses

a global threshold system of maximum and

minimum limits that an angle can take with the

motion planes (global axis system and reference axis

system).

4.3 Assisted Mode

The replication of movement through the 3D model

of the human body is inherent to the application. It is

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

224

as simple as attaching a sensor node in the body

segment which you want to reproduce the

movement. On the other hand, the creation of logs of

rehabilitation sessions and learning processes must

be initiated by the users. As opposed to these

processes, the assisted mode depends on the pre-

existence of a learning file, so that a motion model

can be created. This assisted mode consists in

comparing the performed motion to a stored

reference motion model of the same movement

(learning file). The result of this comparison is

translated to a visual feedback to the user. To start

the assisted mode, the user must first open the

correspondent learning file. After opening the file

and correspondent motion model creation, the

assisted mode and visual feedback can begin.

At home, when using the assisted mode, the

application should be able to inform the user

whether the movement performed previously with

the caregiver, is being well executed. Figure 4

demonstrates the capabilities of the human 3D

model and application to provide feedback to the

user about the correctness of the movement. The

yellow and red colors are used to inform the user

that the allowed limits for the motion were exceeded

(yellow for the first threshold and red for the second

threshold). A detailed view of the angles that are

exceeding their thresholds is shown in the bottom

right side of the window, which presents the values

of the angles that the segment makes with each

motion plane. For this test, a learning file was

created, a simple flexion and extension of the right

arm, and then the user was asked to replicate the

previously taught movement.

Figure 4: Application interface in assisted mode.

4.3.1 Assessment of Range of Motion

In (Braddom, 2006), the author provides guidelines

to assess the efficacy of treatment procedures, to

determine the biomechanical cause of impairment

and to document the integrity of a joint range of

motion. According to the author, normal range of

motion (ROM) varies based on age, obesity, gender,

conditioning and genetics. Moreover, men have a

more limited range when compared to women,

depending on age and specific joint action. In order

to better assess the motion, passive ROM tests

should be performed by the examiner, thorough all

planes of motion, in a relaxed patient. On the other

hand, active ROM tests through all planes of motion,

performed by the patient without assistance from the

examiner, simultaneously evaluate muscle strength,

coordination of movement and functional ability.

The range of motion is measured with a universal

goniometer and should be performed prior to

strength testing. The developed system depicted here

can serve as a test goniometer to assess patient range

of motion and correctness of movement. The planes

of motion mentioned earlier consist in a division of

the human body into three cardinal planes, as shown

in Figure 5. The sagittal plane divides the body into

left and right halves. The frontal (coronal) plane

divides into anterior and posterior halves. Lastly, the

transverse plane divides the body into superior and

inferior parts.

Figure 5: Cardinal planes of motion (adapted from

Braddom, 2006).

The planes of motion are defined according the

coordinate system as follows: the transverse plane is

delimited by the x and y axis; the sagittal plane, by

the y and z axis; and the coronal is bounded by the x

and z axis. Using this global coordinate system, or

planes, defined by the axis position of a reference

module for other at the time of the movement, it is

possible to calculate the range of motion along the

planes. As an example the chest module can serve as

a reference for the arm modules. The coordinate

system used to calculate the three orientation angles

was the right-handed. The forward orientation of the

module is set to be along the y axis, with pitch

defining the rotation on the x axis (





∝



), the roll

APersonalizedRehabilitationSystembasedonWirelessMotionCaptureSensors

225

on the y axis (









) and yaw on the z axis (









according to equations 1, 2 and 3 respectively. After

the matrix multiplications (rotations), the final

matrix of the module’s orientation is found in

equation 4. This matrix is needed in case the

reference for the motion planes is another sensor

node; within this matrix the module’s axis can be

found. In those cases where the reference is the

global system, the axes are set to its origin: x axis to

[1 0 0]; y axis to [0 1 0]; and z axis to [0 0 1], equal

to the identity matrix.

The module’s orientation vector is given by

multiplying the final rotation matrix by its original

position [0 0 1], x axis, y axis and z axis

respectively.







∝





10 0

0 cos∝ sin∝

0 sin∝ cos∝



(1)













cos 0 sin

010

sin 0 cos



(2)













cos sin 0

sin cos 0

001



(3)





,∝,



















∝













(4)

The orientation vector for the reference node

(sensor node or not) is given by the perpendicular

vector (normal vector) of the plane we want to

calculate. In other words, if we want to calculate the

angle between one module orientation vector and the

transversal plane, the orientation vector of that plane

is its z axis. The angle () between the two

orientation vectors is calculated using equation 5.

Therefore, the mentioned angle defines the angle

between a given module orientation vector and a

plane of motion (Dunn et al., 2011).











⋅



















(5)

So far, the three planes of motion gave us three

angles to each module in relation to its reference(s),

but a fourth angle can be calculated, the torsion

angle. This angle reflects the difference between the

one module and its reference(s) direction. As an

example, if both modules are presented northwards,

the angle between them is zero. This angle must be

calculated with another module as reference,

because otherwise there would be no way to predict

the initial state (or direction) of the patient. The

plane used in the reference module is the sagittal,

and on the side of the other module it is used its

direction, y axis. The angle is calculated using

equation 5, with the y axis being considered the

orientation vector of the module; this value was

obtained in the final rotation matrix of the module.

The purpose of calculating all of these possible

angles, for each module, relative to the motion

planes, is to better assess the user’s performance.

4.4 Rehabilitation Session

This rehabilitation session mode was created in the

application in order to create session logs that can

later serve to gauge the patient’s progress during the

rehabilitation process. A new rehabilitation session

must be initiated by the user; only this way the

application will automatically create the new session

file.

Several experiments were carried out to evaluate

the developed system. The one presented in this

paper had three sensor nodes attached to the body,

two of them on the right arm (upper arm and

forearm) and the other on the chest. The goal was to

verify the replication of the body movement in the

3D model present in the application and, at the same

time, produce a rehabilitation session file. Figure 6

and Figure 7 exemplify these experiments by means

of photographs taken while the movement was being

executed.

Figure 6: Right arm movement replication, first position.

Figure 8 presents the angle between the upper and

forearm along the time. The user was requested to

do a simple movement, flex and extend the arm,

multiple times. Information regarding the user name,

session number, type of rehabilitation, superframe

period, number of samples per superframe and date

of session can be seen on the left side. In the future,

more body angles will be added to the chart,

according to the performed movement and including

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

226

only the most relevant angles from those recorded

on the text file.

Figure 7: Right arm movement replication, second

position.

Figure 8: Chart of the angle of the articulation.

5 CONCLUSIONS

This paper outlines the design and development of a

wireless sensor network based on wearable and

easily placed inertial and magnetic sensors, which is

applied to monitoring of physical activities in

clinical settings and home environments. Such a

system can have tremendous application to

complement and enhance physical rehabilitation

procedures. An application capable of collecting

data from the sensors and present them in a real-time

3D model of a human body is under development,

focusing on the special case of assisting elderlies in

the practice of physiotherapy at home.

The main objective of this project consists in

developing a home-based rehabilitation system that,

through constant monitoring of the movement, is

capable of interacting in real-time with the user. The

purpose of this interaction is to assist the user to

perform the prescribed exercise correctly. The

rehabilitation process can be delayed or even cause

more damage if the exercise is poorly performed.

Therefore, this system can help medical specialists

to take into account physical limitations of the

patients in the preparation of physiotherapy sessions,

and to set well-defined objectives with regard to

patient outcomes.

The caregiver presence is not required during the

home-based sessions (assisted mode), but only in the

first session (learning mode). Unlike camera-based

systems, the patient can freely move anywhere under

the coverage of the WSN.

As future work, it is intended to create graph

sessions providing more useful information, such as

other meaningful angles, in order to create

rehabilitation session files that can better serve the

caregiver when assessing the patient’s progress.

Another feature under study is the compensation

of misplacement of the modules in the human body,

that is, if the sensor nodes are not attached correctly

to the chosen body segment, through the automatic

detection and correction of angle offsets with

relation to the correct position.

Studies are also being carried out in order to

improve the existing motion model. So far, a global

thresholds technique is being used, which places

limits to the angles which are used throughout the

movement and provides feedback to the user about

the correctness of his motion. Although this solution

results for simple movements, it is not able to

account for more complex movements. Therefore, a

new motion model based on local thresholds and

state machines is currently being implemented.

ACKNOWLEDGEMENTS

Project “AAL4ALL”, co-financed by the European

Community Fund FEDER through COMPETE –

Programa Operacional Factores de Competitividade

(POFC). FCT – Foundation for Science and

Technology – Lisbon, Portugal, through project

PEst-C/CTM/LA0025/2013.

REFERENCES

Afonso, J., Silva, H., Macedo, P., Rocha, L., 2011. An

Enhanced Reservation-Based MAC Protocol for IEEE

802.15.4 Networks. Sensors, Vol. 11, Issue 4, April

2011, pp. 3852-3873.

APersonalizedRehabilitationSystembasedonWirelessMotionCaptureSensors

227

Aminian, K., Najafi, B., 2004. Capturing human motion

using body-fixed sensors: outdoor measurement and

clinical applications. In Computer Animation and

Virtual Worlds, Volume 15, Issue 2, pp. 79-94.

Bergmann, J., Mayagoitia, R, Smith, I., 2009. A portable

system for collecting anatomical joint angles during

stair ascent: a comparison with an optical tracking

device. In Dynamic Medicine 2009, 8:3.

Braddom, R. L., 2006. Physical Medicine &

Rehabilitation. Third Edition.

Brander, V., Stulberg, S. D., 2006. Rehabilitation After

Hip and Knee Joint Replacement. An Experience and

Evidence-Based Approach to Care. In American

Journal of Physical Medicine & Rehabilitation, Vol.

85, No. 11 (Supplement): S98-118.

Cavallo, F., Aquilano, M., Odetti, L., Arvati, M.,

Carrozza, M.C., 2009. A first step toward a pervasive

and smart ZigBee sensor system for assistance and

rehabilitation. IEEE 11

International Conference on

Rehabilitation Robotics Kyoto International

Conference Center, Japan, June 23-26, 2009.

Dunn, F., Parberry, I., 2011. 3D Math Primer for Graphics

and Game Development. Second Edition pp. 141-143,

ISBN-13: 978-1-4398-6981-9, 2011.

Fuchsberger, M., 2008. Ambient Assisted Living: Elderly

People’s Needs and How to face Them. In

Proceedings of the 1

ACM international workshop on

Semantic ambient media experiences.

Hadjidj, A., Souil, M., Bouadballah, A., Challal, Y.,

Owen, H., 2012. Wireless sensor networks for

rehabilitation applications: Challenges and

opportunities. In Journal of Network and Computer

Applications.

Hamel, M., Fontaine, R., Boissy, P., 2008. In-Home

Telerehabilitation for Geriatric Patients. IEEE

Engineering in Medicine and Biology Magazine.

July/August 2008.

IEEE Std 802.15.4-2006, Part 15.4: Wireless Medium

Access Control (MAC) and Physical Layer (PHY)

Specifications for Low-Rate Wireless Personal Area

Networks (WPANs), September 2006.

Instituto Nacional de Estatística, Statistics Portugal, 2014.

Resident population in Portugal with a decreasing and

aging trend. Press release, 10 July 2014.

Kleinberger, T., Becker, M., Ras, E., Holzinger, A., 2007.

Ambient Intelligence in Assisted Living: Enable

Elderly People to Handle Future Interfaces. In 4

International Conference on Universal Access in

Huma-Computer Interaction, UAHCI.

Lin, J., Kulic, D., 2012. Human pose recovry using

wireless inertial measurement units. In Physiologial

Measurement, Volume 33, Number 12, 2012.

Linz, K., Stula, S., 2012. Demographic change in Europe –

An Overview. Working Paper No. 4 of the

Observatory for Sociopolitical Developments in

Europe.

Lu, T., Liu, H., Chen, R., Chen, Y., 2013. Motion-

Sensisng Based Management System for Smart

Context-Awareness Rehabilitation Healthcare. In

Advances in Internet of Things, 2013, 3, 1-7.

Macedo, P., Afonso, J. A., Rocha, L. A., Simões, R., 2014.

A Telerehabilitation System Based on Wireless

Motion Capture Sensors. International Conference on

physiological Computing Systems, 7-9 January, 2014.

Nitz, J. C., Hourigan, S. R., 2004. Physiotherapy Practice

in Residential Aged Care. ISBN 0 7506 8772 X.

O’Leary, K., UW Health Sports Medicine physician

group, 2011. Rehabilitation Guidelines for Surface

Replacement Arthroplasty (SRA). University of

Winsconsin Sports Medicine.

Stula, S., 2012. Living in Old Age in Europe – Current

Developments and Challenges. Working Paper No. 7

of the Observatory for Sociopolitical Developments in

Europe.

Sun, H., De Florio, V., Gui, N., Blondia, C., 2009.

Promises and Challenges of Ambient Assisted Living

Systems. In Information Technology: New

Generations. ITNG’09.

Texas Instruments, “CC2530 data Sheet” April 2009.

Retrieved from http://www.ti.com.

SENSORNETS2015-4thInternationalConferenceonSensorNetworks

228