Smart Moving Nightstand

For Medical Assistance of Elderly People an Open Project

Nicolas Soucies

1,2

, Jeremie Girouard

and Nizar Ouarti

Institute des Systemes Intelligents et de Robotique, Universite Pierre et Marie Curie, 4 place jussieu, Paris, France

Alten, 221 Boulevard Jean Jaurès, Boulogne-Billancourt, France

Keywords: Elderly People, Health, Remote Sensors, Zigbee, Robot, Open Platform.

Abstract: We present an open mobile platform that aims to benefit from versatile wireless sensors. This mobile

assistant is a robot that can monitor different physiologic data for elderly people. Moreover it has the ability

to determine the distance and potentially the position of the elderly person who use it. As an assistant it can

transport some “objects” as glasses or drugs. Preliminary results show the proof of concept of our approach

with a remote sensor that measures the temperature of the subject. We also present a method to assess the

quality of the RSSI signal in order to determine the distance of a zigbee module attached to the arm of the

subject. These results are the first steps towards a totally autonomous system that is an open platform. In

this platform it will be easy to highlight the interaction or the correlation between the different physiological

data and to move the robot properly in case of alert. It is possible to program different services and to

integrate new sensors remotely. This platform can be convenient for developers and researchers involved in

health technology.

1 INTRODUCTION

In recent years the number of elderly people isolated

at home continues to grow. These people often have

need for medical monitoring and have difficulty to

open outward. One of the solutions is to develop

connected medical sensors to remotely monitor

physiological and medical data. The variety of

available physiological sensors allows to consider a

large number of diseases, including non-

exhaustively:

Automatic blood pressure sensors are now

available which send the data, either wired or

wireless to a smartphone or a computer (Isais et al.,

2003).

Any apparatus that measures the concentration of

sugar in the blood are using the same techniques; it

is imperative to remove the patient's blood on a test

strip. Some of these devices can transmit data via a

USB cable or wirelessly.

There are two ways to achieve the pulse

measurement, with an electrocardiogram (ECG) or

an oximeter. The ECG is generally composed of a

transmitter unit includes a belt worn around the

chest and comprising electrodes for sensing the heart

beat and to transmit information to a receiver, which

can be worn the wrist like a watch. Using this

system can be complicated for elderly people. The

alternative is pulse oximeter which measures the

quantity of oxygen in the blood at the finger. The

measurement of the variation of oxygen in the blood

is an indirect indicator of the pulse rate.

Thermometers are classic physiological sensors

widely used at home. But there is an added value to

be able to record temperature regularly for

monitoring the evolution of temperature. Moreover

some new thermometers were recently design that

can measure the temperature with no contact. The

principle is simply to record the infrared radiation

from the heat source and to convert it in

temperature.

The main drawback of these systems is the

constraint to access the data one by one. Based on

this issue, some researchers proposed the project the

E-Health with the purpose to connect several

physiological sensors to one platform. It consists of

a "Cooking Hacks" card, which is used to interface

all medical sensors, associated with either a

"Arduino" card or a "Raspberry Pi" card. On this

platform, it is possible to connect up to nine medical

sensors. Physiological data can be sent via WiFi,

Bluetooth, 3G, GPRS or ZigBee. In addition to

517

Soucies N., Girouard J. and Ouarti N..

Smart Moving Nightstand - For Medical Assistance of Elderly People an Open Project.

DOI: 10.5220/0004915505170522

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2014), pages 517-522

ISBN: 978-989-758-010-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

providing a hardware architecture, the project also

provides all the software part implemented in C + +.

To ensure code compatibility between the two

architectures, developers have chosen to use the

library "ArduPi".

The principal drawback of this project is the

connection with specific sensors that are dedicated

to the project. Moreover, even if the data can be

processed remotely, the acquisition by sensors is

done with wires. These connections not provide a

convenient ergonomics and limit the number of

possible sensors. Another limitation is the absence

of procedure in the case the patient is experiencing a

serious crisis not allowing him to call for help.

Moreover a disease can be detected earlier by the

correlation of different physiological data. An

automatic system allowing to one hand to centralize

and process the data, and to another hand to move to

the patient to observe his condition, would allow to

send an alert to a health’s professional or the family

to alert them. Another possibility is that the patient

can talk with his relatives and health’s professional

to reassure them in case of false alarm. This kind of

task could usefully be done by a robotic assistant.

Today there are many robots that were developed

to assist people.

Some robots are human’s assistant, but are not

natively equipped with physiological sensors. This is

the case of JAZZ robot, whose main application is

telepresence. It may, in some cases, allow a doctor

to visually observe a patient without being

physically present on the site. The robot Kompaï, for

its part is designed to support the elderly in a home

environment. This robot focuses on multimedia

features and non-medical application. ASIMO is a

robot equipped with technologies that give it a

genuinely independent action. Its name is an

acronym for Advanced Step in Innovative Mobility.

Asimo is an autonomous robot capable of

determining his behavior in unpredicted situation.

Thus, it can coexist with humans. Autonomy allows

him to decide to change his path to avoid a collision

with another person. In some circumstances, his

faculties are superior to those of men. For example,

it is able to track multiple conversations

simultaneously (Mutlu et al., 2005); (Sakagami et

al., 2002).

In contrast, other robots are connected to

physiological sensors. This is the case of the robot

RP-Vita Remote Presence is a medical robot mobile

telepresence designed to be used primarily in a

hospital and communicate with medical instruments

connected to it. It helps to have several medical

officers in connection who may have access to all

information on the equipped patients. This

comprehensive platform is exclusively available to

hospitals. It requires that the hospital has to be

equipped with hardware that can communicate with

the robot and the platform is not suitable for home

use. In this category of robots one can quote

HealthBots that is a project aiming to measure some

physiologic data, but the drawback is that the robot

uses some dedicated sensors (Jayawardena et al.,

2010); (Jayawardena et al., 2012). Another

orientation is the one taken by Robo MD which is to

combine the mobility of a Nao robot with sensors

networks. This approach is mainly oriented to

provide an alert in case of falls situations (Van de

Ven et al., 2010).

The aim of this study is to develop a medical

assistant robotics for elderly people. It seems

essential for such an assistant to have physiological

sensors in large numbers. We decided to connect the

physiological sensors with a ZigBee connection, this

strategy has many advantages. Firstly it allows to be

connected wirelessly with the platform, it also

multiplies the number of accessible sensors, and

then it can detect diseases more accurately by the

correlation of physiological signals by identifying

the type of sensor. An assistant must be able to both:

use multimedia resources in an emergency to

communicate with the older person or to observe his

condition visually. Moreover, given the constraints

due to the humanoid form that greatly complicates

the mechanics of a system; we opted for a more

rudimentary design that can be better accepted by

the elderly. This is the concept of "smart moving

nightstand". This platform is developed to allow an

elderly person to be autonomous while being

connected to the outside world (family and doctors).

It would also carry essential items such as

eyeglasses or medications of any user by monitoring

its essential physiological variables. It is a new open

platform for developers interested to compute

physiological data and offers the services of wireless

robotic assistant.

2 SYSTEM OVERVIEW

The system can be divided into 6 main parts (see

Figure 1: Hardware block diagram of the platform.):

• The wireless communications with Zigbee and

WiFi.

• Medical sensors (scalable to the needs of the user)

• The multimedia part that plays the role of user

interface.

• The processor (SOC ARM 32 bit)

HEALTHINF2014-InternationalConferenceonHealthInformatics

518

• The motor control to move the robot.

• The power to the battery and charging station.

In this study we focus on the first two issues.

Figure 1: Hardware block diagram of the platform.

2.1 Concept

The principle behind this intelligent platform is that

it can blend in with its environment. The aim is to

have a "smart nightstand" that can move

independently in a home environment (apartment or

house). Our goal is that this assistant can navigate,

locate and track the user's system when needed. This

medical assistant may allow a relative or a doctor to

remotely monitor the physiologic data of the user

(thermometers, blood glucose meters, blood

pressure, pulse oximeters, etc.). It also helps when

an alert occurs to visually inspect the patient and if

necessary, communicate with him. This platform

should also allow the user to access a number of

multimedia features such as video conferencing,

play multimedia content (movies, music, etc..). The

technological system choices were made with the

idea that this platform must be accessible to the

greatest number of elderly people.

A first prototype has been developed which

embodies the concept discussed. We present here the

different technologies and techniques to meet the

expected functionality.

2.2 Zigbee for Localization

and Sensors Communication

2.2.1 Zigbee

The Zigbee technology, based on the 802.15.4

standard and it works on the same frequency band as

WiFi, 2.4GHz. This technology has the advantage of

having very low power consumption, one have also

the opportunity to significantly expand a mesh

network (65,000 end-devices). It therefore provides

parallel information from many sensors.

Additionally, this technology can be used in

domotic, allowing the platform to control compatible

equipment. Indeed, one of the advantages of ZigBee

is its interoperability. The ZigBee Alliance has

created this standard to create a consistent

communication between multiple devices. This

standard is actually a layer (ZigBee Pro Stack 2007)

which is placed on top of the 802.15.4 MAC layer

handles addresses. It allows to manage the network

(topology, security, communication, etc). The

Zigbee standard formats the messages sent between

the devices. Each device is classified into categories

and sub-category (Cluster) and is able according to

the categories on which it depends to send or receive

specific messages. All these categories are stored in

different standardized norms. For our part, we will

focus especially on the “Home Automation”

standard and the XBEE hardware that is constituted

of many analogic channels. The standard home

automation can control lights, heating, electrical

outlets, smoke detectors or can provide alarms. The

XBEE hardware is interesting to collect analogic

data from different sensors, in our case medical

sensors.

2.2.2 Zigbee for Localization

The localization of the platform in its environment is

achieved by Zigbee. For Zigbee network, XBee Pro

module was used as a system coordinator. Then,

effective way to equip an apartment with Zigbee,

would be to place an electrical outlet with Zigbee on

each power socket. This will allow it to have a large

mesh network covering the whole apartment and

allow it to achieve an effective localization. To

allow a Zigbee module to be wear by the user, the

Zigbee medical solution ZCare of CLEODE was

chosen. This sensor allows can monitor: pulse,

possible falls of the person, and includes a button to

call emergency. Here we can see the interest of

Zigbee with the interoperability of a system

developed by third parties. This solution allows to

locate the person and to perform medical monitoring

at the same time.

The localization of a ZigBee module for its part

can be determined based on the signal power (RSSI:

Receive Signal Strength Indication) sends to other

modules.

The signal strength varies with the distance

(Blumenthal, et al., 2007):

With: PTX = Transmission power of sender,

PRX = Remaining power of wave at receiver,

SmartMovingNightstand-ForMedicalAssistanceofElderlyPeopleanOpenProject

519

GTX = gain of transmitter,

GRX = Gain of receiver,

λ = wave length,

d = Distance betweens sender and receiver,

PREF = power reference (Typically 1mW)

RSSI in dBm.

The RSSI values range between -45 and -100 dBm

and therefore it is possible (for a signal on 100m) to

trace the evolution of the theoretical RSSI function

of distance (Sugano, 2006).

A number of researches have been done on the

indoor localization Zigbee (Lau, et al., 2009)

(Thomas & Ros, 2005). Based on triangulation

algorithms it seems conceivable given the location

of a Zigbee module in a room provided by Zigbee

module with an error margin of 2m.

In this project, the ZigBee can be used in order to

know in what room of the apartment is the user

through a zigbee bracelet, and know where the

platform is. It would be useful to determine how

accurately our platform could determine the position

of the user in the room with the RSSI signal to come

and watch, and whether this accuracy is maintained

in outdoor conditions.

2.2.3 Zigbee for Sensors Communication

Communication with XBee modules on our platform

is done by receiving messages with UART written in

hexadecimal. The X-CTU software is used to

configure the XBee card. There are two possible

modes, the transparent mode and API mode. The

API mode is more indicated for a network or

identification of multiple devices.

In this project we choose to apply the API mode

that will help for the computation of different

physiological signal by identifying which sensor

sends the data. Moreover another type of application

could be the localisation of a lost sensor which can

be an interesting issue with elderly people.

3 PRELIMINARY RESULTS

Some preliminary results were obtained with our

first prototype to show the potential of our approach.

In this section we show the potential of our method

that can either be a tool for localization and for

sending data.

3.1 Distance Estimation with RSSI

It is possible to estimate the distance between two

ZigBee modules with the help of the RSSI signal

(Received Signal Strength Indication). A relation

exist between the value of the RSSI and the distance.

In the first experiment, we decided to estimate this

relation with our specific hardware. We decide to

estimate the distance between the ZigBee module

and our robotic platform.

3.1.1 Static Determination of Distance

(Indoor)

As state in the introduction, RSSI signal is perturbed

by noise but the RSSI signal is more clear and

discriminant at short distance (<1.5m).

We realised the following measurement at

different distances: 5, 10, 20, 30, 40, 50, 70, 100,

130, 160, 200, 250 and 300 cm. These measures

were repeated 10 times at each position to obtain

statistically exploitable data. The mean value

obtained is represented at Figure 2. This study was

conducted in an office room which can be likened to

a domestic environment. It can be observed that the

RSSI curve is approximately bijective, meaning that

the distance can be evaluated. But there is an

exception at 160 cm that can be explained by

occlusion and reflection of the ZigBee wave due to

the objects included in the room.

Figure 2: Static recording of RSSI (indoor).

3.1.2 Dynamic Determination of Distance

(Outdoor)

In this section we tested with two repetitions the

RSSI value in indoor condition when the mobile

platform was moving to the target (i.e. Zigbee

module) to find whether the motion can lower the

accuracy of RSSI signal. We propose a new manner

to assess the quality of the obtained points by

computing the coefficient of determination (R²)

between these points and a logarithmic curve that

fits the points. If the points follow a bijective and

logarithmic function as expected ideally, the R² will

be close to 1. If the R² is further to 1 it means that

the quality is low and certainly the RSSI signal is

perturbed by occlusion, reflection or low intensity

signal. In this experiment the R² is equal to 0.92 (see

HEALTHINF2014-InternationalConferenceonHealthInformatics

520

Figure 3) that is compared to 0.96 in the static

condition.

Figure 3: Dynamic recording of RSSI (outdoor).

This result has to be compared to the RSSI data

obtained in indoor in dynamic condition. We

obtained a R² of 0.87 (see Figure 4). We can

observed that in this case the curve is much noisy

compared to outdoor condition and also compared to

static condition where 10 samples were recorded.

Figure 4: Dynamic recording of RSSI (indoor).

3.2 Send Data from a Sensor

The XBEE technology can be connected to a shield

that sends data to a UART port. We developed a

software that reads the data from this UART and

record it on our platform. As explained before the

user holds a sensor using the home automation

protocol. Moreover we send wirelessly the data of

temperature during a moment when the user grasps

the thermometer. This thermometer is a device that

we developed for the experiment (see Figure 5).

The temperature of the body where captured

(hand temperature) and sent remotely to the robot

(see Figure 6). One can notice the ambient

temperature around 24° C, the progressive increase

around 27°C, the temperature of the hand during the

grasping and a progressive return to the former value

of ambient temperature.

Figure 5: Communication between our robot and the user.

A: Zigbee Thermometer, B: ZCare device, C: User, D:

Zigbee Coordinator, E: Touch screen and F: Robot.

Figure 6: Temperature measurement with ZigBee

transmission.

4 DISCUSSION AND FUTURE

WORK

We propose an open platform allowing to access the

robot at low level. We demonstrated a first proof of

concept by sending remotely temperature data. We

also presented a method to assess the quality of the

RSSI signal. And we showed that outdoor dynamic

measurement is more reliable than indoor

measurement, probably due to occlusion and

reflection of the wave signal. The determination of

the distance between the robot and a wireless sensor

is essential because this can lead to the localization

of the user that wearing the sensor. Another

important issue is the localization of a lost sensor

which can be solved with the same method. Our

ongoing researches are focused on different type of

sensors. The development of specific algorithms

based on the different sensors is the challenge that

we want to promote with the help of the community

of developers and researchers involved in health

technology.

REFERENCES

Blumenthal, J., Grossmann, R., Golatowsk, F. &

Timmermann, D., 2007. Weighted Centroid

Localization in Zigbee-based Sensor Networks.. IEEE

international symposium on Intelligent Signal

SmartMovingNightstand-ForMedicalAssistanceofElderlyPeopleanOpenProject

521

Processing, p. 1–6..

Isais, R. et al., 2003. A Low-cost Microcontroller-based

Wireless ECG-blood Pressure Telemonitor for Home

Care. s.l., Proceedings of the 25th Annual

International Conference of Engineering in Medicine

and Biology Society (Vol. 4, pp. 3157-3160) .

Jayawardena, C. et al., 2012. Design, implementation and

field tests of a socially assistive robot for the elderly:

Healthbot version 2.. s.l., In Biomedical Robotics and

Biomechatronics (BioRob), 2012 4th IEEE RAS &

EMBS International Conference on (pp. 1837-1842)..

Jayawardena, C. et al., 2010. Deployment of a service

robot to help older people.. s.l., In Intelligent Robots

and Systems (IROS), 2010 IEEE/RSJ International

Conference on (pp.pp. 1837-1842)..

Lau, S.-Y.et al., 2009. A Measurement Study of Zigbee-

based Indoor Localization Systems Under RF

Interference.. New York, In Proceedings of the 4th

ACM International Workshop on Experimental

Evaluation and Character.

Mutlu, B. et al., 2005. Perceptions of ASIMO: An

Exploration on Co-operation and Competition with

Humans and Humanoid Robots.. New York, In

Proceedings of the 1st ACM SIGCHI/SIGART

conference on Human-robot interaction (pp. 351-352).

ACM..

Sakagami, Y. et al., 2002. The intelligent ASIMO: System

overview and integration.. s.l., IEEE/RSJ International

Conference on In Intelligent Robots and Systems (Vol.

3, pp. 2478-2483)..

Sugano, M. K. T. O. Y. &. M. M. ,. 5. 0., 2006. Indoor

localization system using RSSI measurement of

wireless sensor network based on ZigBee standard..

Target, Volume 538, p. 50.

Thomas, F. & Ros, L., 2005. Revisiting Trilateration for

Robot Localization.. IEEE Transactions on Robotics ,

21(1), pp. 93-101.

toto, 2013. zigbee page. (En ligne)

Available at: www.zigbee.org

Van de Ven, A. A., Sponselee, A. M. A. & Schouten, B.

A., 2010. Robo MD: a home care robot for monitoring

and detection of critical situations.. s.l., In

Proceedings of the 28th Annual European Conference

on Cognitive Ergonomics (pp. 375-376). ACM..

HEALTHINF2014-InternationalConferenceonHealthInformatics

522