=Paper=
{{Paper
|id=None
|storemode=property
|title=Integrating Human Observations and Sensor Observations – the Example of a Noise Mapping Community
|pdfUrl=https://ceur-ws.org/Vol-640/paper1.pdf
|volume=Vol-640
}}
==Integrating Human Observations and Sensor Observations – the Example of a Noise Mapping Community==
https://ceur-ws.org/Vol-640/paper1.pdf
Integrating Human Observations and Sensor
Observations - the Example of a Noise Mapping
Community
Theodor Foerster, Simon Jirka, Christoph Stasch, Benjamin Pross, Thomas Everding,
Arne Bröring, Eike H. Jürrens
¹[theodor.foerster,jirka,staschc,benjamin.pross,everdingt,arneb]
@uni-muenster.de
²ehjuerrens@52north.org
Abstract. Human observations have the potential to significantly improve the
actuality and completeness of data about phenomena such as noise distribution
in urban environments. The Human Sensor Web aims at providing approaches
for creating and sharing human observations as well as sensor observations on
the Web. One challenge is the integration of these observations for further
analysis. The aspects presented in this paper are examined by the example of a
noise mapping community.
1 Introduction
Noise maps are currently generated out of a sparse measurement basis using
complex processing steps including simulations. Thus, the information provided for a
specific point of interest is a rough estimation based on this sparse measurement
basis. Human observations have the potential to significantly improve the
measurement basis, which supports the EC directive on the assessment and
management of environmental noise [1]. The effort of integrating human observations
as well as sensor observations is subject of the Human Sensor Web, which aims at
providing a full spatio-temporal data coverage on specific phenomena by
incorporating different types of observations. In this context we distinguish between
human observations which are collected by humans directly (such as a textual
description) and human sensor observations which are collected by sensors carried by
or attached to humans (e.g. continuous measurements by carried smart phones).
The Human Sensor Web adopts technology from the Sensor Web, with a strong
focus on the concepts of VGI [2] and the Digital Earth [3]. The challenge attached to
the Human Sensor Web regarding the integration of human observations and sensor
observations has not been described yet. In this paper we will analyse this challenge
based on the example of establishing a noise mapping community.
Section 2 will describe in detail the envisioned noise mapping community and will
show the difference with the established quake catcher network. The challenges for
the Human Sensor Web are described in Section 3. The paper ends with a conclusion.
2 Theodor Foerster et al.
2 Noise Mapping Community
The integration of sensor observations and human observations is exemplified by
establishing a noise mapping community. Mapping noise in urban environments is
motivated by an according EC directive [1]. In this community, which is currently in
preparation at our lab we envisage that noise data is collected and shared by users
through smart phones. In particular, the smart phones are configured to continuously
send current noise measures to the community. These measures include the volume of
noise, its main characteristics (e.g. frequency distribution) and the location, at which
this noise has been measured. By integrating such data, collected by different users, it
is possible to analyze the noise level and to calculate a full coverage of noise
distribution regarding time and space. In this case the person carrying the mobile
phone does not act as a human sensor but rather as a human sensor platform.
These technical measurements can be supported by human observations that are
sent via smart phones or any other kind of web browser. These human observations
may describe the noise intensity and the source producing the noise. By sending such
observations, these users become sensors themselves. This makes it possible to better
interpret the noise map based on the noise measurements (taken by mobile phones).
An overview of the noise mapping community is depicted in Figure 1. This figure
shows that the different types of observations are integrated into the Human Sensor
Web to create and share noise measurements, which can be visualized as a full
coverage map. In the given map example, the noise intensity is visualized from green
(low noise level) to red (high noise level).
Fig. 1. Overview of the architecture of the noise mapping community.
Integrating Human Observations and Sensor Observations 3
Another example of such a community is the quake catcher network 1 [4], which
measures earth quakes through acceleration sensors as built-in in current laptops. The
essential difference of the quake catcher network over the noise mapping community
is that it does not incorporate human observations. This difference is also the
challenge faced by the noise mapping community and is described in Section 3.
3 Challenge
When establishing a Human Sensor Web such as the noise mapping community
one challenge which appears to be essential is the integration of human observations
and sensor observations. This integration is important to provide full coverage data
but it is challenging due to their different nature. While sensor observations usually
are well calibrated and errors can be automatically detected (for instance [5]), human
observations are not quality assured and might be error-prone. Therefore, mechanisms
to automatically detect errors in the data are required. One possibility is to validate
human observations by these sensor observations. For example, noise sensors can be
used to detect erroneous noise observations. Additionally, mechanisms to assign trust
levels to human observers depending on the quality of their previous observations
need to be elaborated [6]. This requires developing validation algorithms that take
into account observations from conventional technical sensor networks but also trust
and metrics for determining whether observations are suitable for the validation
process.
In a second step, established sensor network architectures on the Web such as
OGC’s Sensor Web Enablement framework [7] need to be investigated for further
usage. This especially applies to the scalability of such frameworks. Scalability will
become an important factor, as the amount of noise data might become critical, if the
community becomes intensively used. In this regard, cloud computing can be one
solution to increase scalability of such architectures. Besides that, reducing the
amount of transferred data by the means of aggregation might be a further promising
approach to increase scalability and performance.
When tackling the described challenge of the Human Sensor Web further
challenges need to be addressed, such as (in prioritized order):
Enabling semantics of human observations
Designing ergonomic user interfaces
Investigating and stimulate incentives of people to participate in such a community
Handling of human cognition and resulting uncertainties
Ensuring security and privacy aspects
Handling of unstructured information provided by human observers.
1 Quake Catcher Network website: http://qcn.stanford.edu/
4 Theodor Foerster et al.
4 Conclusions & Outlook
In this paper we introduce the Human Sensor Web for building a noise mapping
community. By integrating sensor observations measured by carried smart phones and
by taking into account (fuzzy) human observations the noise mapping capabilities of
conventional sensor networks can be significantly enhanced.
We identify a series of challenges that will need to be solved before such a noise
mapping system can become functional. Currently, we focus on the flexible
combination of the Human Sensor Web with conventional sources of sensor
measurements as well as geospatial data.
Regarding the challenges described within this paper, we will continue our work
with the design and implementation of the noise mapping community. We expect to
make use of the OGC Sensor Web Enablement concepts to integrate the different
kinds of (sensor) data sources. Based on this initial implementation, further research
topics like incentives for increasing the participation of users, the analysis of data
quality/reliability, security and the handling of unstructured information can be
addressed. Future research will also investigate the benefits of this Human Sensor
Web approach to noise mapping over conventional approaches.
References
1. European Parliament: Directive 2002/49/EC of the European Parliament and of the
Council of 25 June 2002 relating to the assessment and management of
environmental noise, (2002).
2. Goodchild, M.: Citizens as Sensors: the World of Volunteered Geography.
GeoJournal. 69, 211-221 (2007).
3. Craglia, M., Goodchild, M., Annoni, A., Camara, G., Gould, M., Kuhn, W., Mark,
D.M., Maguire, D., Liang, S., Parsons, E.: Next-generation Digital Earth.
International Journal of Spatial Data Infrastructure Research. 3, 146-167 (2008).
4. Cochran, E.S., Lawrence, J.F., Christensen, C., Jakka, R.S.: The Quake-Catcher
Network: Citizen Science Expanding Seismic Horizons. Seismological Research
Letters. 80, 26-30 (2009).
5. Everding, T., Jürrens, E., Andrae, S.: In-stream Validation of Measurements with
OGC SWE Web Services. Second International Conference on Advanced
Geographic Information Systems, Applications, and Services. pp. 93-98IEEE
Computer Society (2010).
6. Bishr, M., Mantelas, L.: A trust and reputation model for filtering and classifying
knowledge about urban growth. GeoJournal. 229-237 (2008).
7. Botts, M., Percivall, G., Reed, C., Davidson, J.: OGC Sensor Web Enablement:
Overview And High Level Architecture. OGC, Wayland, MA, USA (2007).