FUNCTIONAL INTEGRATION FOR THE OBSERVATION WEB

Sven Schade, Frank Ostermann and Laura Spinsanti

European Commission – Joint Research Centre, Institute for Environment and Sustainability, Ispra, Italy

Keywords: Sensor web, Semantic integration, Interoperability, Algebra, Functional programming, Forest fire.

Abstract: The integration of sensor-based information has been a research topic for many years. International

standards, such as the Sensor Web Enablement (SWE) suite of specifications that will soon be released in its

second version, have been developed, and major syntactical and structural challenges have been overcome.

Solutions for addressing semantic aspects of interoperability have been suggested, but mature applications

are still missing. The advent of user generated content for the geospatial domain, Volunteered Geographic

Information (VGI), which can be considered as readings of virtual sensors, makes it even more difficult to

establish formal systems for the combination of information that is based on heterogeneous sensing

methods. This paper proposes a novel approach of integrating conventional sensor information and VGI,

which is exploited in the context of detecting forest fire events. In contrast to common logic-based semantic

descriptions, we present a formal system using algebraic specifications as a more elegant, illustrative and

straight forward solution.

1 INTRODUCTION

The integration of sensor-based information has

been a research topic for many years (Schade, 2005;

Sheth et al., 2008; Janowicz et al., 2011). Standards,

such as the Sensor Web Enablement (SWE) suite of

specifications that is currently under major revision

(Bröring et al., 2011), help to establish syntactical

and structural interoperability. This includes

possibilities for integrating information produced by

physical sensors and by environmental simulations

(Maué et al., 2011).

Several approaches for addressing semantic

interoperability have been proposed. Most notably,

the W3C recently released a sophisticated ontology

on observations and measurements (Janowicz and

Compton, 2010), and a light-weight approach for the

semantic enablement of the Sensor Web has been

proposed (Janowicz et al., 2011). Still, mature

At this stage (August 3rd, 2011), the SWE Common 2.0

data model already became a standard and the second

version of Observations and Measurement is close to its

official release, while SensorML 2.0 is still under

debate. With respect to service specifications, the SWE

common model and the Sensor Observation Service

(SOS) are available as standards, while discussions on

the Sensor Planning Service (SPS) are still ongoing.

solutions that illustrate the use of such work for

solving challenges of geospatial information

integration are missing.

During the past decade, we have witnessed a

surge of geographic information provided by the

public to the public via the internet. Resembling

virtual sensors, citizens provide this Volunteered

Geographic Information (VGI) (Goodchild, 2007)

through the web by posting images or videos (e.g. on

Flickr or YouTube), blogging or micro-blogging

(Twitter), surveying and updating geographic

information (OpenStreetMap), or playing games

(FourSquare). Considering the increase in mobile

internet access through smartphones and the number

of (geo)social media platforms, we can expect the

amount of VGI to continually grow in the near

future. This new wealth of VGI has several

advantages over the traditional, authoritative

gathering, maintaining and disseminating of

geographic information. First, it is more up-to-date,

because a larger number of ’surveyors’ reports new

information or changes to existing information in

near real-time. Second, VGI can be very rich in

content, providing already pre-processed

information instead of raw data. As information

498

Schade S., Ostermann F. and Spinsanti L..

FUNCTIONAL INTEGRATION FOR THE OBSERVATION WEB.

DOI: 10.5220/0003698404980504

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (SSW-2011), pages 498-504

ISBN: 978-989-8425-80-5

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

portals such as EyeOnEarth

successfully illustrate,

the provided information often complements the

data coming from traditional sensor networks. The

EyeOnEarth portal, which is hosted by the European

Environmental Agency, provides access to

measurement stations in air and water, but also

reports on air quality and water quality that have

been generated by laymen.

However, there are clearly several problems

associated with VGI. The technology-driven

development leads to frequent changes in the data

structure, since new platforms emerge, old ones

disappear, and prevailing ones modify their user and

programming interfaces. Further, VGI frequently is

rather unstructured in nature, and quality control

proves difficult. Even comparatively well-structured

and quality-controlled platforms such as

OpenStreetMap have to deal with these issues.

Therefore, the integration of VGI with existing

sensor networks and spatial data infrastructures is a

challenging task. This increased diversity of

information channels and provided messages makes

it even more difficult to establish formal systems for

information combination. A conceptual solution of

using SWE for integrating VGI with the Sensor Web

has been suggested for creating a (more general)

Observation Web (DeLongueville et al., 2010), but

again semantic aspects have not been considered

explicitly.

In this paper, we propose a novel approach to

integrate conventional sensor information and VGI.

Contrary to common logic-based approaches, we

base our developments on another formalization

paradigm from software engineering: algebraic

specification (Ehrich and Mahr, 1985).

The remainder of this paper is organized as

follows. We briefly illustrate the proposed solution

to sensor integration in the next section, while we

argue for the use of algebraic specifications for

ontology engineering and present related work in

section 3. The formalization of our approach to

sensor integration is provided in section 4. This first

application of this approach is based on several

assumptions and simplifications, which are

discussed in section 5. We conclude the paper with a

summary of our main findings and an outline of

future work. Throughout the paper, the VGI use case

of forest fire detection serves as example.

Official Web portal available at

http://www.eyeonearth.eu/ (last accessed on August 3

2011)

2 DESIGN OF THE

INTEGRATION APPROACH

We consider the application of specific, stepwise

processing of a given raw data set as a core

principle. The different layers of value-added

information can be illustrated as in Figure 1, where

the centre represents the initial content and each

additional surrounding layer represents the results of

one processing step. For example, the raw data

might be air temperature measures (for intervals of

one minute), and the first processing step might

provide daily averages, the next weekly averages,

etc. (Figure 1 a). We may also think of data coming

from different sources, for example measured by

diverse sensor networks, such as air temperature,

wind-direction, cloud-cover and humidity values. In

this case, the first processing step might be a merger

of pieces of information into a complex measure as

the fire risk index (Figure 1 b).

Figure 1: Layers of value-added information, a) averages

of air temperature; b) fire risk index.

Alternatively, contents provided by two different

sources, e.g. satellite images from the MODIS

satellite (a conventional physical sensor) and VGI

posts on Flickr (http://flickr.com), may be provided

separately. As the information gets processed,

resulting layers might overlap. For example, first

MODIS images could be analyzed for temperature

hotspots, and some of these hotspots might be

categorized as forest fires. At the same time, VGI

might be analyzed for hotspots as well. In social

media, these hotspots could be purely thematic in

nature, such as an increase of words like ‘fire’ in

messages, but in the case of sufficiently accurate

VGI, the hotspots could correspond to spatio-

temporal clusters (Spinsanti and Ostermann, 2011)

and subsequently, some of these hotspots might also

be categorized as forest fires (Figure 2). Notably, we

can arrive at the same kind of (forest fire) event

using different information channels.

FUNCTIONAL INTEGRATION FOR THE OBSERVATION WEB

499

Figure 2: Detecting forest fires using MODIS and VGI.

It is worth noticing the following characteristics

of this approach:

1. We can only move from the inner layers to the

outside.

2. The information becomes more specialized with

distance to the centre, i.e. application specific

context is introduced increasingly.

3. Information can only be integrated on a shared

layer.

In section 4 we will use these characteristics,

together with an ontology for observations, to

formalize a system for information integration.

Before, i.e. in the next section, we discuss the use of

algebra as a tool for (formal) system engineering.

3 ENGINEERING FORMAL

SYSTEMS WITH ALGEBRAS

Ontologies have been suggested as the basis for

semantic interoperability of information systems

(Guarino, 1998). Following Guarino’s

characterization of an ontology (in the Artificial

Intelligence sense), as an “engineering artifact,

constituted by a specific vocabulary used to describe

a certain reality, plus a set of explicit assumptions

regarding the intended meaning of the vocabulary

words” (Guarino, 1998), assumptions can be stated

using any formal theory. We use algebraic

specifications of Abstract Data Types (ADTs)

(Sommerville, 2007), in which the (algebraic) theory

of an ADT describes its abstract behavior, whereas

models are given by concrete data types (Ehrich and

Mahr, 1985). In other words, we use ADTs together

with their (algebraic) specification as ontologies.

Additionally, we provide (algebraic) functions to

define transitions between ADTs. The resulting

formal system will implement the integration

approach that has been outlined above.

The decision to use an algebraic approach

compared to common approaches, which apply

Description Logics or First-Order Logic, for

ontology engineering depends on the intended goals.

In our case, we face a data integration challenge

involving classical sensors and VGI. Encapsulation,

as a main feature of ADTs (Sommerville, 2007),

provides us with the required abstraction

mechanisms. The functional equations that are used

in algebraic specification can be directly applied for

mapping from sensor- and VGI-specific models to

an integrated theory of observations. In order to

support clean ontology engineering, the observation

theory can even be aligned with an upper-level

ontology, as we will see in our examples in section

4. These two possibilities provide clear benefits

compared to logic-based approaches, which are for

example strong in instance re-classification.

The above mentioned principles relate closely to

concepts of No-SQL data bases in information

science (Agrawal et al.,2008). Moreover, the history

of using algebraic specifications and functional

programming for the geospatial sciences dates back

at least twenty years. It has been first suggested for

user interface design of Geospatial Information

Systems (GIS) (Kuhn and Frank, 1991) and soon has

been extended to conceptual modeling (Car and

Frank, 1995), including the principle of

measurement-based GIS (Goodchild, 1999). The

explicit use of this approach for (spatio-temporal)

ontology engineering dates back at least five years

(Schade et al., 2004; Frank, 2007; Kuhn,2007).

Recent works direct these ontology developments to

sensing (Kuhn, 2009) and data integration (Schade,

2010).

4 FORMALIZATION OF THE

INTEGRATION APPROACH

As indicated above, we build our formal system for

integrating sensor data using algebraic

specifications; in particular, we use the functional

programming Haskell (Peyton, 2003) for

formalization. As we present the solution directly in

an executable language, this corresponds also to the

implementation of the desired system.

4.1 Algebraic Specification of the

Observation Ontology

Before we can define the transitions between layers

(by functions) and introducing data types for

intermediate and final integration results, we have to

establish an ontology that captures the core

principles of observation, i.e. the inner layer of the

diagrams presented in the previous section. This

should not only cover physical sensing procedures

KEOD 2011 - International Conference on Knowledge Engineering and Ontology Development

500

and environmental simulation, but also VGI.

Since a basic observation algebra is already

available (Kuhn, 2009), we can re-use ADTs, such

as a construct for measurement values, i.e. for raw

data that is the result of a measurement (in Haskell,

algebraic data types are introduced by specifying

constructor functions. The keyword ‘data’ is

followed by the type name, an equal sign and the

constructor function(s). The first element of this

function is its name. ‘

|’ is used for separating

multiple constructer functions for a single ADT):

data Value = Boolean Bool |

Count Int |

Measure Float Unit

| Category String

In addition, we induce an ADT for representing

user generated content (UGC), such as a text

message, a photo or a audio-video. For simplicity

reasons, we include photos and videos by reference,

i.e. using http URIs (Coates et al., 2001):

data UGC = Message String |

Photo URI | Audio URI

| Video URI

An observation consists of either a measurement

value or user generated content, each combined with

a position and time. We thus extend the basic

observation data type with a special construct for

VGI:

data Observation = Measurement

Value Position

ClockTime |

VGI UGC

Position

ClockTime

For later implementation of our example, we add

a construct for MODIS satellite images. Again for

simplicity reasons, we define the image as a list of

observations, where each observation represents a

pixel of the image with its associated values (we use

Haskell type synonyms to give previously defined

data types a new name.):

type MODISimage = [Observation]

4.2 Generating

Value-Added Information

Now, as the foundations are available, we focus our

attention on the transitions from the raw data

(innermost layers in figure 1), to any added value,

i.e. processed information (outer layers the

diagrams). The forest fire example serves for

illustrations.

We introduce constructs, which represent the

results of a transition (the next outer layer in the

diagram), and functions, which represent the actual

transition between two layers, such as the creation of

hotspots. In analogy to the observations ontology

above, the former can be seen as ADTs, which

provide unique entry points to each layer, such as

hotspots and forest fires. They thereby encapsulate

the manifold possibilities in which a single instance

might have been produced.

As a first example, we create a type for capturing

hotspots as a specific kind of value-added

information. Each hotspot represents a set of

observations:

type Hotspot = [Observation]

As an additional benefit of using an executable

functional programming language for formalization,

all Haskell build-in operations are available for

manipulating hotspots (as we will see later). We

might define additional operations on hotspots at

will, but for the moment, we concentrate only on the

function for transitions, i.e. on that function, which

generates hotspots from the information that is

available on lower layers. We define a constructor

function for hotspots in a re-usable manner, i.e. in a

way that we can create hotspots out of any

observation collection, independent of the nature of

the observation. Such reusability is a benefit but is

not mandatory for each transition function.

For us, a hotspot is characterized by high spatio-

temporal density, and may optionally include a

'filter', such as a specific threshold for a measure or a

specific category. We specify the desired minimum

density in space or time, and an optional filter as

parameters of the hotspot function; we omit the

implementing algorithm here, as this is out of the

scope of this paper (Haskell functions can be

specified using signature declarations. Here, a

function name is followed by ‘::’ and a list of

parameters separated by ‘->’. The last parameter of

a signature stands for the output. Parentheses can

be used to include functions, such as a filtering

function, as parameters.):

hotspots :: Value -> Value ->

([Observation] ->

Observation ->

[Observation]) ->

[Hotspot]

We follow a similar principle for the next (the

forest fire) layer. We use the ‘data’ construct of

Haskell for introducing an ADT, because we want to

define special functions on this data type and

FUNCTIONAL INTEGRATION FOR THE OBSERVATION WEB

501

because we intent to use class instantiation as a tool

for ontology alignment later on (see section 4.4):

data ForestFire = ForestFire

Hotspot

Next, we introduce a function to describe the

transition from the lower layers, i.e. from a

collection of hotspots to a collection of forest fires.

In the example, we categorize all hotspots as forest

fires. The implementation also shows the application

of common Haskell functions on the ‘Hotspot’ ADT.

We use the ‘map’ operator, which applies a function

(here the Forest Fire constructor function) to all

elements in a list:

forestFires :: [Hotspot] ->

[ForestFire]

forestFires hs = map ForestFire

e can also use this example to show how specific

operations, so called observers can be added to a

(newly defined) ADT. The signature of a co-

occurrence function for forest fires, which returns a

list that contains all collections of forest fires, which

overlap in space and time may be defined as:

ffCoOccurrence :: [ForestFire] ->

[[ForestFire]]

4.3 Common Access to Value-Added

Information

The approach outlined above provides a

straightforward solution to the integration problem.

Information from two separate sources, as for

example illustrated for MODIS and VGI in Figure 2

(above), may be merged as soon as the derived

information can be provided via a shared ADT.

Considering our example, this would be the data

type representing forest fires. Co-occurrences of

forest fires can now be calculated without required

knowledge about the sources that lead to the

information about a particular fire.

If we would for example have all MODIS and

VGI data available for the 2010 forest fire season in

France, then we could find out how many forest fire

events were detected by both sources, and how many

were detected by only one. Such additional

information, which can be derived from using the

two sensor sources, MODIS and VGI in this case,

can be used to further calibrate the involved sensors

and thus to improve measurement quality (Spinsanti,

and Ostermann, 2011).

For real-time applications this means that we

become able to process (and in particular compare)

information about events, such as forest fires,

seamlessly, i.e. without any artificial barriers that

might have been imposed by diverse use of sensors

or user generated content. In this sense, we solved a

semantic integration issue in the Observation Web.

It is worth noticing that we remain able to trace

back the generation process of a data set. Having co-

occurring forest fires identified, we can reveal if a

particular instance has been created from a MODIS

image or from VGI by uncovering the used

constructor function. Among other information, we

may get to know if statements of two separate

sources confirm or contradict each other.

4.4 Alignment with DOLCE

In order to further specify the semantics of forest

fires (as events), we consider an alignment with

Foundational Ontology. Based on previous

experience, we select Descriptive Ontology for

Linguistic and Cognitive Engineering (DOLCE)

(Masolo et al., 2003) as the best candidate. In

DOLCE, events are defined as special kinds of

perdurants, which extend in time by accumulating

different temporal parts. Each perdurant can be

associated with a duration in which it occurred.

Again following Kuhn (Kuhn, 2009), DOLCE

concepts can be formalized as follows (the ‘class’

construct in Haskell introduces type classes for a set

of (abstract data) types sharing some behavior. Type

classes are named, followed by a parameter for the

types belonging to the class. A ‘where’ clause

introduces a block of functions, which capture the

shared behaviors of the type class):

class PERDURANTS perdurant where

getDuration :: perdurant ->

Duration

class PERDURANTS event => EVENTS

event

The alignment of the forest fire ADT can be

achieved by instantiating the type classes introduced

above; we omit the implementation of the duration

function:

instance PERDURANTS ForestFire

where

instance EVENTS ForestFire

KEOD 2011 - International Conference on Knowledge Engineering and Ontology Development

502

5 DISCUSSION OF THE

INTEGRATION APPROACH

By using concatenated functions as opposed to

(description) logic constructs, the proposed

integration approach follows the idea of

algebraically specifying GIS and the principle of

measurement-based information systems

(Goodchild, 1999). The advent of No-SQL data

bases (Agrawal et al., 2008) indicates a trend of such

solutions even for mainstream IT. On top of these

known concepts, we extend the observations

ontology with user generated content (VGI in our

case) and we apply the approach to the data

integration. The encapsulation by abstract data types

and the use of functional equations as transformation

mechanisms are powerful characteristics of this

algebraic approach. It has the potential to solve

many of the semantic interoperability problems,

which continue to grow with the current trend of

user generated content and the ongoing interests for

information integration for quality improvement.

Although promising, algebraic solutions are still

rare in the Semantic Web and they miss a support in

the context of geospatial web applications. This

results in some difficulties, such as (i) missing tool

support for input and out put data management for

the web; (ii) rare examples that illustrate the

combined use of logic-based and algebraic

specifications for knowledge engineering; and (iii)

disconnectivity between the algebraic specification

and the Semantic Web community. Research in this

direction should be extended in order to avoid

focusing solely on logic-based efforts towards a

Semantic Sensor Web and running the risk of

stagnation.

The presented work lays the function for further

investigations. For example others already suggests

using Haskell for common error modeling and

elegant propagation (Navratil et al., 2008), but

further examinations are required in this field,

especially exploiting the interplay between the

‘quality’ of VGI and of classical measurements.

Another open issue is the investigation of place.

Our approach currently assumes that geospatial

components are introduced in form of geographic

locations, with precise geometric attributes. VGI

however, might merely include vague place names

instead of precise locations, introducing uncertainty

on the geographical side. The interplay between

place and location should be considered as a major

aspect to improve the suggested integration

approach.

The presented algebra does also not explicitly

account for (spatio-temporal) level of detail, i.e.

scale. Here lies a major area for improvement,

because different information sources are often

captured on diverse scales. A functional approach

eases scale changes, but still more practical

examples, have to be explored.

6 CONCLUSIONS AND

OUTLOOK

We have established a formal system of the semantic

integration of observation-based information and

showed a successful approach to the challenge of

integrating them in a forest fire scenario. This

illustrates the functionality that we envision for a

future Observation Web. The proposed approach

still requires maturity. Therefore, we plan to

experiment with a more detailed workflow for

processing forest fire information, as for example

presented in (Ostermann and Spinsanti, 2011). In

addition, more cases will be investigated in order to

show that our general assumptions hold.

As the current implementation used only parts of

a previously developed ontology for observations in

Haskell, we are still investigating complete re-use.

This should be established next and is a short term

goal.

Our intermediate goals are (i) the examination

regarding an integrative quality model (including

propagation) for classical sensor based information,

results of environmental simulations and VGI; and

(ii) experimentations on scale transitions.

ACKNOWLEDGEMENTS

This work was partially funded by the exploratory

research project "Next Generation Digital Earth:

Engaging the citizens in forest fire risk and impact

assessment" from the Institute for Environment and

Sustainability of the European Commission – Joint

Research Centre and under the European

Community’s Seventh Framework Programme

(FP7/2007-2013), Grant Agreement No. 284898

(ENVIROFI project).

REFERENCES

Schade, S.: Sensors on the Way to Semantic

Interoperability. GI-Days 2005: Geo-Sensor Networks

FUNCTIONAL INTEGRATION FOR THE OBSERVATION WEB

503

- From Science to Practical Applications, ifgi-Prints

Bd. 23. Münster (2005).

Sheth, A. et al.: Semantic Sensor Web. IEEE Internet

Computing (2008).

Janowicz, K. et al.: A RESTful Proxy and Data Model for

Linked Sensor Data. International Journal of Digital

Earth (2011, in print).

Bröring, A. et al.: New Generation Sensor Web

Enablement. Sensors 2011, 11(3) (2011).

Maué P. et al.: Geospatial Standards for Web-enabled

Environmental Models. Internal Journal for Spatial

Data Infrastructures Research (IJSDIR), 6 (2011).

Janowicz, K. and Compton, M.: The Stimulus-Sensor-

Observation Ontology Design Pattern and its

Integration into the Semantic Sensor Network

Ontology. 3rd International workshop on Semantic

Sensor Networks 2010 (SSN10) (2010).

Goodchild, M. F.: Citizens as voluntary sensors: spatial

data infrastructure in the world of Web 2.0.

International Journal of Spatial Data Infrastructures

Research, 2 (2007).

DeLongueville, B. et al.:Digital earth’s nervous system for

crisis events: real-time sensor web enablement of

volunteered geographic information. International

Journal of Digital Earth ,3(3) (2010).

Ehrich, H.-D. and Mahr, B.: Fundamentals of Algebraic

Specification 1: Equation and Initial Semantics.

Springer (1985).

Spinsanti, L. and Ostermann, F. O.: Retrieve Volunteered

Geographic Information for Forest Fires. Proceedings

of the 2nd Italian Information Retrieval (IIR)

Workshop (2011).

Guarino, N.: Formal Ontology in Information Systems.

International Conference on Formal Ontology in

Information Systems, FOIS’98, Trento, Italy (1998).

Sommerville, I.: Software Engineering. Pearson

Education Limited (2007).

Agrawal, R. et al.: The Claremont report on database

research. SIGMOD Record (ACM) 37(3) (2008).

Kuhn, W. and Frank, A. U.: A Formalization of Metaphors

and Image-Schemas in User Interfaces. In Cognitive

and Linguistic Aspects of Geographic Space, NATO

ASI Series, Dordrecht, The Netherlands, Kluwer

Academic Publishers (1991).

Car, A. and Frank, A. U.: Formalization of Conceptual

Models for GIS using Gofer. Computers,

Environment, and Urban Systems, 19(2) (1995).

Goodchild, M. F.: Measurement-Based GIS. International

Symposium on Spatial Data Quality, Hong Kong,

Department of Land Surveying and Geo-Informatics,

Hong Kong Polytechnic University (1999).

Schade, S. et al.: Comparing Approaches for Semantic

Service Description and Matchmaking. 3rd

International Conference on Ontologies, DataBases,

and Applications of Semantics for Large Scale

Information Systems (ODBASE) (2004).

Frank , A.: Towards a Mathematical Theory for Snapshot

and Temporal Formal Ontologies. AGILE 2007,

Aalborg, Denmark (2007).

Kuhn, W.: An Image-Schematic Account of Spatial

Categories. Spatial Information Theory, 8th

International Conference, COSIT 2007. Melbourne,

Australia (2007).

Kuhn, W.: A Functional Ontology of Observation and

Measurement. GeoSpatial Semantics ·Third

International Conference (GeoS) (2009).

Schade, S.: Ontology-Driven Translation of Geospatial

Data. AKA, Heidelberg, Germany (2010).

Peyton Jones, S.: Haskell 98 Language and Libraries - The

Revised Report. Cambridge University Press,

Cambridge (2003).

Coates, T. et al.: URIs, URLs, and URNs: Clarifications

and Recommendations 1.0 (2001).

Spinsanti, L. and Ostermann, F. O.: Validation and

Relevance Assessment of Volunteered Geographic

Information in the Case of Forest Fires. 2nd

International Workshop On Validation Of Geo-

Information Products For Crisis Management, Ispra,

Italy (2011).

Masolo, C. et al.: D18 - Ontology Library (final version).

Deliverable of the WonderWeb Project (2003).

Navratil, G. et al.: Lifting Imprecise Values. 11th AGILE

International Conference on Geographic Information

Science (2008).

Ostermann, F. O., Spinsanti, L.: A Conceptual Workflow

For Automatically Assessing The Quality Of

Volunteered Geographic Information For Crisis

Management. Proceedings of AGILE Conference

(2011).

KEOD 2011 - International Conference on Knowledge Engineering and Ontology Development

504