A PERVASIVE NUTRITIONAL MONITORING AND ADVISE

SYSTEM

NutriMe

Vítor Basto, João Varajão and António Cunha

Department of Engineering, University of trás-os-Montes e Alto Douro, Quinta de Prados, Vila Real, Portugal

Keywords: Nutrition, Pervasive Nutritional Monitoring, Nutritional Advise System.

Abstract: It is well known, widely accepted, scientifically proved and published by major governmental and non-

governmental organizations worldwide (e.g. WHO - World Health Organization), that nutritional

misbehaviour in so called developed countries, is a major cause of diseases, morbidity and death. The

phenomenon is mainly felt in aged populations, but a significant increase has also been detected more

recently in young populations. This paper presents a proposal to tackle serious social and behavioural

problems related to aging and nutrition. NutriMe is presented as a nutritional monitoring and advising

system to help individuals to monitor and correct their behaviours. We also propose NutriMe as the main

component for a public national observatory on nutritional profiles for public health analysis purposes.

1 INTRODUCTION

Several economical, social and cultural factors that

took place in the last decades in developed countries

have strongly influenced human diseases profiles.

Among those factors, globalization and urban

prevalent and increasing lifestyles are worth to

mention. One of the reflexes turning out from those

changes is related to nutritional (mis)behaviours

(Lopes et al., 2006). Although other reflexes are also

subject of research (e.g. smoking, pollution, etc.), we

will not take them into account in our study,

knowing that nutrition is referred as one of the most

important factors.

The outcome of several years of research about

nutritional reflexes on health/diseases, lead to a

present common sense assumption that proper

nutritional monitoring and advise is need, must be

continuous, rigorous and customised for each

individual, according to biological, medical, and life

style parameters (Lopes et al., 2006).

The extent of many harmful reflexes (e.g.

morbidity and mortality) caused by incorrect

nutritional behaviours on health, have been

estimated in several studies. World Health

Organization (WHO) reports that 80% of

cardiovascular disease cases, 90% of diabetes

mellitus type 2 and 33% of all types of cancer could

have been prevented by adopting healthier lifestyles,

which includes correct nutrition, regular physical

activity and non-smoking (WHO, 2006). WHO also

states that: “A change in dietary habits, physical

activity and tobacco control, have a major impact in

reducing the rates of these chronic diseases, often in

a relatively short time”.

Nutritional monitoring and advise is therefore

important both in an individual point of view for

individual behaviours correction, and in a global

point of view, essential for global policy definitions

and for nutritional education planning.

NutriMe, the system proposed in this paper, is a

system designed to tackle the above presented

issues. NutriMe can also be integrated with the

Smart pantry project - Diet module (Alves et al.,

2006), whose features consist of a subset of NutriMe

features in the context of a smart house project

focused on accessibility and inclusion.

In section 2 a general characterization is made of

the Smart pantry project and explained its possible

integration with the NutriMe system, followed by

section 3 which presents the NutriMe system.

Section 4 shows a NutriMe prototype and some

issues related to the distributed data model are

presented in section 5. Finally, conclusions and

future work are referred in section 6.

149

Basto V., Varajão J. and Cunha A. (2008).

A PERVASIVE NUTRITIONAL MONITORING AND ADVISE SYSTEM - NutriMe.

In Proceedings of the International Conference on e-Business, pages 149-154

DOI: 10.5220/0001906001490154

 SciTePress

2 SMART PANTRY PROJECT

Developed countries aged populations is raising

several and severe problems related to home daily

elder people tasks, such as physical access and

handling of home stored goods, nutritional control

and health care, etc. The Smart pantry project

(Alves et al., 2006) was developed with the intention

to fulfil some of those needs. Its main features are

related to stock control, triggering advertises to

replace depleted/missing products, create automatic

customized shop lists, make products physically

accessible to people with reduced mobility,

suggesting menus tailed to user preferences and

medical profiles (diets), all supported by a central

database system (Barrias et al., 2008). Figure 1

presents a modular architecture for the smart pantry

as proposed in (Alves et al., 2006).

Figure 1: Modular architecture for the smart pantry

Source: (Alves et al., 2006).

In this system, users enter products using the

entry module, which automatically identifies the

product, presents its features to the user and asks

him for confirmation/validation. After validation, the

system stores the product according to its storage

conditions (temperature and humidity), size, expiry

dates, etc.

Based on the storage information in the database,

the stock management module manages the

existence of stored products (e.g. expiry dates, so

that they can be disposed or used), a ‘shopping list’

is also presented based on predicted needs, on-line

shopping with supplier’s systems integration

(Cardoso et al., 2007), history of menus, etc.

The storage module consists of a robotic storage

system, adapted to user needs (accessible) and to

available space and conditions. It may be based on

rotating shelves, suspended elements, or other

systems such those employed in large-scale storage

systems.

In order to produce a prototype, two basic

solutions are under development: one thought to be

part of traditional kitchen furniture (a numeric-

control system drawing its inspiration from a bucket

chain); another will be a robotic system of the ‘mini-

warehouse’ variety, to be installed in its own

compartment – the ‘pantry’ (modular systems of

greater capacity).

An extremely important element is the HCI

(nicely) module (human computer interface module)

that will be programmed according to user profiles

and specific needs, so that the system can interact

with each person according to his personal

physical/psychological skills and impairments

(image-based, text-based, voice based, simplified vs.

full featured versions).

The diet module has a set of menus that allow the

counselling of users, accounting for different

criteria, such as the number of people (family size),

the status of the food stock, or the way for eating

adequately.

Smart pantry project diet module features are

considered as part of NutriMe project as the home

based monitoring and advising features. Nutrime

intends to cope with ubiquitous nutritional

monitoring also in other contexts like restaurants,

canteens, bars and pubs, vending machines (food,

beverage, cigars automatic selling machines), etc.

Nutrime extends the concept of nutritional

monitoring to ubiquitous individual monitoring and

advising system.

3 NUTRIME

NutriMe is a distributed software and distributed

data based system that collects nutritional

information from different sources. In addition to

individual nutritional information gathering, it also

links, relates data, ensures consistency and

integration of syntactic and semantic data models

from different sources (food suppliers – home,

restaurants, vending, etc.). NutriMe uses that

information not only for individual real-time

nutritional monitoring and advising, but also for the

purpose of feeding a nutritional national database

(nutritional ‘observatory’). The observatory

database allows for classification, segmentation and

prediction of nutritional profiles. Based on data

analysis and knowledge extraction (data mining)

from this database, public health policies and

strategies can be better supported and deployed.

Figure 2 shows the components of NutriMe

(conceptual view).

ICE-B 2008 - International Conference on e-Business

150

Figure 2: Components of NutriMe (conceptual view).

Each component relies on different technologies

and targets different purposes and features of a

global nutritional monitoring and advising system.

NutriMe conceptual model, its individual

components roles and descriptions, follow:

 Home desktop component, supports all the

features mentioned in the Smart pantry project

– Diet module. The major features specified

are related to user profile characterisation

(age, gender, physical activity profile,

professional profile, medical profile, etc.),

alimentary items characterisation

(identification, common description,

nutritional data composition, etc.), diets,

healthy profiles characterisation, nutritional

reports generation and nutritional advise;

 Mobile device component, implements a mobile

version of the Home desktop component and

features for interoperation and integration with

all the other system components that provide

(or collect) nutritional data for the individual,

like the Home desktop, Restaurant, Automatic

selling machines and Nutritional observatory

components. It is intended to download and

run mobile code made available from the other

system components (e.g. in restaurants),

behave like an extended GUI (graphical user

interface) of those systems and interchange

(collect/provide) information about individual

user consumed items from those system

components. The consumed items identified

and/or described are to be found in the mobile

device database for nutritional monitoring,

processing and advise purposes. The mobile

component is also able to look through

consumed items description provided by the

other system components, map them into the

nutritional characterisation and run the

respective nutritional monitoring, evaluation/

classification and advise algorithms;

 Restaurant component, our concept of a “smart

restaurant” system component includes the

following major features: provide mobile

applications/code to be downloaded into

customer mobile devices allowing for

customer multimedia interactive menu

selection (e.g. food, drink, desert, etc.), for

delivery selection (e.g. tables and location of

the customers to be selected based on

interactive maps), for customer identification

and profile management (e.g. personal data for

invoice and receipts), detailed and electronic

invoice issuing (detailed info about consumed

items) sent to the customer mobile device,

electronic payments, etc.;

 Automatic selling machines (vending)

component, this is a generic component

representing any other system component

providing and/or collecting nutritional

information (e.g. beverage and food automatic

selling machines). Each of these components

must provide mobile code to be run on user

mobile devices or implement a compliant

generic communication protocol profile (to be

defined) for integration with the mobile

device;

 Nutritional observatory component, includes

features of data collection from individual

devices (desktop and mobile devices) and

restaurant systems, data synthesis, nutritional

profiles analysis based on multiple criteria

(e.g. age, location/geography, profession,

etc.), allowing for population risk

classification and evaluation concerning

nutritional behaviours and diseases prevalence

analysis.

NutriMe intends to promote healthy nutritional

behaviours by the means of ubiquitous nutritional

monitoring in an individual and population basis

(reporting individual and population nutritional

warnings). In addition, it is intended to provide

detailed individual nutritional advising (suggesting

detailed meals according to nutritional principles

and user preferences), global behaviour synthesis,

risk evaluation and classification using data mining

techniques.

Figure 3 shows UML (Unified Modeling

Language) Use Cases specification for NutriMe.

It shows the system actors, features, their

interactions and relationships (Fowler, 2003).

A PERVASIVE NUTRITIONAL MONITORING AND ADVISE SYSTEM - NutriMe

151

Food co nsum er

Manual explicit

consumption

istration

Restaurant

mana

Setup restaurant

virtual space

interface

Food

suppliers

stems

Health

observatory

mana

Hea lth

observatory

stem

Nutritional monitoring

ortin

and advisin

Nutritional

misbehaviour

alerts

Food s upplie r

mana

Virtua l

interfaces setu

Vending

machines

mana

Setup vending

machines virtual

interface

Consum er

rofile setu

Nutrit ional data

observator

feed

Group nut r it ional and

health data analysis

and data minin

Restaurant systems

Vending

machine

stems

Automatic

consumption

istration

Physical and health

rofile setu

<<include>>

Diet

rofile setu

<<include>>

Nutrit ionist

Indiv idual nutritiona l a nd

health data analysis and data

minin

<<ext end>>

Figure 3: NutriMe Use-Cases Diagram.

Use Case “Consumer profile setup” allows the

user to provide his personal data (e.g. gender, age,

height, weight) to the system, defining nutritional

profiles, diets and preferences (e.g. vegetarian,

athlete, diabetic, etc).

Use Cases “Nutritional misbehaviour alerts”

and “Nutritional monitoring reporting and

advising” allows for automatic warnings from

NutriMe towards the user, allows the user to access

statistics about his behaviour and provides

nutritional advising. In addition, NutriMe provides

extra nutritional expertise and tools for knowledge

extraction (clustering, classification, prediction)

from monitoring data for nutrition professionals

(“Nutritionist”) decision support in “Individual

nutritional and health data analysis and data

mining” Use Case.

An observatory component of NutriMe is also

feed by nutritional monitoring data (Use Case

“Nutritional data observatory feed”) for “Group

nutritional and health data analysis and data

mining” purposes.

“Virtual interfaces setup” Use Case supports

user interface, data and functional integration

features between NutriMe and “Food supplier

systems”. For instance, it is possible for “Food

consumers” to issue meal orders from NutriMe that

are transmitted electronically and processed by

“Restaurant systems”. When “Food consumers”

interact with “Food supplier systems” through

NutriMe “Automatic consumption registration” is

possible. “Automatic consumption registration” is

the preferred way of nutritional data gathering in

NutriMe, but “Manual explicit consumption

registration” by “Food consumers” are also

available in NutriMe.

NutriMe simplified data model is presented in an

Entity-Relationship diagram in Figure 4. The

conceptual data model will be deployed as a

distributed relational database system as detailed in

section 5. As a distributed software and distributed

data based system that collects nutritional

information from different sources and

heterogeneous technologies, concerns of

consistency, correctness, integration of syntactic and

semantic data models are to be considered.

Figure 4: NutriMe Entities-Relationships Diagram.

An innovative implementation of NutriMe

involving flexible and adaptive information systems

integration based on desktop and mobile devices is

presented in section 4.

4 PROTOTYPE

Inclusion and accessibility have been major

concerns during all development life cycle of

NutriMe. Software design and technology selection

were essential to construct a solution that promotes

device and software mobility, usability, high

availability and continuous monitoring.

Mobile technologies such mobile phones/PDAs

proved recently to ease cumbersome daily task.

ICE-B 2008 - International Conference on e-Business

152

Recent devices reached significant processing

storage and communication capabilities, allowing

increasing features and new applications. The

massive adoption of this kind of devices is

considered today one of the greatest success of

information and communications technologies

market acceptance (Paes, 2006).

The above mentioned technical factors, growing

market acceptance, easy of use and ubiquity of this

kind of mobile technologies, lead us to select this

computing and communication technologies as the

basic support for the main NutriMe component.

Figure 5 shows the technological solution

proposed for NutriMe system main components,

including physical/logical elements, interoperation

and interactions.

Figure 5: Technical scenario for NutriMe mobile

component prototype.

NutriMe software (mobile and desktop)

components are hosted in PC like and mobile

devices (e.g. mobile phones/PDAs).

NutriMe desktop and mobile versions have been

developed with Microsoft .NET 2005 Framework

and Microsoft .NET Compact Framework

respectively. Microsoft SQL Server 2005 and

Microsoft SQL Server 2005 Compact were de

DBMS (database management systems) adopted for

desktop and mobile support respectively. Data

replication, integrity and synchronisation are

supported by both mentioned DBMS (master-slave

data replication and synchronisation) which is the

most common strategy adopted for similar scenarios.

This strategy shows optimal trade-offs concerning

quality of service (availability, punctuality, etc.) and

resources consumption/usage (storage, processing

and communication) within strongly connected,

weakly connected and connectionless situations.

Support of multiple communications technologies

(GPRS, UMTS, Wi-Fi, Bluetooth), actually quite

common in mobile devices, enforces applications

robustness allowing not only for horizontal but also

vertical handoffs.

Next we present some prototype features and

graphical user interface of NutriMe. User personal

profile form is presented in Figure 6 (left picture) for

user characterisation features of Definition of

Personal and Nutritional Profiles Use Case. And

interactive consumption registration form is also

presented in Figure 6 (right picture) for

Consumption Monitoring Use Case features.

Figure 6: NutriMe personal profile form and Interactive

consumption registration form.

NutriMe user interface also provides friendly

graphical classifications of nutritional and health

ratios, such as the BMI - body mass index (Figure

6), which are subject of time evolution graphical or

text based presentation and analysis.

Supporting distributed relational database

technologies and strategies for NutriMe desktop and

mobile prototypes are presented in section 5.

5 DISTRIBUTED DATA MODEL

Available commercial DBMS (e.g. Microsoft SQL

Server) allow for flexible data management,

enhancing application robustness, high availability

and performance, being especially important in

distributed systems involving mobile devices with

non continuous communication connectivity.

Replication strategies are usually adopted (and very

effective) to overcome periods of non connectivity.

However, data replication needs specific

mechanisms for integrity enforcement among the

copies spread throughout the several components of

a distributed system (e.g. mobile and desktop

equipment). While subject to isolated processing

during non connectivity periods, data replicas in

different processing devices can evolve to divergent,

eventually inconsistent states. Automatic

A PERVASIVE NUTRITIONAL MONITORING AND ADVISE SYSTEM - NutriMe

153

mechanisms must be activated to rollback or roll-

forward the replicas to get the overall distributed

system back to a consistent state. The mechanisms

adopted in our system for this purpose are based in

the two-level master-slave transaction model (Liu et

al., 1999).

One level is ruled by desktop workstations, the

second level is ruled by the mobile devices (one or

more devices). Data convergence actions take place

when a communication channel is available between

desktop equipment (master replicas) and mobile

devices (slave replicas).

First, the master (re)executes all transactions

made available by the slave, corresponding to all the

actions performed by the slave during non

connectivity periods. Second, the master notifies the

slave the successful reconciliation transactions to be

committed. Inconsistent transactions are tracked

back until its root (causal graph node), and undone

until the overall distributed system gets into a

consistent status.

The presented two-level master-slave replication

strategy seemed to us a suitable solution for the

NutriMe system. It revealed optimised trade-offs

according to quality of service and resource usage,

in scenarios similar to ours.

6 CONCLUSIONS

This paper presents a software based system –

NutriMe, which addresses ubiquitous nutritional

monitoring and advising supported by several

cooperating software components distributed

through desktop and mobile devices, databases and

applications. The system intends to introduce

increasing computational and communication

pervasive features, assisting in simple daily tasks

with low or non intrusive reflexes (low or no

perception or interaction required from users).

Because aged populations are common in developed

countries, and they have usually special needs,

concerns of accessibility and usability have been of

major importance in the design and implementation

of NutriMe prototype. Individual and public interest

(e.g. health, economic) of this kind of systems are

worth to mention. They constituted our first interest

and justified further work and research on this topic.

Future work milestones include information

gathering and processing in public spaces context,

multimodal interfaces for impaired people,

introduction of data mining techniques for

customised individual user advise, extraction and

analysis of general nutritional tendencies and

patterns.

REFERENCES

Lopes C., Oliveira A., Santos A., Ramos E., Severo M.,

2006. Consumo alimentar no Porto. Porto: Serviço de

higiene e Epidemiologia - Faculdade de Medicina do

Porto.

WHO, World Health Organization. 2003. Global strategy

on diet, physical activity and health. Available:

URL:http://www.who.int/hpr/NPH/docs/gs_global_str

ategy_general.pdf

Alves J., Cunha A., Trigueiros P., Barroso J., Cruz J.,

2006. Sistema de gestão inteligente de uma despensa.

In DSAI 2006, Proceedings of Conference Software

Development for Enhancing Accessibility and Fighting

Info-exclusion, Vila Real.

Barrias A., Cunha F., Varajão, J., Cunha A., 2008. Smart

Pantry: specification of a management system using

UML. In IADIS e-Society 2008, Proceedings of IADIS

International Conference e-Society 2008, Algarve.

Fowler K., 2003. UML distilled: a brief guide to the

standard Object Modeling Language, Addison-

Wesley. 3

edition.

Liu P., Ammann P., Jajodia S., 1999. Incorporating

Transaction Semantics to Reduce Reprocessing

Overhead in Replicated Mobile Data Applications. In

ICDCS' 99, Proceedings 19th IEEE International

Conference Distributed Computing Systems, Austin.

Cardoso J., Jaulino R., Varajão J., Fernandes V., Cunha

A., 2007. Smart pantry: integration models with

electronic commerce solutions, In DSAI 2007,

Proceedings of Conference Software Development for

Enhancing Accessibility and Fighting Info-exclusion,

Vila Real.

Paes C., Moreira F., 2006. Aprendizagem com

Dispositivos Móveis: Aspectos Técnicos e

Pedagógicos a Serem Considerados Num Sistema de

Educação.URL:http://grace.tech-x-pert.org/blogs/

files/2007/06/aprendizagem-com-dispositivos-moveis-

na-aspectos-tecnicos-e-pedagogicos-a-serem-

considerados.pdf

ICE-B 2008 - International Conference on e-Business

154