Research Article

International Journal of Distributed

Sensor Networks
2017, Vol. 13(1)
An integrated system for building Ó The Author(s) 2017
DOI: 10.1177/1550147716689101
structural health monitoring and early journals.sagepub.com/home/ijdsn

warning based on an Internet of things

approach

Jun Wang1,2, Yongfeng Fu3 and Xiaokang Yang1

Abstract
The intelligent security monitoring of buildings and their surroundings has become increasingly crucial as the number of
high-rise buildings increases. Building structural health monitoring and early warning technology are key components of
building safety, the implementation of which remains challenging, and the Internet of things approach provides a new
technical measure for addressing this challenge. This article presents a novel integrated information system that com-
bines Internet of things, building information management, early warning system, and cloud services. Specifically, the sys-
tem involves an intelligent data box with enhanced connectivity and exchangeability for accessing and integrating the
data obtained from distributed heterogeneous sensing devices. An extensible markup language (XML)–based uniform
data parsing model is proposed to abstract the various message formats of heterogeneous devices to ensure data inte-
gration. The proposed Internet of things–based integrated information system structure was applied for monitoring an
actual pit excavation engineering site. Three early warning levels were implemented according to rules based on the
threshold value, which determined the specific safety personnel to be notified. The proposed Internet of things–based
integrated information system is demonstrated to improve the effectiveness of monitoring processes and decision mak-
ing in construction informatics applications. Our work highlights the crucial importance of a systematic approach toward
integrated information systems for effective information collection and structural health monitoring.

Keywords
Construction informatics, system integration, Internet of things, early warning system, structural health monitoring

Date received: 11 August 2016; accepted: 3 December 2016

Academic Editor: Yee Wei Law

Introduction in buildings over the entire life cycle continues to be a

key challenge for scientists and engineers.
In recent years, incidents involving the structural col-
lapse of buildings have raised a common concern,1 and 1
Institute of Image Communication and Information Processing, Shanghai
structural monitoring, modeling, and management
Jiao Tong University, Shanghai, China
have provided measures enabling the early warning of 2
The 52nd Institute of China Electronics Technology Group Corporation,
building-structure defects. These measures are activity Hangzhou, China
3
developed in the wide-ranging field of structural health Zhijiang College, Zhejiang University of Technology, Hangzhou, China
monitoring (SHM), which applies scientific and engi-
Corresponding author:
neering principles to prevent the loss of life and prop- Xiaokang Yang, Institute of Image Communication and Information
erty. However, the development of measures for Processing, Shanghai Jiao Tong University, Shanghai 200240, China.
effectively monitoring the process of structural change Email: xkyang@sjtu.edu.cn

Creative Commons CC-BY: This article is distributed under the terms of the Creative Commons Attribution 3.0 License
(http://www.creativecommons.org/licenses/by/3.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (http://www.uk.sagepub.com/aboutus/
openaccess.htm).
2 International Journal of Distributed Sensor Networks

The field of construction informatics focuses on warning model and a case study involving the safety
information processing activities, visualization, and the monitoring of an actual construction site are presented
conveyance of construction information and has pro- in sections ‘‘Early warning model’’ and ‘‘Case study
vided a feasible approach for addressing this challenge. and discussion,’’ respectively, and we finally conclude
Construction informatics incorporates the fields of our work by outlining its primary contributions in sec-
computer science and technology, system analysis engi- tion ‘‘Conclusion.’’
neering, and integrated information systems (IISs), all
of which have been applied for data acquisition, data
processing, and information exchange since the 1970s.2 Related work
Since 2000, the proliferation of automatic data acquisi-
IIS applications involve a broad range of technologies
tion technologies, such as radio frequency identification
that are widely used in the fields of computer and infor-
(RFID) and sensor technologies, has facilitated the
mation sciences. In this section, we review construction
development of decision support systems (DSSs) and
informatics, IISs, and existing early warning
construction information systems (CISs) and has also
technology.
brought new vitality to structural monitoring and
management.
The rapid development and wide application of con- Construction informatics
struction informatics have greatly promoted the build-
ing monitoring industry and enhanced management Construction informatics, also known as construction
efficiency and effectiveness. Nevertheless, most infor- information technology (IT) or communication and
mation systems based on construction informatics have information technologies in construction, was first pro-
been designed specifically to manage the building con- posed in the InfoMate model.4 It mainly focuses on the
struction process, and few actually address SHM prac- specific issues related to processing, representation, and
tices. The Internet of things (IoT) is an approach first interchange of specific information employed in the
proposed by Ashton3 that offers a promising frame- field of construction.5 Since its introduction, construc-
work for developing industrial systems and applica- tion informatics has developed as a methodology fos-
tions. Significant advancements, particularly in the past tering the integration and coordination of numerous
few years, have been achieved in key IoT technologies informatics technologies. It has improved decision
such as sensor technologies, sensor networks, detection, making to closely link informatics knowledge with
and intelligent computing. All these technologies facili- building, civil engineering, structural engineering, AEC
tate the deployment of IoT devices in the building envi- (architecture, engineering, construction), and other dis-
ronment for SHM applications such as early warning ciplines associated with the shaping of the built envi-
systems (EWSs). IoT-based applications are capable of ronment. Informatics-related topics are applicable in
promoting the entire process of structural monitoring, several civil engineering disciplines, for example, prod-
modeling, and management and also of supporting uct modeling, integration, concurrent engineering, vari-
early warning and sustainable decision making. An ous analysis and monitoring methods, as well as
IoT-based application can provide a strategy for the distance working and learning. Therefore, more cost-
implementation of EWSs, capable of detecting the effective and forward-looking solutions can be pro-
development of building-structure defects and bring vided using this methodology in various types of struc-
new vitality to SHM activities. This study proposes a tural problems.
novel IoT-based integrated information system (IIIS) Numerous researchers have introduced an IT-
application, which is presented herein as a new para- oriented view in construction and proposed new direc-
digm for the development of SHM and EWS applica- tions of research.6,7 Ontologies such as model and
tions. The proposed IIIS seeks to improve the efficiency schema have been identified by Ekholm.8 These ontolo-
of building safety monitoring and management tasks, gies divided the research themes from an IT-oriented
and the results from a case study demonstrate that the view, such as visualization,9,10 engineering manage-
proposed system is both valuable and efficient for com- ment,11 integration,12 and IT strategies.13 Accordingly,
plex tasks in construction monitoring and management. construction informatics can help address various types
The remaining sections of this article are organized of damage detection problems and provide more cost-
as follows. After discussing related work in section effective and forward-looking solutions in SHM.
‘‘Related work,’’ we provide an overview of the IoT However, these studies have focused on computer-
framework in section ‘‘System architecture,’’ present aided design construction or information support for
the employed data acquisition and transmission engineering management activities using conventional
approach in section ‘‘Data acquisition and transmis- IISs and have generally neglected IoT technology such
sion,’’ and propose a unified data packet parsing model as multi-sensor fusion, data integration, and decision
in section ‘‘Unified data packet parsing.’’ An early models.
Wang et al. 3

Present state of IIS development al.26 summarized the level of remote sensing technology
IISs have been studied from various perspectives such utilization and the potential trends for future EWSs. In
as integration scope, integration tools, integration point Salzano et al.,27 a seismic EWS was implemented to
of view, architectures, and technologies. Core IIS tech- analyze the relationship between structural monitoring
nologies and typical applications have been discussed in and the seismic reliability of equipment. In addition,
the literature,14,15 and IISs have been shown to provide other EWSs have been developed, including the InSAR
promising solutions to complex tasks in structural mon- monitoring system,28 the GBInSAR-based method,29
itoring and management. A number of research topics and seismic-event-detection,30 which utilize their own
have evolved from the perspective of CIS, such as early warning strategies based on the data analysis of
building information management (BIM),16,17 data real-world building structures.
integration, and decision-making systems. BIM is the EWS applications have promoted remarkable
foundation of early CIS and typically excels in spatial achievements in SHM, and advanced SHM is expected
analysis, information visualization, and construction with increased IIS integration facilitated by the applica-
modeling, but falls short in data acquisition and man- tion of IoT approaches. Considerable work however
agement. Data integration is a challenge for IIS. remains for establishing a practical and advanced IIIS
Sensors and actuators employed in structural monitor- for building safety control, and many issues must be
ing areas are highly heterogeneous and are therefore considered, particularly those aspects involving theore-
implemented with different data protocols and different tical research and warning strategies.
configuration methods. Hence, some researchers have
proposed to solve the problem of data source heteroge-
The motivations of this article
neity with a unified protocol parsing method.18
The IOT approach has been suitably applied to an The previous review of related work indicates that the
IIS by providing sensing in a distributed and self- efficiency of decision making in building monitoring
managing fashion.19 So far, numerous studies have and management has been effectively improved by the
paid close attention to the applications of IISs in differ- important progress in both construction informatics
ent areas within the framework of IoT-based IISs.20–22 and IIS, and the development and wide application of
An IIIS comprises a set of foundational concepts and various CISs. However, the structures of some systems
techniques that facilitates the information integration are outdated, and urgent demand exists for additional
process and includes methods for solving complex studies regarding next generation IISs that focus on the
problems in the development of construction infor- following aspects:
matics applications. Furthermore, web services and
integrated models have been used to create accessible (1) New technologies, such as SHM, CIS, and
interfaces and manage the integrated data sets in EWS, have not been well integrated into exist-
IIISs.23 ing systems.
It is noted that these studies have primarily sought (2) The problems of association and fusion
to develop IIIS for specific applications; however, less between IISs and data acquisition, data source
work is available on IIIS construction informatics. heterogeneity, and lack of acquisition terminals
Some shortcomings exist in both the architectures and with strong interoperability and IoT network
applications of previous CISs, and the problems of het- connectivity have been rarely studied.
erogeneity and integration created by IoT-based data (3) Compared to the single and fixed warning rules
have not been well addressed. employed in past EWSs, a flexible early warn-
ing model is required for improving emergency
response and level of preparedness.
The application of EWSs for SHM
EWSs are multifunctional systems consisting of sen- To address the above problems, we develop a novel
sors, signal processing, and communication facilities IoT-based IIS with high-level integrated, smart, and
that aim to provide timely information useful for sche- sustainable abilities because IoT reflects the near future
duling repairs and maintenance to minimize the impact trends in informatics. The features of our system are as
of structural degradation.24 An EWS combines a wide follows: (1) an intelligent data box (iDataBox) data
range of sensors and evaluation methods to provide acquisition terminal with strong interoperability and
accurate forecasting and enables users to improve the IoT network connectivity is used to connect different
level of emergency preparedness, as well as provides a sensor devices and to collect, preprocess, and forward
more intelligent service for users and enterprises.25 sensor data; (2) an unified data parsing framework is
At present, numerous studies have investigated the proposed to address the data source heterogeneity issue;
application of EWSs for SHM. The work in Quansah et and (3) two types of early warning model are employed
4 International Journal of Distributed Sensor Networks

Figure 1. Overall structure of the proposed IoT-based integrated information system (IIIS).

in our IIS to better support decision making based on modules. The analog or digital signals obtained from
warning signals derived from real-time parameters. sensors are conditioned in the local data processing
and control unit prior to being transferred to the data
processing and management layer. This layer is also a
System architecture multi-standard communication computing platform,
The rapid development of IT and information commu- which is compatible with different hardwire/wireless
nications technology (ICT) over the past decade has laid protocols such as those employed in Ethernet, Zigbee,
the foundation for implementing integrative informa- Wi-Fi, and 3G/4G networks.
tion architectures in IISs for construction informatics The data management layer functions as a transition
applications. Figure 1 presents the architecture of the auxiliary layer that involves the efficient administration
proposed IIIS in detail. The architecture of the proposed and management of structural health data and facili-
IIIS comprises three network layers as follows: (1) a sen- tates the smooth operation of the monitoring service
sor data collection layer, (2) a data processing and man- system. The data management module employed in this
agement layer, and (3) an SHM service layer. layer unifies the data packet format and supports static
The sensor data collection layer is the basis of the and dynamic data management by describing data
entire network, and its functionality is presented in packets according to a predefined datagram semantic
greater detail in Figure 2. This layer employs a number description model (DSDM) based on the extensible
of different types of sensors fixed at the pivotal loca- markup language (XML) syntax. Thus, device data can
tions of a large-scale, complex structure to acquire data be parsed according to the model, which is described in
reflecting the stress and health state of the structure. detail in a later section. Furthermore, cloud computing
This layer comprises data sensing and acquisition services are available to the supervision department and
devices, signal conditioning and processing units, data other interested parties, providing an efficient method
storage devices, and hardwire/wireless transmitting for ensuring data security.
Wang et al. 5

Figure 2. Structure of the sensor data collection layer.

The SHM service layer is directly linked to profes- keyboard is embedded into the lid of the data box. The
sional construction facilities such as the supervision data transfer module comprises a Wi-Fi unit, a Zigbee
department, construction units, and other interested unit, and an RS-485 port. Four 64-pin jacks (M1–M4)
parties. For example, this layer enables a safety supervi- are integrated in the middle of the iDataBox for the
sor to manage a large assortment of monitoring equip- installation of measurement modules for conducting
ment efficiently. The health state of the building as well various measurements such as vibration measurements,
as the historical data for a specific structural parameter resistance measurements, and voltage measurements
can be readily inspected. Through the self-contained for indicating the health state of the structure. The 32
EWS, the safety supervisor can analyze and process col- terminal ports are designed to be compatible with all
lected health-state information and make informed and types of sensors such as steel stress gauges, earth pres-
accurate maintenance and repair decisions based upon sure cells, and other third-party sensors/devices, which
structural conditions lying outside of normal ranges. can be connected to the iDataBox via various hard-
wire/wireless technologies.

Data acquisition and transmission

A large number of different types of distributed sensors
Data communication protocol
are typically employed in practical SHM applications. We designed a custom protocol to transfer data
Under this condition, the means of managing the vari- between the iDataBox and the master station. The data
ous distributed sensors and fusing the varieties of data communication comprises a series of protocols such as
actually detected by sensors is a particularly difficult the query command (Info), start data collection com-
problem that must be fully addressed. To this end, an mand (Start), end data collection command (End), and
iDataBox is integrated with data acquisition and read acquisition results command (Read). The Read
transmission in the proposed application. The command accepts data from a measurement module
iDataBox is the central terminal of the integrated sys- and adds a header creating a data segment. The data
tem for SHM. segment is then encapsulated into a datagram and
exchanged with the master station. The data processing
associated with the Read command is illustrated in
Architecture of the iDataBox Figure 4 and incorporates the following data segments:
The iDataBox is equipped with a high-performance
and open-platform-based tablet personal computer (1) Date (3 bytes) and Time (3 bytes) identify the
(PC) and wireless transmission units, so the iDataBox data acquisition time, including the year,
is fully available to building inspectors and on-site risk month, and day, and hours, minutes, and
examiners for daily monitoring. The components and seconds.
interfaces of the iDataBox are illustrated in Figure 3. A (2) Module ID (2 bytes) identifies the measure-
master controller module with a display module and ment module ID: M1, M2, M3, or M4.
6 International Journal of Distributed Sensor Networks

Figure 3. Hardware architecture and interfaces of the iDataBox.

Figure 4. Read command processing of the data communication protocol.

(3) Module Type (2 bytes) identifies the type of individual packet protocol is set by the equipment man-
measurement module. The default value of this ufacturer. As such, a multitude of communication pro-
field is ZX, which denotes a vibration wire tocols are used, each of which generally employs a
sensor. different structure. This is, to a large extent, the cause
(4) The Data Type (1 byte) field has a dual role. If of data heterogeneity. Therefore, to improve the effi-
the SYN flag is set (N), then the data represent ciency of system development for system engineers and
new sensor collection data. If the SYN flag is facilitate the efficient access of data for safety supervi-
clear (0), then this data segment has been sors, a flexible, semantic data model is required to sup-
received by the master station. port the access of heterogeneous data, particularly in
(5) D1 (2 bytes) specifies the data value of the first the big-data environment of an IoT-based application.
sensor in the measurement module. The semantic data model is expected to be self-
(6) T1 (2 bytes) specifies the data value of the first explanatory and to support diverse packet protocols as
temperature sensor in the measurement well as provide flexible and efficient data sharing.
module. Furthermore, for an SHM service, the data must be
(7) Checksum (2 bytes) is used for error checking accessible anytime and in any kind of working environ-
the header and data using the cyclic redun- ment. Based on these features of SHM services, we con-
dancy check. clude that the implementation of support for the access
of heterogeneous IoT data must have the following
capabilities:
Unified data packet parsing
In SHM, data from different sensor devices are invari- (1) To support data access in an environment of
heterogeneous packet protocols;
ably encapsulated in different formats because the
Wang et al. 7

Figure 5. Unified data packet parsing (UDP) framework.

(2) To be useful in building real-time application (4) The datagram parsing post-processing module
systems; post-processes the structured data objects
(3) To be capable of parsing and processing new parsed by the unified semantic parsing driver
protocols rapidly. module.
(5) The datagram description file describes the
To provide the above-mentioned capabilities, we data protocol based on the DSDM for the par-
present a DSDM that provides a formal unified descrip- sing of sensing device data packets by the uni-
tion of IoT data packet formats that is independent of fied semantic parsing driver module.
the individual protocol. This description model allows (6) The description file editor is more than just an
for the implementation of parsing and processing pro- authoring tool and also provides other func-
gramming that can avoid code parsing and processing tionality such as syntax coloring, keyword
for different protocols separately and makes the pro- selection, and document tree browsing.
gramming more flexible and adaptable.
The framework for unified data packet parsing
(UDP) proposed here is illustrated in Figure 5. The
modules given in the figure are explained in greater
Flowchart of the approach
detail as follows: The general processing of the UDP framework for data
packet parsing is to transform the original data to uni-
(1) The communication module establishes con- versal data in terms of both syntax and semantics
nections with physical devices and receives data (herein denoted by the ! symbol) along a chain of
packets from the devices. data transformation as follows
(2) The unified semantic parsing driver module
loads the datagram description file and parses Original data ðunstructuredÞ !
the received packet data for generating struc- Unified data ðsemi  structuredÞ ! Structured data
tured data objects.
(3) The DSDM refers to the common definition of The final structured data are then stored in a data-
the data packet structure characteristics and base. The key task of data transformation is parsing the
elements. The model emphasizes the self- original data and unifying access by means of a univer-
explanation of data values for accessing data sal syntax, as discussed in the next subsection.
ubiquitously. The data received from each IoT Based on the datagram description and parsing pro-
device is notated and defined in XML format, cess, the UDP approach is divided into several stages,
so that it is self-described and can be accessed given as datagram description file configuration, data
from the Web. acquisition, datagram parsing, and post-processing. A
simple flowchart of the approach is given in Figure 6.
8 International Journal of Distributed Sensor Networks

Figure 6. Simplified flowchart of the UDP approach.

DSDM based on XML Definition 1. The syntax of DSDM

XML is rapidly becoming the standard for data UD :¼ hDatagram, Sequence
exchange between machines. XML schema express :¼ element, KeyMappings
shared vocabularies and allow machines to carry out :¼ KeyItem, AttrSet :¼ hAttributesii
user-defined rules by providing a means for defining (1) UnifiedDef (UD). The root node of the seman-
the structure, content, and semantics of XML docu- tic description model and has attributes format
ments. As such, XML is useful for integrating struc- and device_type.
tured, semi-structured, and unstructured data in an IoT (2) Datagram. A primary node tag of the datagram
environment. Therefore, the DSDM representation is and has the delimiter attribute.
based on XML syntax. In general, semantics refer to (3) Sequence. A container containing one or more
the explanation of data symbols, such as the meaning elements (defined in (4)) and has attributes such
of classes, elements, and constraints. In this article, as delimiter, element_size, start, and end.
DSDM describes the meaning of the features in a (4) Element. Represents the structure information
device datagram. A DSDM is an abstract expression of a data element and has attributes such as
and definition of a device datagram and includes fea- name, size, and pattern.
tures such as a set of predefined keywords, the built-in (5) KeyMappings. A container containing one or
schema type, expression, and operator function, as more keyitems (defined in (6)) and has no
summarized in the following definition. attributes.
Wang et al. 9

Figure 7. The syntax structure of the datagram semantic description model (DSDM) for the UDP framework.

Table 1. XML description of the DSDM. Table 2. Datagram received from the iDataBox in response to
a Read command.
\UnifiedDef format = ‘‘ascii |hex’’ devicetype =
‘‘DeviceType’’. 15/07/20 08:09:25 M1ZXN M1 = 2789.6 T1 = 210.0
\Datagram delimiter = ‘‘;’’. M2 = 4456.0 T2 = 20.5 M3 = 6635.7 T3 = 18.9 M4 = 4311.0
\Element name = ‘‘column1’’ size = ‘‘’’type = ‘‘string’’/. T4 = 25.0 M5 = 4564.8 T5 = 20.0 M6 = 3210.5 T6 = 40.0
...... M7 = 3333.6 T7 = 215.0 —— checksum0d0a
\Sequence start = ‘‘&&’’end = ‘‘&&’’delimiter = ‘‘;’’.
\Element name = ‘‘$Key’’pattern = ‘‘{=,Key,Value}‘‘/.
......
\KeyMappings.
\KeyItem name = ‘‘Key1’’ type = ‘‘number’’/.
(2) pattern = yymmddhh:mm:ss represents a
...... combined time and date element, which is used
\/KeyMappings. to represent the time data format.
\/Sequence element_size = ‘‘2’’.
\Element name = ‘‘timestamp’’ size = ‘‘’’type = ‘‘datetime’’/. Semantic description is used for expressing the data-
\Element name = ‘‘!end’’ value = ‘‘endchars’’/.
\/Datagram.
gram structure of devices, data element analytical
\/UnifiedDef. transformations, and packet protocol. Because each
item of the data is defined with ontology, the data can
be explained at a semantic level and integrated with
other instances of similar or related concepts.
(6) KeyItem. An optional tag that is used to define
some data element sequences that have an
equivalent regular pattern and has attributes XML model implementation
such as name, type, and pattern. The implementation is illustrated by some examples.
(7) AttrSet. Sets the attributes of the DSDM. Table 2 presents a received datagram responding to a
read command, as illustrated in Figure 3. Table 3 pre-
DSDM elements have been described, and their rela- sents an iDataBox datagram description file, which
tionship to the overall task is expressed in Figure 7. In describes the read command of the data communica-
addition to the preceding definitions, XML syntax tion protocol. The unified semantic parsing driver mod-
always comprises XML tags, and other semantic ele- ule parses the datagram into XML syntax according to
ments, as shown in Table 1.The tag \pattern. is used the DSDM. The module ID, sensor type, and update
to identify a series of specific character strings, an type are identified by pattern = {delimiter, key1, key2,
example of which is presented in Table 1. Some pattern key2, key3}, which is denoted in the table as M1ZXN,
examples are presented as follows: where the module_id is M1, the sensor_type is ZX, and
the update_type is N.
(1) pattern = {delimiter, Key, Value}, for exam-
ple, T1= 20.5 can be expressed as Early warning model
{=,Key,Value}, where the equals sign (=) is
A warning rule is based on changes in the characteris-
the delimiter, the key is T1 (denoting the tem-
tics of a monitored item and functions according to a
perature), and the value is 20.5.
user-defined warning model. To meet different levels of
10 International Journal of Distributed Sensor Networks

Table 3. Datagram description file of the iDataBox.

\UnifiedDef format = ‘‘ascii’’ devicetype = ‘‘XXX-iDataBox-Node’’.

\Datagram delimiter = ‘‘’’.
\!–the Element identifies the data-acquisition time: 15/07/20 08:09:25–.
\Element name = ‘‘ collectedtime’’ type = ‘‘datetime’’ pattern = ‘‘yy/mm/dd hh:mm:ss’’/.
\!—The Sequence identifies the module number, sensor type and update type:M1ZXN –.
\Sequence element_size = ‘‘3’’pattern = ‘‘{!none,module_no,#ZX,update_type}’’.
\Element name = ‘‘module_id’’type = ‘‘string’’/.
\Element name = ‘‘sensor_type’’type = ‘‘string’’/.
\Element name = ‘‘update_type’’type = ‘‘string’’size = ‘‘1’’/.
\/Sequence.
\!–The sequence specifies the data value of the sensor: M1 = 2789.6 T1 = 210.0 M2 = 4456.0 T2 = 20.5 M3 = 6635.7
T3 = 18.9 M4 = 4311.0 T4 = 25.0 M5 = 4564.8 T5 = 20.0 M6 = 3210.5 T6 = 40.0 M7 = 3333.6 T7 = 215.0 –.
\Sequence until = ‘‘—’’.
\Element name = ‘‘$Key’’alt_name = ‘‘frequency’’ pattern = ‘‘{=,Key,Value}’’ type = ‘‘string’’/.
\Element name = ‘‘$Key’’alt_name = ‘‘temperature’’ pattern = ‘‘{=,Key,Value}’’type = ‘‘string’’/.
\/Sequence.
\!–The element identifiesthe checksum: checksum0d0a–.
\Element name = ‘‘checksum_endchar’’pattern = ‘‘{!none,%checksum,#0d0a}’’type = ‘‘string’’/.
\/Datagram.
\/UnifiedDef.

Figure 8. The early warning strategy.

urgency, each type of warning model is divided into system, the security issue may represent a significant
three levels, which are presented in Figure 8. The three security risk that may potentially lead to an exceedingly
warning levels are blue, yellow, and red. Whenever a serious safety condition. Therefore, the red warning
given threshold established for a monitored item is message transmits notification of the hazard to the
exceeded, a warning signal must be transmitted to the supervision department. In addition, while the warning
interested parties for enacting adequate countermea- system provides real-time notification, it also records
sures. The yellow warning level transmits a warning the alarm information and corresponding data stream
message to the security supervisor of the construction for later query.
project. The orange warning level transmits the security The following two types of warning models were
issue to the project manager. When the system issues a defined according to the relationship between the real-
red warning, the highest of the three-level warning time value (R.Value) of a monitored item (the measured
Wang et al. 11

Figure 9. Monitored points in the excavation pit structure.

value in real time), the threshold value (T.Value) (estab- Table 4. Monitoring points and monitoring objects for the pit
lished independently), and the base value (B.Value) of a structure illustrated in Figure 9.
monitored item (its standard operating value).
Monitoring point Monitoring objects
(1) The measured value model compares R.Value A Gas pipe line
directly with T.Value. The model has two rules: B Soil behind the wall
R.Value . T.Value or R.Value  T.Value. C Continuous wall
(2) The difference value model allows the user to D Surrounding roads
E Support columns
observe the variation in the value of a monitored
G Surrounding pipe
item with respect to its base value. According to
differences in the monitored item, the rule can
be set as |R.Value-B.Value| \ T.Value or
|R.Value-B.Value|  T.Value.
preprocess the data obtained from the sensors installed
in the foundation pit.
Case study and discussion
In this section, we present the results of a case study to Results
demonstrate the performance of the proposed IIIS.
Figure 11(a) displays the graphic interface of the moni-
toring platform for a typical early warning process
Case study using the measured value model. The x-axis represents
Instability in an underground pit wall structure or a time, and the y-axis represents the strut force of the sup-
foundation pit excavation accident can result in serious port column at point E13. The yellow warning thresh-
multi-pile displacement when a high-rise building is old value (Tyellow :Value) is 4000 kN, the orange warning
constructed on soft ground. Accordingly, the employ- threshold value (Torange :Value) is 5000 kN, and the red
ment of an EWS is indispensable under this condition. warning threshold value (Tred :Value) is 9000 kN. Figure
The IIIS developed in this work was tested in an actual 11(a) clearly shows that the three warning levels are sep-
pit excavation engineering site for a six-story office arately trigged by the changing strut force according to
building with a composite frame structure. Figure 9 the various threshold settings. The message includes
illustrates the positions of the monitoring points in the monitoring point information and a description of the
pit structure. Settlement of the building elements was abnormality. Subsequently, the appropriate supervisory
detected by a variety of sensors, such as steel stress personnel can log into the database via a remote termi-
gauges and earth pressure cells, and the monitoring nal and evaluate the necessity for checking the moni-
points of the pit structure are listed in Table 4. tored site. Moreover, the alarm history is regularly
Figure 10 shows the IIIS deployment in the pit excava- saved in the server, and safety regulators can employ it
tion engineering site. Here, the iDataBox collects and as a reference for project acceptance.
12 International Journal of Distributed Sensor Networks

Figure 10. Test setup at the pit excavation engineering site.

Figure 11. Monitoring results obtained from the graphical interface of the monitoring platform: (a) the strut force of the support
column at point E13 and (b) the soil pressure of the soil located behind the wall at point B9.

The functionality of the sensors was tested over a measured soil pressure was caused by the operation of
period of 2 years throughout the entire construction heavy machinery or the storage of building materials in
phase, and a specific time period of monitoring data the area around the sensor.
was selected for discussion. Figure 11(b) displays the
early warning process using the difference value model
for monitored changes in the soil pressure of the soil Discussion
located behind the wall at point B9. In addition, we The characteristics of the proposed IIIS were compared
analyzed the factors influencing the warning messages with those of the user-centric design for adoption of
received. In some cases, the abnormal change in the Internet of things (UDA-IoT)20 and the shared early
Wang et al. 13

Table 5. Comparisons with similar systems.

Features Our solution UDA-IoT20 SEWS31

Integrated heterogeneous hardware O 3 O

Data model Uniform data parsing model Resource model MySQL
Software architecture J2EE, restful Restful Not given
Restful services O O 3
Early warning model O 3 O
Device parameter configurability O 3 3
Dynamic graphical interface O 3 3

UDA-IoT: user-centric design for adoption of Internet of things; SEWS: shared early warning system.

warning system (SEWS) deployable SEW suite (DSS),31 development efficiency, particularly when
which are listed in Table 5. As shown in the table, the incorporating new heterogeneous sensor
proposed IIIS has several advantages over one or both devices with the monitoring platform.
of the other systems, such as support for integrated het- (3) The early warning functionality of the pro-
erogeneous hardware, the employment of an early posed IIIS was applied for monitoring an
warning model, device parameter configurability, and a actual pit excavation engineering site. The sys-
dynamic graphical interface. tem was demonstrated to be highly modular-
The features of our proposed IIIS are reflected in the ized and extensible. The IIIS allowed for the
following aspects: (1) our model can provide support setting of multiple alarm levels and for the
for integrated heterogeneous hardware by abstracting inline modification of the individual threshold
IoT data into a unified access interface using a unified values; thus, it was shown to be a very flexible
data parsing model, which is another highlight of our system.
solution, (2) two types of early warning models are
employed in our IIS to better support decision making The better prevention and preparation provided by
based on warning signals derived from real-time para- EWSs greatly contributes to avoiding the loss of life
meters, (3) the proposed IIIS employs stable software and property. Accordingly, the new IIIS introduced in
architecture and provides portability and maintainabil- this work clearly enhances the contribution expected
ity, (4) the device parameters are configurable in the from an EWS and provides a new paradigm for the
software interface, and (5) the warning results can be future development of structural monitoring and early
obtained from a dynamic graphical interface, as shown warning applications. Meanwhile, future work can
in Figure 11. focus on applying the proposed IIIS to other smart city
applications such as city safety monitoring and infra-
Conclusion structure supervision.

This article presented a novel IIS based on IoT that was

designed to improve the efficiency of complex tasks in Declaration of conflicting interests
construction informatics applications. The main contri- The author(s) declared no potential conflicts of interest with
butions of the work are as follows: respect to the research, authorship, and/or publication of this
article.
(1) The iDataBox data-acquisition terminal was
developed to connect various types of sensor
Funding
devices and to collect, preprocess, and forward
sensor data. The iDataBox services a construc- The author(s) disclosed receipt of the following financial sup-
tion healthcare station by providing strong port for the research, authorship, and/or publication of this
interoperability and IoT network connectivity article: This work was supported in part by the Foundation
for Innovative Research Groups of the National Natural
and is therefore fully available for building
Science Foundation of China (no. 61221001 and no.
inspectors and on-site risk examiners for daily 61304125).
monitoring.
(2) We proposed a unified data parsing framework
to address the issues associated with data References
source heterogeneity. A DSDM based on 1. Giangarra PP, Metrovich B, Schwitters MM, et al. Smar-
XML was defined. The semantic description of ter bridges through advanced structural health monitor-
datagrams can profoundly enhance application ing. IBM J Res Dev 2011; 55(1.2): 9.1–9.10.
14 International Journal of Distributed Sensor Networks

2. Eastman C, et al. An outline of the building description sys- 18. EPC Tag Data Standard. GS1 standard version 1.8. 2014,
tem. Institute of Physical Planning, Carnegie-Mellon Uni- http://www.gs1.org/sites/default/files/docs/epc/
versity, 1974, http://files.eric.ed.gov/fulltext/ED113833.pdf TDS_1_8_Standard_20140203.pdf
3. Ashton K. That ‘‘Internet of Things’’ thing. RFID J, 19. Miorandi D, Sicari S, Pellegrini FD, et al. Internet of
http://www.rfidjournal.com/articles/view?4986 things: vision, applications and research challenges. Ad
4. Björk BC. Information technology in construction— Hoc Netw 2012; 10(7): 1497–1516.
domain definition and research issues. Int J Comput 20. Xu B, Xu LD, Cai H, et al. Ubiquitous data accessing
Integr Des Constr 1999; 1: 1–16. method in IoT-based information system for emergency
5. Turk Z. Construction informatics: definition and ontol- medical services. IEEE T Ind Inform 2014; 10(2):
ogy. Adv Eng Inform 2006; 20: 187–199. 1578–1586.
6. Ida S and Mikami B. Information technology for con- 21. Islam SMR, Kwak D, Kabir MH, et al. The Internet of
struction: recent work and future directions. Electron J things for health care: a comprehensive survey. IEEE
Inform Tech Constr 2002; 7: 245–258. Access 2015; 3: 678–708.
7. Brandon P and Betts M. The Armathwaite initiative: the 22. Xu L. Information architecture for supply chain quality
formation of a global construction IT network. University management. Int J Prod Res 2011; 49(1): 183–198.
of Salford, 1997, http://citeseerx.ist.psu.edu/viewdoc/ 23. Tao F, LaiLi Y, Xu L, et al. FC-PACO-RM: a parallel
download?doi=10.1.1.453.3805&rep=rep1&type=pdf method for service composition optimal-selection in
8. Ekholm A. Themes and projects of ‘‘IT bygg&fastighet cloud manufacturing system. IEEE Trans Ind Inform
2002’’—an analysis as a background for further initiatives. 2013; 9(4): 2023–2033.
Lund Institute of Technology, Lund University, 2002, 24. Waidyanatha N. Towards a typology of integrated func-
http://www.caad.lth.se/fileadmin/projekteringsmetodik/ tional early warning systems. Int J Crit Infrastruct 2010;
research/Other_projects/Ekholm_themes_projects.pdf 6(1): 31–51.
9. Ebrahim AB, Mosly I and Abed-Elhafez IY. Building 25. Li S, Xu L, Wang X, et al. Integration of hybrid wireless
construction information system using GIS. Arab J Sci networks in cloud services oriented enterprise informa-
Eng 2016; 41: 3827–3840. tion systems. Enterprise Inform Syst 2012; 6(2):
10. Fukushi M. Visualization of construction information 165–187.
using the BIM model. Concrete J 2015; 53(5): 503–509. 26. Quansah JE, Engel B and Rochon GL. Early warning
11. Wang HH, Boukamp F and Elghamrawy T. Ontology- systems: a review. J Terr Obs 2010; 2(2): 24–44.
based approach to context representation and reasoning 27. Salzano E, Di Carluccio A, Agreda AG, et al. Seismic
for managing context-sensitive construction information. early warning systems for the process industry. In: Pro-
J Comput Civil Eng 2011; 25(5): 331–346. ceedings of the 12th international symposium on loss pre-
12. Li J, Shao K and Zhang L. The design of construction vention and safety promotion in the process industries,
information integration model based on BIM. Construct Edinburgh, Scotland, 22–24 May 2007, pp.1–7. UK:
Tech 2015; S2: 688–691. IChemE.
13. Albert A, Hallowell MR and Kleiner BM. Emerging stra- 28. Traglia FD, Intrieri E, Nolesini T, et al. The ground-
tegies for construction safety & health hazard recognition. based InSAR monitoring system at Stromboli volcano:
In: Proceedings of the 14th IEEE international conference linking changes in displacement rate and intensity of
on cognitive informatics & cognitive computing, Beijing, persistent volcanic activity. B Volcanol 2014; 76(2):
China, 6–8 July 2015, pp.152–161. New York: IEEE. 1–18.
14. Izza S. Integration of industrial information systems: 29. Pratesi F, Nolesini T, Bianchini S, et al. Early warning
from syntactic to semantic integration approaches. Enterp GBInSAR-based method for monitoring Volterra (Tus-
Inf Syst 2009; 3(1): 1–57. cany, Italy) city walls. IEEE J Select Topic Appl Earth
15. Panetto H and Cecil HJ. Information systems for enter- Obs Remote Sens 2015; 8(4): 1753–1762.
prise integration, interoperability and networking: theory 30. Federici F, Alesii R, Colarieti A, et al. An accelerometer
and applications. Enterprise Inform Syst 2013; 7(1): 1–6. digital front end for efficient seismic event detection sup-
16. Eastman C, Teicholz P, Sacks R, et al. BIM handbook: a port in a wireless sensor node. In: IEEE Sensors, Valen-
guide to building information modeling for owners, manag- cia, Spain, 2–5 November 2014, pp.265–268. New York:
ers, designers, engineers and contractors. Hoboken, NJ: IEEE.
John Wiley & Sons, 2008. 31. Rainieri C, Fabbrocino G and Cosenza E. Integrated
17. Li S, Isele J and Bretthauer G. Proposed methodology for seismic early warning and structural health monitoring of
generation of building information model with laser scan- critical civil infrastructures in seismically prone areas.
ning. Tsinghua Sci Tech 2008; 13(S1): 138–144. Struct Health Monit 2011; 10(3): 291–308.