Materi Bimtek Penulisan Artikel Jurnal Ilmiah Internasional 2017
II.1 Related Work in User Mobility
In the current literature, the term ‘user mobility’ can be found in three research areas; 1.
Immersive Virtual Environment, 2. cellular phone wireless mobile networks and 3. Intelligent
Environment including nomadic computing, ubiquitous and pervasive computing. User mobility
in this thesis has only a small relationship to user mobility in an Immersive Virtual Environment,
which deals with how the users control their viewpoints to move in virtual space (Frees and
Kesser 2003). This study also has only a loose relationship to user mobility in cellular phone
wireless mobile networks, which enable a mobile user to communicate with others regardless of
location. User mobility in this area has more mature technology than in the Intelligent
Environment area. This cellular environment uses user mobility information as follows:
assisted user mobility management (traffic routing) (Shen, Mark et al. 2000; Kravets,
Carter et al. 2001),
managed network resources, such as resources allocation, call admission control,
congestion and flow control (Bellavista, Corradi et al. 2000; Shen, Mark et al. 2000),
predicted user mobility (Liu, Bahl et al. 1998; Chan and Seneviratne 1999; Akyildiz and
Wang 2004),
user mobility patterns or movement pattern schemes (Zonoozi and Dassanayake 1997;
Chan and Seneviratne 1999; Cayirci and Akyildiz 2002)
User cell change processes in micro-cells and macro-cells (Liu, Bahl et al. 1998; Cho,
Chung et al. 2000; Del-Re, Fantacci et al. 2000)
Most research into user mobility in the Ubiquitous Computing area focuses on the problem
of host mobility, to allow the user access to the same service while moving. However, to do this
the user needs to carry the same mobile host. Cui et al. argue that this is only a special case of
user mobility (Cui, Nahrstedt et al. 2004). In the Intelligent Environment, user mobility not only
includes host mobility, but also includes the case where a user is free to switch from one host to
another as well and move from one location or computing environment to another. Further, in the
case of user mobility, a user is free to access his personalised service anytime, anywhere, through
any possible mobile or fixed device (Sousa and Garlan 2002).
While the use of movement prediction seems to be a promising approach for improving the
efficiency, reliability and adaptability of wireless networks, the actual user mobility client
patterns are not yet well understood (Chan and Seneviratne 1999).
In the literature, current approaches to user mobility in ubiquitous computing are based on
one of five techniques, none of which fully achieves the ubiquitous computing goal. One
approach is to support as much of a user’s computing needs as possible on a mobile machine. A
second approach is to compute via remote access to a computing server that stores a user’s
personal state and preferences, as much as can be done with VNC or XDM (X-terminals). A third
approach is to provide standard applications that are ported to and installed in all environments.
Those applications are extended to become aware of user intention and mobility. A fourth
approach provides standard virtual platforms (such as the Java Virtual Machine) that enables
mobile code to follow the user as needed (Sousa and Garlan 2002). The last approach is to
provide user access from one terminal to another actual terminal, which is not a virtual terminal
such as VNC does, in his working environment. This fifth approach requires the system to
authenticate a user entering the system and to organise a user’s working environment
accordingly.
There are two problems with these approaches. First, since to some degree the assumption of
homogenous computing baseline is used, this cannot take full advantage of the diverse
capabilities of each environment, such as external displays, processors, and I/O devices. Second,
the lack of ability to handle dynamic variations to capabilities and resources in the context-aware
environment without overburdening the user with manual tuning and reconfiguration (Sousa and
Garlan 2002; Cui, Nahrstedt et al. 2004).
A user mobility use mostly in wireless environment, which the availability of large
bandwidth, low error rates and always-on connectivity exists as it has in a wired environment.
Mobile user devices need to quickly detect and adapt to drastic changes in the characteristics and
available resources of the wireless environment. In this study, middleware technology is
considered for this purpose, as intermediary to providing higher-level network services and
abstract service environments for distributed applications.
The middleware technology supports transparent service to users. This transparency builds a
new form of awareness of an environment that allows the execution context and adapts to
middleware behaviour. Unfortunately in executing resources and execution environments from
direct application participation often results in premature termination of the problematic
application if available resources are depleted. An alternative approach is to develop an eventdriven mobile middleware that supports a degree of flexibility by allowing direct participation of
applications in adapting to changes in resources. In the formulation of a context-aware
middleware, suitable control mechanisms were required to directly participate in resources
adaptation in response to the dynamic operating environment (Chan, Chuang et al. 2004). The
scripting of an event service including the declarative language can be developed based on XML,
which supports synchronous event call-backs, to express the conditions that are bound to a
specific event in order to build a composite event from a set of primitive events.
The early event model has been introduced on agent-oriented software engineering, such as
ROADMAP (Juan, Sterling et al. 2002), Gaia (Wooldridge, Jennings et al. 2000; Juan, Pearce et
al. 2002), Prometheus (Padgham and Winikoff 2002). In event model, the object abstractions are
on the flow and the processing of events is applied across different levels leading to the ultimate
notification of events to the service objects. For example, Chuang et al. developed event platform
in WebPADS (Chuang, Chan et al. 2002). When the WebPADS client starts executing, a defaultservice chain is created, which is based on the description in XML configuration. The servicechain map is attached to an environment monitor, which regulates the time and conditions that
determine when reconfiguration of the service is to take place (Chan, Chuang et al. 2004), which
can be used for monitoring user mobility.
The user mobility problem in Ubiquitous Computing has significant challenges in
developing Active Office, for example, developing infrastructure with a variety of wired and
wireless sensors, fusing sensor data using a spatio-temporal database (see Section 8.4.1), and the
use of machine learning (see Section 5.2) for a variety of user locations and user activities, the
use of event-driven mobile middleware for a variety of user mobility. In developing context
aware systems, it also challenges because of the use of already existing devices (old sensors and
networks) and new embedded devices, to develop context-aware applications, which require
toolkits designed to enable maximum capability of the devices/sensors. The toolkit design should
follow the requirements of the context aware applications, which are:
a. ease of deployment
b. network and sensor scalability through sensor fusion rather than in precision
c. enabling wider acceptance through better design for user needs
d. increasing human trust in the system’s care for user privacy and security.
This study contributes to user mobility in the Ubiquitous Computing area, especially in
indoor space, which is a user mobility model based on user mobility patterns in the history of the
predicted and proximate users’ locations in an Active Office. The location history that was
collected used wired and wireless sensors and scalable context processing.
In the current literature, the term ‘user mobility’ can be found in three research areas; 1.
Immersive Virtual Environment, 2. cellular phone wireless mobile networks and 3. Intelligent
Environment including nomadic computing, ubiquitous and pervasive computing. User mobility
in this thesis has only a small relationship to user mobility in an Immersive Virtual Environment,
which deals with how the users control their viewpoints to move in virtual space (Frees and
Kesser 2003). This study also has only a loose relationship to user mobility in cellular phone
wireless mobile networks, which enable a mobile user to communicate with others regardless of
location. User mobility in this area has more mature technology than in the Intelligent
Environment area. This cellular environment uses user mobility information as follows:
assisted user mobility management (traffic routing) (Shen, Mark et al. 2000; Kravets,
Carter et al. 2001),
managed network resources, such as resources allocation, call admission control,
congestion and flow control (Bellavista, Corradi et al. 2000; Shen, Mark et al. 2000),
predicted user mobility (Liu, Bahl et al. 1998; Chan and Seneviratne 1999; Akyildiz and
Wang 2004),
user mobility patterns or movement pattern schemes (Zonoozi and Dassanayake 1997;
Chan and Seneviratne 1999; Cayirci and Akyildiz 2002)
User cell change processes in micro-cells and macro-cells (Liu, Bahl et al. 1998; Cho,
Chung et al. 2000; Del-Re, Fantacci et al. 2000)
Most research into user mobility in the Ubiquitous Computing area focuses on the problem
of host mobility, to allow the user access to the same service while moving. However, to do this
the user needs to carry the same mobile host. Cui et al. argue that this is only a special case of
user mobility (Cui, Nahrstedt et al. 2004). In the Intelligent Environment, user mobility not only
includes host mobility, but also includes the case where a user is free to switch from one host to
another as well and move from one location or computing environment to another. Further, in the
case of user mobility, a user is free to access his personalised service anytime, anywhere, through
any possible mobile or fixed device (Sousa and Garlan 2002).
While the use of movement prediction seems to be a promising approach for improving the
efficiency, reliability and adaptability of wireless networks, the actual user mobility client
patterns are not yet well understood (Chan and Seneviratne 1999).
In the literature, current approaches to user mobility in ubiquitous computing are based on
one of five techniques, none of which fully achieves the ubiquitous computing goal. One
approach is to support as much of a user’s computing needs as possible on a mobile machine. A
second approach is to compute via remote access to a computing server that stores a user’s
personal state and preferences, as much as can be done with VNC or XDM (X-terminals). A third
approach is to provide standard applications that are ported to and installed in all environments.
Those applications are extended to become aware of user intention and mobility. A fourth
approach provides standard virtual platforms (such as the Java Virtual Machine) that enables
mobile code to follow the user as needed (Sousa and Garlan 2002). The last approach is to
provide user access from one terminal to another actual terminal, which is not a virtual terminal
such as VNC does, in his working environment. This fifth approach requires the system to
authenticate a user entering the system and to organise a user’s working environment
accordingly.
There are two problems with these approaches. First, since to some degree the assumption of
homogenous computing baseline is used, this cannot take full advantage of the diverse
capabilities of each environment, such as external displays, processors, and I/O devices. Second,
the lack of ability to handle dynamic variations to capabilities and resources in the context-aware
environment without overburdening the user with manual tuning and reconfiguration (Sousa and
Garlan 2002; Cui, Nahrstedt et al. 2004).
A user mobility use mostly in wireless environment, which the availability of large
bandwidth, low error rates and always-on connectivity exists as it has in a wired environment.
Mobile user devices need to quickly detect and adapt to drastic changes in the characteristics and
available resources of the wireless environment. In this study, middleware technology is
considered for this purpose, as intermediary to providing higher-level network services and
abstract service environments for distributed applications.
The middleware technology supports transparent service to users. This transparency builds a
new form of awareness of an environment that allows the execution context and adapts to
middleware behaviour. Unfortunately in executing resources and execution environments from
direct application participation often results in premature termination of the problematic
application if available resources are depleted. An alternative approach is to develop an eventdriven mobile middleware that supports a degree of flexibility by allowing direct participation of
applications in adapting to changes in resources. In the formulation of a context-aware
middleware, suitable control mechanisms were required to directly participate in resources
adaptation in response to the dynamic operating environment (Chan, Chuang et al. 2004). The
scripting of an event service including the declarative language can be developed based on XML,
which supports synchronous event call-backs, to express the conditions that are bound to a
specific event in order to build a composite event from a set of primitive events.
The early event model has been introduced on agent-oriented software engineering, such as
ROADMAP (Juan, Sterling et al. 2002), Gaia (Wooldridge, Jennings et al. 2000; Juan, Pearce et
al. 2002), Prometheus (Padgham and Winikoff 2002). In event model, the object abstractions are
on the flow and the processing of events is applied across different levels leading to the ultimate
notification of events to the service objects. For example, Chuang et al. developed event platform
in WebPADS (Chuang, Chan et al. 2002). When the WebPADS client starts executing, a defaultservice chain is created, which is based on the description in XML configuration. The servicechain map is attached to an environment monitor, which regulates the time and conditions that
determine when reconfiguration of the service is to take place (Chan, Chuang et al. 2004), which
can be used for monitoring user mobility.
The user mobility problem in Ubiquitous Computing has significant challenges in
developing Active Office, for example, developing infrastructure with a variety of wired and
wireless sensors, fusing sensor data using a spatio-temporal database (see Section 8.4.1), and the
use of machine learning (see Section 5.2) for a variety of user locations and user activities, the
use of event-driven mobile middleware for a variety of user mobility. In developing context
aware systems, it also challenges because of the use of already existing devices (old sensors and
networks) and new embedded devices, to develop context-aware applications, which require
toolkits designed to enable maximum capability of the devices/sensors. The toolkit design should
follow the requirements of the context aware applications, which are:
a. ease of deployment
b. network and sensor scalability through sensor fusion rather than in precision
c. enabling wider acceptance through better design for user needs
d. increasing human trust in the system’s care for user privacy and security.
This study contributes to user mobility in the Ubiquitous Computing area, especially in
indoor space, which is a user mobility model based on user mobility patterns in the history of the
predicted and proximate users’ locations in an Active Office. The location history that was
collected used wired and wireless sensors and scalable context processing.