IBC2023 Tech Papers: Daily Context-Adaptive Presentation driven by Personal Data Store

Abstract

Smart environments are expected to enable broadcasters to develop “context-adaptive content presentation” technology, through which heterogeneous smart and Internet-of-Things (IoT) devices in the environment work together to autonomously select and present appropriate content according to a user’s context. However, it is impractical for a single broadcaster to implement all the functions in heterogeneous user environments. Therefore, a loosely coupled mechanism that allows devices and software developed by different manufacturers and service providers to work flexibly in various combinations is required. Thus, we designed a system architecture that realizes content presentation based on the user’s moment-to-moment situation by utilizing context recognition and a personal data store. We prototyped the system and confirmed its feasibility even if the context estimation and content presentation modules were constructed independently. The result demonstrates the scalability of our system architecture and the potential impact of adaptive media in the smart era.

Introduction

The Internet has facilitated people’s ability to choose and access information fit to their preferences based on a wealth of available data. However, individuals have limited time to access information in their daily lives, meaning that access to information of relatively low- interest decreases as access to information of high-interest increases. Therefore, certain users may have less access to information that should be widely shared, such as socially relevant public and political topics, resulting in unequal levels of knowledge in relation to those with more access (1). This can lead to a split in public opinion. Broadcasters primarily transmit various information through broadcasting. However, owing to changes in people’s information contact styles, it has become difficult to disseminate such information sufficiently through broadcasting alone. Therefore, we are exploring ways to extend the broadcasting service platform and boost the convenience of the services, thereby increasing the opportunities to present information to users.

One possible approach is smart technologies for living environments using the Internet-of- Things (IoT) and smart devices. Efforts are underway across industries to create smart cities, homes, and other living environments (2). These smart environments are expected to enable users to receive services in a personalized manner in their current context with minimal efforts. If our living environment changes in such a way, broadcasters can make their services readily accessible to broader users than today. For example, suppose a user watches a video on TV in the living room but interrupts viewing and moves to another room. If the smart environment detects the user’s movement and automatically continues the video on the device in a new location, the user is encouraged to continue watching it. In another example, a user performs daily activities (e.g., brushing their teeth) and feels bored. By detecting such user requests and presenting content (e.g., the latest news) through a nearby device (e.g., a smart mirror or communication robot), the system can motivate the user to watch the content while performing daily activities (3). It contributes to the broadcaster’s mission of delivering information about the public interest to the audience.

The difficulty of constructing a complex system is a challenge for broadcasters in realizing context-adaptive content presentation. The system requires several complicated procedures: detection of events based on IoT sensor data, estimation of the current situation, selection of an appropriate presentation method, and presentation through the coordinated control of smart/IoT devices. Implementing all the functions in heterogeneous user environments is impractical for a single broadcaster. Therefore, a loosely coupled mechanism that allows devices and software developed by different manufacturers and service providers to work flexibly in various combinations is required.

In this study, we have designed a “context-adaptive content presentation” system architecture to realize content presentation according to the user’s moment-to-moment situation by utilizing knowledge-driven context recognition and a personal data store (PDS). The architecture realizes context-adaptive presentation, even if the context estimation and content presentation modules are constructed independently.

The context-adaptive approach to content presentation proposed in this study was developed as one of the technologies comprising web-based broadcast media, which we are exploring as part of this research. Web-based Broadcast Media is a new media realized by combining broadcasting with the web. In designing the architecture of the context- adaptive presentation system, we considered consistency with a set of technologies that we are developing for web-based broadcasting media. This study also provides an overview of web-based broadcast media and their technology groups.

The remainder of this paper is organized as follows: Section 2 provides an overview of web- based broadcast media and their technologies. Section 3 describes the requirements for a context-adaptive presentation system. Section 4 describes the structure of our proposed context-adaptive presentation system. Section 5 describes the implementation of a typical use case of a context-adaptive presentation in a smart home. Additionally, the evaluation of the proposed system is described. Finally, Section 6 presents our conclusions and discusses future work.