Bayesian and behavior networks for context-adaptive user interface in a ubiquitous home environment
Highlights
► This paper presents a work on using Bayesian and behavior networks to generate an adaptive user interface for smart home environments. ► The needs come from the complexity of device control in a smart home environment. ► This paper proposes to use context information to infer more relevant controls. ► The approach is a mix of induction-based (machine learning) and deduction-based (rule inference). ► We compared the adaptive user interface with the fixed user interface by surveying fourteen subjects.
Introduction
People use remote controls in order to control appliances. As more appliances are purchased, the number of controllers also increases. For example, users have to control several devices including a cable TV receiver, DVD player, television, audio equipment, video players, and DivX players in order to use modern home theater systems. Controllers for these devices have different interfaces based on their manufacturers and models even though they look similar. Therefore, it is not easy to get accustomed to all of these controller interfaces (Nichols, Rothrock, Chau, & Myers, 2006). Even though the controllers have various functions, only a few of those functions are practical for use. If the ubiquitous home environment becomes more common, most home devices including lights, boilers and home appliances will also be controlled using a remote controller. Therefore, a context-adaptive user interface that provides users with necessary functions should be developed.
The personal universal controller (PUC), which automatically generates user interfaces in PDAs or smart phones, has recently been studied. Nichols and Myers of Carnegie Mellon University developed such a system, which can generate an optimized controller interface for smart phones using a hierarchical menu navigation system (Nichols & Myers, 2006) and introduced HUDDLE, a system that can automatically generate task-based user interface (Nichols et al., 2006). In addition, they verified through a user survey that the automatically generated interface had superior usability in terms of time efficiency and consistency than did the general interfaces for device control (Nichols, Chau, & Myers, 2007). These studies generated and provided a useful PUC, but they did not take into consideration the context for its uses. In a ubiquitous environment, it is very important to consider the context in order to provide relevant services and information (Lee, Seo, & Rhee, 2008).
This paper presents a context-adaptive user interface. In a modeled ubiquitous home environment with location and device information, the proposed method used the Bayesian network to predict the necessity of the devices given the context and the behavior network to select the necessary functions given the required devices in the current context. The adaptive user interface consists of these selected functions based on a presentation template. Modeling using Bayesian networks provides good performance by effectively incorporating the domain knowledge (Kleiter, 1996). We also used the behavior network so that the generated user interfaces could be adapted using the user input (usage log of the devices).
For this experiment, we created a synthetic log for the use of the devices based on the experimental scenario, and then we inferred the necessities of these devices. We then evaluated the proposed method based on the time required to conduct the predefined tasks. Fourteen subjects were asked to perform ten tasks using both the conventional fixed home user interface and the proposed adaptive home user interface. The results indicated that the subjects dealt with the general tasks faster when using the proposed user interface than when using the conventional interface.
Section snippets
User Interface in a ubiquitous home environment
Recently, universal controllers or PDA software that can control several devices have been introduced (Nichols & Myers, 2006). In order to use these controllers and software, people have to assign the infra-red signal of the remote control to the corresponding button and then implement it. Home network technology, which allows users to access home appliances wherever they are, has been studied for the control of devices without the use of additional processes, as noted above (Filibeli, Ozkasap,
Context-adaptive user interface for a ubiquitous home environment
Fig. 3 summarizes the process for the proposed adaptive user interface generation for use with ubiquitous home devices. To begin with, the Bayesian network infers the necessary devices in the current context. Based on this result and the description of the devices and controllers, the behavior network is constructed in order to select the necessary functions for each device. Finally, the user interface for a given controller is generated using the user interface template.
Experiments
We conducted experiments in a simulated ubiquitous home environment for evaluation. Comparing the proposed adaptive UI with the conventional fixed UI, we confirmed that the adaptive UI had a superior performance.
Conclusions and future works
In this paper, we proposed an adaptive user interface generation method using Bayesian and behavior networks. The Bayesian network was used to select the necessary devices in a home environment, and the behavior network was used to select the functions that would be included in the adaptive user interface. For evaluation, we modeled a general home environment with devices and functions, and evaluated the user interface generated by the proposed method. Comparing to the fixed user interface, we
Acknowledgements
The authors would like to thank Dr. Han-Saem Park for preparing for this manuscripts. This research is supported by Ministry of Culture, Sports and Tourism (MCST) and Korea Creative Content Agency (KOCCA) in the Culture Technology (CT) Research & Development Program, and by Korea Communications Commission (KCC) as a project, ”Development of UX based Smart TV Environmental Status Recognition Technology.”
References (25)
- et al.
ConaMSN: A context-aware messenger using dynamic Bayesian networks with wearable sensors
Expert Systems with Applications
(2010) - et al.
Landmark detection from mobile life log using a modular Bayesian network model
Expert Systems with Applications
(2009) - et al.
Interaction framework for home environment using speech and vision
Image and Vision Computing
(2007) Propagating imprecise probabilities in Bayesian networks
Artificial Intelligence
(1996)- et al.
Visualization and interaction of pervasive services using context-aware augmented reality
Expert Systems with Applications
(2008) - et al.
Context-aware end-to-end QoS qualitative diagnosis and quantitative guarantee based on Bayesian network
Computer Commucations
(2010) - et al.
The integration of home-automation and IPTV system and services
Computer Standards & Interface
(2009) - et al.
The psychology of human computer interaction
(1983) - Chen, Z. -Y. et al.(2006). Using semi-supervised learning to build Bayesian network for personal preference modeling in...
- et al.
A Bayesian method for the induction of probabilistic networks from data
Machine Learning
(1992)
The method of paired comparison
Understanding and using context
Personal and Ubiquitous Computing
Cited by (22)
Towards the design of personalized adaptive user interfaces for smart TV viewers
2023, Journal of King Saud University - Computer and Information SciencesAQUArIUM - A suite of software measures for HCI quality evaluation of ubiquitous mobile applications
2018, Journal of Systems and SoftwareCitation Excerpt :Even if the application provides precise results, it cannot cope with the uncertainty that is human nature. Several studies have been performed for usability evaluation (Kim et al., 2003) (Acampora et al., 2012) (Song and Cho, 2013) (Lee et al., 2008). In general, these studies did not use software measures for usability testing that consider the characteristics of ubiquitous applications such as context-awareness.
A BBN-based Method to Manage Adaptive Behavior of a Smart User Interface
2016, Procedia CIRPMethodology to Develop a Home Energy Management System Architecture
2022, Computacion y SistemasResearch on Visual Search Cognitive Characteristics of Adaptive Interface
2021, Tien Tzu Hsueh Pao/Acta Electronica SinicaPrototyping adaptive systems in smart environments using virtual reality
2019, International Journal on Interactive Design and Manufacturing