Research Project

PI, 2024.8 - 2025.7

In recent years, Web APIs designed based on REST (REpresentational State Transfer) have become mainstream. The IT industry, aiming to rapidly adapt to market changes, quickly release new features, and enhance service reliability and efficiency, has gradually shifted towards using containerized and Microservices Architecture (MSA) for application development and deployment. In this context, Cloud Native technologies are increasingly becoming a trend. Many applications are transitioning from monolithic to microservices architectures, and REST-based inter-service distributed transactions have become a critical issue. The project's main goal is to develop a REST-based distributed transaction mechanism for containerized microservices environments, addressing three critical aspects: achieving transaction agnosticism, handling single-point failures in transaction services, and ensuring varied isolation levels. The design will utilize the Sidecar container pattern for the first two challenges and incorporate previous research findings to tackle the third. The project will conclude by experimentally verifying the functionality, performance, and practical feasibility of this mechanism. The designed prototype system will be made available for future researchers and industry R&D teams for reference and use.

PI, 2023.8 - 2024.7

MQTT (Message Queuing Telemetry Transport) is a popular industrial standard that specifies a message-oriented application layer protocol designed for the publish/subscribe style of communications among IoT devices. Because of the nature of the message-oriented architectural style, MQTT systems are flexible and robust. However, the architectural style also brings challenges to the design of distributed transactions on MQTT. A transaction refers to a unit of work groping a set of operations, where these operations are either all committed or aborted. Also, ACID (Atomic, Consistent, Isolated, Durable) properties must be ensured when processing transactions. Compared to the traditional RPC-based distributed transaction technology, there is little literature that focuses on this topic and the queued transaction technology is less mature and incomplete. We have identified several research issues (mainly isolation issues) based on the findings of the previous project on MQTT transactions. The objective of this proposal is to deal with the isolation issues by introducing the concept of a “Coordination Kernel.” The research results will be implemented as a prototype system and design cases/scenarios to be referenced by future researchers or industry R&D teams. We will also conduct a set of experiments on the performance of transaction processing based on the prototype system.

PI, 2022.8 - 2023.7

Web of Things (WoT) refers to a “Web-ify” Internet of Things (IoT) environment designed based entirely on the REST (REpresentational State Transfer) architecture, where all devices have URIs that can be used not only for addressing but also for interacting with the resources. The main reason for applying the WoT concept to an IoT system is to take advantage of the welldeveloped Web technologies as well as the high flexibility and scalability of the Web. Meanwhile, although BLE is a widely adopted non-IP network protocol, developing BLE systems is not easy for most application developers. This project proposes a mechanism to transparently weave BLE devices into the WoT environment, called WeBLE (Web of BLE devices), so that the application developers can access and manage BLE services the same way as they are accessing WoT services, reducing the development efforts and management burden. WeBLE can be searched, managed, and accessed by WoT applications on IP networks through the WeBLE Gateway. During the research period of the project, we will focus on the two main issues: Findability and Accessibility. The research team will explore the design mechanisms including internal and external service management, directory services, IP-BLE address mapping, service mapping, and call forwarding, etc. We will implement the design prototype and conduct experiments to verify the design. The outcome of this project can be a baseline reference for researchers or R&D teams of related topics.

Co-PI, 2022.1 - 2023.12

“Surveillance capitalism” have human unconsciously become lost the freedom to be human. People believe that they can operate a machine under will. Yet, the logic of surveillance capitalism and instrumentarianism power actually has controlled people's behavior in an illegible way and has people become lost their solid sense of self. The implications of this project include the true reflection of society and the age of surveillance capitalism, and the realization of the politics of space in cyberspace. Two ways of this project are being investigated. One of them is to explore “the politics of space” in cyberspace, the other is to explore the “body politics” of virtual reality.

PI, 2021.8 - 2022.7

MQTT (Message Queuing Telemetry Transport) is a message-oriented communication protocol designed for IoT devices. From the perspective of software architecture, MQTT adopts the Message-Oriented architectural style, which decouples temporal and spatial relationships among messaging endpoints. On one hand, the decoupling enhances system flexibility and robustness, on the other hand, decoupling brings challenges to the design of message-oriented transactions. Compared to the general RPC-based distributed transaction technology, there is little literature that focuses on this topic and the message-oriented transaction technology is less mature and incomplete. The objective of this proposal is therefore to propose a flexible message-oriented transaction scheme for IoT environments. Specifically, in response to the variation of device capability in IoT environments, we propose a multi-mode approach to support different situations. Also, to improve the versatility of the design, the scheme is built based on MQTT 5. The research results will be implemented as a prototype system and design cases/scenarios to be referenced by future researchers or industry R&D teams.

PI, 2020.8 - 2021.7

MQTT (Message Queuing Telemetry Transport), a message-oriented application layer protocol designed for limited network bandwidth and resource-constrained devices, has received considerable attention and is widely adopted in the Internet of Things industry and the smart environments. The QoS (Quality of Service) mechanism of MQTT facilitates the applications to dynamically change the strategy of transmission impromptu. However, instead of end-to-end delivery guarantee, the QoS mechanism in the current version of MQTT only ensures the quality of delivery among the broker and endpoints. Besides, there is no way for the publishing endpoints to get the results of delivery. Also, the specification does not address the messaging retain policy when an endpoint is reconnected. The problems mentioned above lead to an unstable MQTT system and make the applications hard to debug. This proposal aims to deal with the aforementioned problems. Specifically, we extend MQTT (without modifying the packet format) so that it supports end-to-end guarantee delivery, receipt acknowledgment, and recovery mechanisms for the re-connect endpoints. Finally, the research results will be implemented as a prototype system and design cases/scenarios to be referenced by future researchers or industry R&D teams.

PI, 2019.8 - 2020.7

Due to the popularity of the Web, Web and related technologies have a profound impact on the research and development of the Internet of Things (IoT) and Smart Environments. Consequently, the concept of WoT (Web of Things) emerges by combining the Web, REST (Representational State Transfer), and IoT. Similar to other service-oriented systems, it is also essential to manage and coordinate software and hardware components dynamically. Among the existing service management mechanisms, the mDNS/DNS-SD specification developed by the IETF is considered to be promising as a service management mechanism for WoT, because it extends the DNS packet format and mechanisms and can be used without additional configuration of other system components. Also, the encoding scheme of DNS is relatively compact and thus requires less memory. However, mDNS/DNS-SD lacks many core service management functions such as service description and service invocation. Besides, the query mechanism in DNS-SD in inflexible and inefficient. As a result, this project aims to propose an integrated solution to deal with the problems mentioned above. We are going to design a mechanism that embeds the W3C WTM standard into DNS-SD, provides service access based on the REST principle to implement WoT service invocation mechanism, and develop flexible preference expression service search. Finally, we are going to optimize the performance of interactions among service components and verify the functionality, performance and practical feasibility through experiments and application scenarios. Finally, the research results will be implemented as a prototype system and design cases/scenarios to be referenced by future researchers or industry R&D teams.

Co-PI, 2019.8 - 2020.7

Over the past few years, blockchain technology has attracted much attention. The smart contract is a computer program that implements and executes transactions and business logic on the blockchain. Unlike traditional general-purpose programming languages, the modularization techniques of programming languages for the smart contracts are still immature. Specifically, a suitable modularization mechanism for crosscutting concerns is still not available. From software engineering’s perspective, the modularization of crosscutting concerns is a fundamental issue that can dramatically affect the contracts' maintainability and readability. As a result, reducing the redundant crosscutting logic is still a challenging task. This project developed a practically viable approach that supports Aspect-Oriented Programming (AOP) for smart contract development on Ethereum-based blockchains. We designed and implemented a set of extension to Solidity, a programming language runs on the Ethereum-based blockchains, that facilitates AOP semantics and behaviors. As a proof of concept, we realized the weaving module for the aspect extension via a preprocessor in an extended Solidity compiler. The feasibility of the proposed approach are demonstrated by a few case studies.

PI, 2018.8 - 2019.7

Recently, Data-centric and MOM (Message-Oriented Middleware) architectural styles have been widely adopted when designing Pervasive systems and IoT (Internet of Things) applications. As a result, MQTT (Message Queuing Telemetry Transport), a compact and efficient data exchange protocol for devices with limited computing capability and network bandwidth, receives considerable attention both in academia and industry. A Pervasive system relies on environmental data collected by the wireless sensors inter-connected by LLN (Low-power and Lossy Networks). Appearently, the data must be transferred from LLN to the IP network. Apparently, a gateway is the key to cope with the impedance mismatch between heterogeneous networks. In this project, we propose two architectures, namely, the Generic Gateway Integration and the Customized Gateway Integration for dealing with common issues when designing MQTT-LLN gateways. Specifically, we will investigate the efficient mechanisms for gateway discovery and binding, adapting MQTT protocol features, and deal with the issues caused by transmitting data between MQTT and LLN. The proposed designs will be realized as reference implementations of MQTT-LLN gateway prototypes. Validations and experiments will also be performed to verify the effectiveness of the proposed approach.

Co-PI, 2017.8 - 2018.7

As technology progresses, various types of wearable devices have surfaced in recent years and found practical applications in different areas. In a previous 3-year project titled “Interactive Performance Using Wearable Devices: Technology and Innovative Applications”, we have applied wearable technologies to assist the design and implementation of digital interactive performance from different perspectives, including platform infrastructure, network communication, interface design, and application development. Three different categories of interactive performances, namely, concert, dance and mobile virtual reality, have been conducted from 2015 to 2016. All performances have received positive acclaims. However, the tools that we have developed are not yet readily available to directors or performers with little background in computer hardware/software. In this project, we propose to extend and promote our previous research results by constructing a user-friendly platform that integrated essential components for scripting digital interactive performances with low entry barrier. The integrated tool, along with the instruction materials, will be tested and validated by arranging experiments in selected elementary or junior high schools.

PI, 2017.8 - 2018.7

Until recently, Smart Home technologies are still not widely deployed in most people’s living spaces. The main reason is that operations management mechanisms for Smart Home such as remote deployment, monitoring, and maintenance are not well-studied and only a few attempts have so far been made toward this aspect. CWMP, proposed by Broadband Forum, is a promising standard for realizing a Smart Home operations management platform. Last year, we investigated critical real-world operations issues of Smart Home, namely, newly installed, module purchasing and download, service starting, service update, service diagnosis, failure recovery, usage statistics, and billing. Based on these experiences, several design issues, namely, poor performance and scalability, poor domain model design and inappropriate web callback architecture, have been identified. The objective of this proposal is, therefore, to deal with the issues mentioned above by suggesting a set of new ways to design CWMP functionalities. The overall approach is based on the RESTful architectural style. Finally, the proposed designs will be realized as an operations management platform prototype. Validations and experiments will also be performed to verify the effectiveness of the proposed approach.

Co-PI, (Subproject II: Key Software Engineering Technologies for Developing Blockchain-Integrated Application Services) 2017.5 - 2020.4

In this project, we collaborated with Fubon Financial Holdings, a pioneer in the application of blockchain technology in Taiwan, to develop the key technologies and tools required for enterprise blockchain applications and conducted field validation to strive for official commercial operations. First, high-performance consensus algorithms had always been a major bottleneck in the application of blockchain technology, with the primary goal of improving the transaction throughput of private chains. In permissioned closed systems, we used mechanisms to prevent specific errors, allowing for the consideration of weak Byzantine faults and the opportunity to design more efficient consensus algorithms. Due to the nature of permissioned consensus algorithms, the fewer the faulty nodes, the higher the throughput. However, most permissioned consensus algorithms made assumptions about the number of faulty nodes. Therefore, as the number of faulty nodes decreased, these assumptions were improved, and we also designed a benchmarking tool to evaluate the performance of various consensus algorithms.

From a software engineering perspective, the learning and development threshold for blockchain applications remained high and cumbersome, with system integration and architecture still in a rudimentary stage, lacking systematic smart contract verification and testing mechanisms. To address these challenges, we developed middleware for automated deployment and management. Secondly, we explored the integration of blockchain with the Internet of Things (IoT) by investigating, analyzing, and implementing important non-functional requirement (NFR) attributes of the service architecture style for IoT integration, applying it to construct IoT integrated application scenarios. To tackle the challenge of the lack of systematic smart contract verification and testing mechanisms, we developed an automated verification tool supporting smart contract languages and collaborated with Fubon's domain experts to apply the Behavior-Driven Development (BDD) method to develop application scenarios and verify their effectiveness.

Finally, in the fast-paced financial trading market, the advantages and convenience of digital currency transactions were irreplaceable by cash transactions. However, balancing user privacy with regulatory oversight had always been a dilemma. On the other hand, Know Your Customer (KYC) played a very important role in financial services, whether in terms of the suitability of customer investments or from the perspective of anti-money laundering. If a bank failed to adequately understand and verify customer information, it could cause significant losses to customers or create operational difficulties for the company. We leveraged blockchain technology to design a blockchain-based digital cash payment and KYC mechanism.

PI, 2016.8 - 2017.7

The vision of Smart Home has been depicted clearly in 1995 by Bill Gates and the technologies that facilitate Smart Home services have also been researched and developed for over 20 years. Nevertheless, these technologies are still not widely deployed in most people’s living spaces. The main reason is that operations management such as remote deployment, monitoring, and maintenance are not well studied and developed, and only few attempts have so far been made toward this aspect. Moreover, most of the related works either focuses solely on home-related or on service-provider-side-related issues. In this project we aim to investigate six most critical real-world operations issues of Smart Home, namely, newly installed, module purchasing and download, service starting, service update, service diagnosis, failure recovery, usage statistics and billing. The proposed approaches will be realized in a real home network. Formal validations and experiments will also be performed to verify the effectiveness of the proposed approach.

Co-PI, 2016.8 - 2017.7

In recent years, with all kinds of embedded computing and sensing device miniaturization, we are able to have the device of a variety of embedded computing power in everyday items, furniture and household appliances. The so-called Smart Home vision is now considered practical gradually. As a result, the programming models and tools for Smart Home are also emerging on the research agenda for software researchers these days. Among other things, trigger-action programming (TAP) is a simple yet popular programming model that enables users to create rules that automate behavior of smart homes, devices, and online services. However, existing TAP systems trade expressivity for simplicity and thus lack the capability to support state manipulation often required in Smart Home and similar contexts, such as digital interactive performance. Recently, due to the advance of information and communication technology, art and culture are combined with technology mediated interaction between actors and their audience to create a new style of digital interactive performance. However, there are still many technical barriers in developing digital interactive performance. Essentially, the scripts for such performance are usually fixed, or with limited programmability. As a result, one performance will be repeated played or requires a lot of efforts to modify it. This project aims to address the issue of programmable scripts with a visual editor and supporting execution environment. In particular, we focus on helping performance director to develop programmable scripts that enable the flexible interaction between physical characters and their virtual counterparts. Our editing tool allows end-users to write scripts using drag-and-drop elements we developed on top of the Blockly framework. Besides detailed descriptions of the design and implantation of our tool, preliminary evaluation results are also presented.

Co-PI, 2015.8 - 2018.7

With the aging population, the number of dementia patients is rapidly increasing, and the impact and burden of dementia on individual health, family economy, and national finances are becoming increasingly severe. Caregivers not only have to manage the daily lives of dementia patients but also respond to their individual needs, dealing with diverse and challenging neuropsychiatric symptoms. The complex care requirements often impose significant stress and burden on caregivers. In light of this, this project utilizes mobile technology to “observe,” “recognize,” “empathize,” “eliminate crises,” and “promote healthy behaviors” for dementia patients, caregivers, and medical teams. By constructing a smart care environment, utilizing first-person perspective video recording and activity analysis, and developing a personalized living and social information platform, the project aims to assist with the daily care of dementia patients, reduce caregiver stress and care costs, and improve the quality of life for dementia patients. Additionally, the project will explore the impact of “filial piety obligations” and “family resilience” on dementia care within families, further assessing whether the developed mobile technology provides tangible benefits.

PI, 2015.8 - 2016.7

We can perceive the advent of the Smart Environments attributed to rapid emerging of embedded and tiny intelligent devices and sensors. However, due to the significant complexity of deployment and maintenance efforts, the degree of market acceptance is still low. Specifically, there is little research addressing the robustness and failsafe issues for Smart Home systems. Consequently, the research paradigm in this field has been migrating from prototyping smart homes to daily maintenance issues in real-world scenarios. In this project, we aim to investigate two of the most critical real-world issues of Smart Home deployment, namely, the bootstrapping and robustness of the system. We propose a bootstrapping procedure and an enhanced hybrid failure detection and recovery architecture. These schemes will be realized based on a popular service management protocol in home network, UPnP (Universal Plug and Play). Formal validations and experiments will also be performed to verify the effectiveness of the proposed approach.

Co-PI, 2015.8 - 2016.7

In recent years, with the rapid development of ICT technology, all kinds of embedded computing and sensing device miniaturization, we were able to have the device of a variety of embedded computing power in everyday items, furniture and household appliances. These miniature-sensing devices can sense the environment and user-related information to form the so-called context. With the context information, we can develop application systems that will infer from the context the presumed intention of the user to provide appropriate services, thus forming the so-called context-aware smart environment. However, due to the distributed and concurrent nature of the devices operating in a smart environment, it is not easy to develop such context-aware applications using traditional imperative programming. Developers have to deal with many complex issues such as asynchronous events and temporal requirements, which are tedious and error-prone. Indeed, many researchers have proposed to follow the reactive programming style to develop applications for smart environments. This project aims to develop a Scala-based domain-specific reactive language for programming smart environments that supports functional reactive programming. In this project, we shall first formulate the domain-specific abstractions appropriate for programming smart environments based on functional reactive programming. Then we shall implement these domain abstractions and their associated operations into Scala using language-embedding techniques, thus resulting in a domain-specific embedding language (DSEL) in Scala. In the meantime, we shall also develop a middle-level facility based on Scala Akka library to provide the interface between our language and the underlying smart environments that works as a simulation environment for our project. Finally, we shall implement a few applications using sensors on top of the simulation environment to demonstrate and evaluate our language and its facilities.

Co-PI 2014.11 - 2018.4

The decreasing cost of printers and materials has unleashed the creativity of designers and average users alike, making self-fabrication an affordable, enjoyable and educational activity. Introducing 3D printing technology to the classroom is now a consensus among educators and government officials. The current tools available on the market, however, are not designed for educational purposes in classroom settings. A stable software platform with low maintenance cost and friendly user interface should be devised to suit the needs of the target users: middle and elementary students. Toward this objective, our team consisting of experts from computer science and instructional technology has designed and implemented an integrated solution consisting of the following components: 3D scanning, intuitive 3D modeling, editing and merging, model searching, file sharing, 3D caricature generation and 3D printing. Curriculum has been created and activities have been planned according to the functionalities provided by our platform. We performed user study and gathered feedback and comments during the field test phase. In the last year of the project, we have built an integration platform by integrating software services developed by all sub-projects. Users are able to access these services via common Web APIs. One of the main objectives of this project is to design a 3D modeling tool to scan and create the photographic 3D models, making it easy to incorporate all kinds of designs curricular activities. We focused on user experience research and analysis, and as a starting point, tried to establish a standardized user behavior analysis and evaluation criteria. In the last year, we have also developed a cross-platform intuitive modeling system, QModel Creator, to provide a straightforward and easy-learning modeling system to creative cubic style model for K-12 students. We have also expanded the functions of QModel Creator by editing mortise-tenon connector to compose multiple parts into a complex model. Besides, we implement a 3D model retrieve system to find the best match 3D model from a 2D sketch image. In addition, we have developed a collaborative learning system called “3D Model Co-Learning Space” to support the storage, sharing, display, and discussion of 3D models. On our Resource Integration Portal website, we also provide basic 3D printing working knowledge and teaching plans for teachers. In this project, we have also developed and implemented three innovative courses for Earth Science and Science and Technology classes in a senior school in Taipei. In these courses, we have designed a 3D-printed earth model that has adjustable axis for exploring the principles of seasons and day and night in science inquiry. A module of roller coaster launcher and rails were also developed for students to test their 3D-printed cars. The software tools and teaching materials developed in this project have been released to the public at no cost in the hope that they can become active catalysts for accelerating the widespread adoption of 3D additive manufacturing curriculum in our educational systems.

Co-PI, 2014.9 - 2017.7

It is commonly agreed that Taiwan’s performance art and culture industry is in the leading position of the Chinese culture area. However, the position is being challenged in recent years due to the raise of China’s culture industry and the lack of economies of scale in the domestic market. Because of the popularity of digital technologies, the sale of sound records is significantly decreased. On the other hand, concert-like performance has become the main source of profits of the performance art and culture industry. In traditional concerts, the audiences play passive roles since they just sit and watch the performance. However, in modern concerts, interacting with the audience has become one of the critical factors of a successful performance. Recent trends suggest that wearable computing technologies can be massively used to promote the degree of interaction and to make a performance more attractive. For instance, motion capture technologies have been widely used in accurately capturing the postures of performers in the movie, game and computer animation industries. More recently, some researchers also propose to use wearable computing devices in the real-time interactive performance. Taking the opening show of the 2014 Winter Olympics for an instance, the dancers wear a lot of sensing devices which fit them to the background snow leopard animation characters. To sum up, the objective of this project is to investigate the approaches that are able to combine wearable computing technologies with the digital interactive performance from the platform, network, interface, and application’s perspectives. The results of this project are able to help Taiwan’s performance art and culture industry to survive and success in the keen competition.

PI, 2014.8 - 2015.7

A Pervasive-computing-enriched smart environment, which contains hundreds of embedded and tiny intelligent devices and sensors coordinated by service management mechanisms, is capable of anticipating intensions of occupants and providing appropriate services accordingly. Although there is a wealth of research achievements in recent years, the degree of market acceptance is still low. The main reason is that most of the devices and services in such environments depend on specific platform or technology, making it hard to develop an application by composing the devices or services. Meanwhile, the concept of Web of Things (WoT) is becoming popular recently. Based on WoT, the developers are able to build applications based on well-known web tools or technologies. Consequently, the objective of this project is to design and to implement a WoT-driven service management mechanism for a small scale smart environment. Formal validations and experiments will also be conducted to verify the effectiveness of the proposed approach. We expect that the results of this project can be a solid foundation of realizing the vision of “end user programmable smart environments”.

Co-PI, 2014.8 - 2015.7

Software as a service (SaaS) is an emerging service model of cloud computing. Its central defining characteristic is the ability for clients to use a software application on a pay-as-you-go subscription basis. However, to be economically sustainable, a SaaS application must leverage resource sharing to a large degree by accommodating different clients of the application while making it appear to each that they have the application all to themselves. In other words, a SaaS application must be a multi-tenant application. There are many facets in the design of multi-tenant SaaS applications. In this project, we focus on developing multi-tenant data architecture that not only enables sharing database among tenants but also allows tenants to customize their table schema, which is a non-trivial task. A general approach to support such architecture is a middleware-level facility that supports the mapping of multiple single-tenant logical schemas in the application to one multi-tenant physical schema in the database. We follow this approach and propose to develop a SQL rewriting mechanism that can transparently transform tenant-specific logical queries into corresponding physical queries for Universal Table, a widely adopted and industry verified schema-mapping technique. Our SQL rewriting rules will be formalized in terms of standard operations in relation algebra so that it will be easy to formulate analytical complexity metrics for predicting the overhead of schema mapping in the engine. We shall conduct a series of performance experiments to evaluate the proposed schemes. The performance results will be compared with the analytical complexity results derived from the formalized SQL rewriting rules. Besides, as object-relational mapping (ORM) tools are the mainstream tools in application development, we shall integrate our SQL rewriting engine into an open source ORM tool to form a data management framework for multi-tenant SaaS applications. A salient feature of this framework will be the support of user customization of domain. Finally, we shall develop a multi-tenant SaaS application using our framework to demonstrate our approach. Performance overhead of this application will also be measured and reported.

Co-PI, 2012.8 - 2015.7

Despite the prolonging of citizens’ life expectancy, the elderly cancer survivors still need to face disease itself and the discomfort and stress brought by the treatment. How to use emerging technology to enhance the quality of life for elderly cancer survivors in Taiwan is obviously one of the most important issues in the next few decades. Among them, physical activity and health management, social activity, and quality of sleep are the key points to determine the quality of life of elderly cancer survivors. This project aims to integrate the experts from the fields of computer science, electrical engineering, cancer medicine, nursing and psychology to build a Cancer-survivor HEalth Enhancing and Recovery System (CHEERS) focusing on “Physical Activity and Health Management”, “social mobility,” and “sleep quality” which are the key topics for the quality of life of elderly cancer survivors. First we focus on the construction of a friendly home long-term care system for elder cancer survivors, which is capable of gather bio and context information faster and more accurate. Based on the information mentioned above, the system then infers Health Alert Context (HAC) which helps care workers determine if the treatment is harmful or being invalid. We will develop a private cloud analysis mechanism to support care workers performing decision-making and to improve the understanding of long-term history of cancer evolution and statistics. Elderly cancer survivors after surgery and chemotherapy, in addition to worry about the illness, are more likely to be faced the frustration, disappointment, depression, and other psychosocial problem. The side effect of chemotherapy may bring fatigue, hair loss, and even restrict the mobility of elderly, further impact the social skills of them. Our second goal is to enhance the social support among elderly and their family and friends and to promote the mobility of the elderly through social network, mobile devices, and interactive multimedia. We propose three social interactive interventions, including web-based interaction, group interaction, and out-side interaction. We aim to lead the elderly to go out-side and face the crowd. It provides a channel for the elderly and their family and let them could communicate and care about each other and will strengthen the social support and pursue to a better quality of life. Cancer survivors usually suffer from the side-effects of the treatment. Sleeping problem is among them the most serious one. Therefore, we propose two systems to deal with this problem: Breath Relaxation Multimedia system and Intelligent Sleep Diary System (i-Sleep Diary). Breath Relaxation Multimedia system assists the patients to practice breath relaxation exercise before sleeping; i-Sleep Diary can help the doctors monitor patients through their sleeping cycle and give advices to improve their sleeping quality. In order to evaluate the effect of the developed CHEERS technologies on promoting the quality of life (QoL) and to explore the associations among “attitude toward aging,” “attitude toward cancer,” and QoL, we also establish three measurement tools such as the “Taiwanese Cancer Survivors Quality of Life” scale. We plan to initiate a clinical application model that integrates the CHEERS technologies, the QoL evaluation tools, and a clinical feedback-consultation process, attempting to maximize the enhancement of QoL of the elderly cancer survivors. The model could not only monitor and improve the related symptoms, but the valuable observation data could also help the clinical professions while discussing with the survivors. By combining the subjective data measured by the QoL evaluation tools and the observational data, clinical professions could help the survivors to clarify their problems and to enhance their QoL more effectively.

PI, 2013.8 - 2014.7

A Pervasive-computing-enriched smart environment, which contains hundreds of embedded and tiny intelligent devices and sensors coordinated by “service management mechanisms”, is capable of anticipating intensions of occupants and providing appropriate services accordingly. Based on the experiences of realizing several real-world smart environments, it is observed that to acquire high-level contexts, the key to provide attentive services such as human activities, usually involves analyzing and identifying “causality” and “temporal ordering” relationships among a bulk stream of sensor readings. Besides, there are relatively few researchers in this field investigating the issue. We notice that Complex Event Processing (CEP) useful for dealing with the issues mentioned above. From CEP’s point of view, incoming sensor data is a stream of events, the logical units for analyzing and transforming events form a Complex-event Processing Network (CPN). In this project, we aim to investigate both theoretical as well as empirical aspects for integrating CEP concepts with PerSAM, a formal service application management model. Formal validations and experiments will also be performed to verify the effectiveness of the proposed approach.

PI, 2012.9 - 2013.7

We can perceive the advent of the Smart Environments attributed to rapid emerging of embedded and tiny intelligent devices and sensors. However, most of current systems for Smart Environments are still being developed based on the so-called “the system is the application” philosophy, causing the developers to take care of all technical details from ground up. There is relative few research focuses on the middleware aspects, so that the independent research achievements or results are not interoperable. In addition, the concept of the “Ambient”, which refers to a bounded place with computing capability, is one of the most important issues when implementing a system for Smart Environments. An Ambient is typically composite-capable (i.e. an Ambient can be embedded within one another) and has complex structure. Residents may have different preferences when they are located in different Ambient. Hence, the results of a service composition are not optimized if the concept of Ambient is not taken into account. As a result, we aim to investigate an Ambient-oriented service management issues systematically from both theoretical and practical aspects. After the issues have been clarified, we will propose a new Ambient-oriented service management scheme and then realize the scheme based on a popular service management protocol in home network, UPnP (Universal Plug and Play). Formal validations and experiments will also be performed to verify the effectiveness of the proposed approach.

Postdoc 2011.1 - 2012.1

In recent years, green smart energy-saving technologies for home environments have gained increasing attention. These technologies must be easy to deploy, low-cost, and non-intrusive to the user's life. To enhance practicality, the technology must also consider the living environment shared by multiple users, ensuring that energy-saving adjustments do not cause discomfort to the residents. The primary goal of this project is to design a practical green smart home system that achieves energy savings by simultaneously considering environmental information and employing an adaptive user comfort evaluation. Additionally, we have designed a series of highly integrated and easy-to-deploy multi-dimensional sensors and an energy-saving inference decision engine to achieve optimal energy-saving control.

Postdoc 2008.1 - 2011.10

According to the data from Council for Economic Planning and Development, the aging rate of population in our country is the 2nd place in the world. In 2005, one survey also pointed out that there are about 580,000 person-time of hospitalization of elders aged 65 or more in one year. Persistent or transient decline of cognitive, nutritional, and physical status was noted in elder patients during hospitalization or six months post hospitalization. Declines of cognition and ADL functions also lead to increase of average stay days in hospitals, increase of re-admission rate, and increase of death rate, especially for the elder patients requiring gastrointestinal surgery. The research team led by Prof. Li-Chen Fu in National Taiwan University has built great partnership over various domains in our previous project, in which a human-centric recovery environment was constructed for several kinds of patients. In our new project, we further target elder (aged 65 or more) patients who receive gastrointestinal surgery. We aim at developing effective intervention plans to prevent declines of their physical, cognition, and nutrition functions so that the quality of their life will be highly improved. Subproject I will build an intelligent care environment composed of two components: In the intelligent environment, we will establish a non-invasive wireless sensor network system to collect information about elders and environments for behavior and emotion recognition. The other component, intelligent caring robot, will play the role of a “digital carer,” who always accompanies the elders, provides suitable services, and even serves as an agent for family, doctors, and nurses at the remote site. Subproject II, intelligent cognitive multimedia, proposes the concept of a “traveling ward,” which consists of virtual carriage, reminiscence multimedia, and reminiscence network. Virtual carriage will make the patient feel as if he is driving for a trip. During the virtual trip, we can do the reality orientation. Reminiscence multimedia will collect users’ memories through an interactive story editing platform, and then the organized contents will be presented by our developed eFovea display system to help the nurses to do reminiscence therapy. Based on reminiscence multimedia, reminiscence network will provide an intuitive interaction interface to facilitate communication between patients in different places so as to maintain their social relationship. The physician, nursing, psychosocial, and physical therapy specialists of our sub-project III will evaluate the cognitive, nutritional, and physical changes in the above-mentioned patients. The risk factors will be examined, and then effective prevention plans and post-surgery therapy will be proposed. Meanwhile, a prototype of persuasive mug for hydration of elderly patients and a prototype of persuasive brush for oral health of elderly patients will be created to enhance the postoperatively nutritional status. The robotic and multi-media technology from the other two sub-projects, the persuasive objects, and physical therapy will cooperate to construct a clinical care system for postoperatively elder patients. Finally, we will verify the achievement through clinical evaluation by the specialists in this project.

Doctoral Research Associate 2008.2–2012.1

Responsibilities: I was one of the earliest members to join this project. I was responsible for writing the section on “Smart Care Technology” in the project proposal and participated in the establishment of the INSIGHT Living Lab. I designed the architecture of the integrated platform for the INSIGHT Living Lab. After the center was established, I mainly handled the R&D, results integration, and major demonstration activities related to “Smart Care Technology.” From 2008 to 2009, I was a member of the INSIGHT Tech Team.

The rapid advancement of information and communication technology (ICT) has not only brought revolutionary changes to society as a whole, ushering humanity into a new era of the information society, but also significantly impacted individual lives. How to leverage ICT to mitigate the impacts of declining birth rates and an aging population, as well as to alleviate the pressures of global environmental changes and energy shortages, has become a key focus of global smart living technology R&D and applications. In view of this, our center, focusing on smart homes, has proposed five main research and development axes: “Smart Living Technology and Humanities Issues,” “Smart Care Technology,” “Home Healthcare,” “Energy and Sustainable Environment,” and “Wireless Sensor Networks.” These axes form the research blueprint for the next five years. Based on the current R&D achievements of three integrated projects—“Smart Living Space Technology,” “Advanced Research on Wireless Sensor Network Technology,” and “Future Product Concept Design of Advanced Engineering Technology”—we integrate various R&D units from industry, government, academia, research institutions, non-governmental organizations, and international partners to present and demonstrate future scenarios of smart living technology in the “home.” Most successful examples of integration and innovation in business operations have strong core R&D capabilities. This center adopts this concept, leveraging strong core R&D capabilities combined with external creativity and the center's operational mechanisms to create a new vision of smart living.

Website: [http://insight.ntu.edu.tw/](http://insight.ntu.edu.tw/)

Doctoral Research Associate 2010.11 - 2011.6

Responsibilities: This project was independently executed by me, and I conducted eight technical transfer workshops at ITRI.

Traditional business-to-business (B2B) integration is mostly implemented based on a direct-connection model. As more services are deployed within the software, this approach can lead to difficulties in system management and functionality expansion. To address this issue, IBM proposed the concept of the Enterprise Service Bus (ESB).

An ESB is centered around a platform that integrates heterogeneous services based on a message-oriented architecture and common standards, allowing various heterogeneous components to communicate and call each other in a loosely coupled manner. The main functions of an ESB include coordinating service calls between heterogeneous components, integrating message transmission between these components, and managing message flow.

In 2005, the Java community completed the first common standard for ESB, known as “Java Business Integration (JBI),” codename JSR-208. Based on the JSR-208 standard, several open-source projects for implementing ESB have been launched in recent years, with more mature and stable versions released around 2008. Among these, Apache ServiceMix has become one of the most widely adopted.

This research project utilized Apache ServiceMix as the core to design a cloud service integration platform software architecture based on JBI, integrating the following (open-source) technologies:

• Apache ServiceMix (ESB)
• Apache ActiveMQ (JMS)
• Apache Camel (Message Routing Engine)
• Apache ODE (BPEL4WS)
• Apache CXF (SOAP/Restful Web Services)
• Apache Felix/Karaf (OSGi)

Doctoral Research Associate 2010.1 – 2010.12

Responsibilities: In this project, I was a team member, primarily involved in conceptualizing and writing the project proposal, and leading certain key technical implementations.

The project primarily integrates cloud computing technology to develop ADL (Activities of Daily Living) recognition and context-aware service inference technology. The system utilizes sensors in the home to collect environmental parameters, which are uploaded to the cloud. The cloud then rapidly infers the necessary results for ADL, which are fed back into the system’s ADL inference model for enhancement. The main components include:

• Collecting sensor data through the Wireless Sensor Network (WSN) in a smart home.
• Performing ADL recognition based on the sensor data to understand user activities and needs.
• Selecting services based on user background information to provide timely and effective services.
• A user-friendly remote home care platform as an application example.

The prototype of this project uses Hadoop and HBase to set up a cloud cluster and utilizes the MapReduce framework and Apache Mahout for data analysis.

Doctoral Research Associate 2009.1 – 2009.12

Responsibilities: I was the team leader for this project, overseeing and monitoring all project executions, writing the project proposal, mid-term, and final reports, and delivering oral reports at all project-related meetings. I independently developed the GLIF Engine and Compiler.

In recent years, Taiwan has experienced rapid economic development, improved living and education standards, and widespread healthcare, leading to a significant increase in the elderly population. Statistics indicate that by 2020, the elderly population in Taiwan will reach 14%, making it the second fastest-aging country in the world after Japan. This situation poses a significant burden on the government’s national health insurance and the lives of children caring for elderly parents. We anticipate that smart home and telecare applications and services will flourish in the future, ranging from daily living, health care to home safety, developing an integrated home care service network that will be a highly anticipated technology field.

Currently, most commercialized home care service applications are limited to the measurement and uploading of physiological information, which can only indirectly achieve the function of “care services.” To establish a better home health care environment, this research project aims to apply the GLIF (GuideLine Interchange Format) medical care guideline standards to the home health care environment. By integrating professional care knowledge with home environment sensors, devices, and networks, we developed the Home-care Guideline Execution Engine (H-GLEE) technology. This engine serves as the core for building an “Intelligent Home Health Care Management Platform,” on which various home care service components and scenarios can be developed and deployed to verify and improve H-GLEE functions.

Doctoral Research Associate 2008.3 – 2009.2

Responsibilities: As the team leader, I oversaw and monitored all project executions and delivered oral reports at all project-related meetings. I also developed the SWE (OGC Sensor Web Enablement) protocol stack adapter to convert sensor data into OGC Sensor Web Enablement O&M format.

The primary objective of this commissioned research project is to provide an application platform for wireless sensor networks in future smart homes based on the OGC Sensor Web Enablement (SWE) Web services architecture. Through a Web software interface, it offers management, control, and data access functionalities for the wireless sensor network system. This project focuses on sharing and integrating information obtained from the wireless sensor network in smart homes with the external world through SWE, providing several demonstration application scenarios. The technologies developed in this project enable telecom operators to extend their service scope further into the home life of users, leveraging their existing telecom and network platform integration advantages to offer more home life services to consumers.

Doctoral Research Associate 2006.8 – 2009.7

Responsibilities: After July 2007, I became the team leader, overseeing and monitoring all interdisciplinary research activities and delivering oral reports at all project-related meetings.

The “home” serves not only as a shelter for individuals and families but also fulfills psychological needs and provides comfort. While current technology can offer a wide range of convenient functions, it still falls short in supporting human-centric needs. The goal of establishing an “Attentive Home” is to provide anticipatory and proactive computer technology in the home environment to meet human physiological, emotional, and cognitive needs, thereby achieving maximum comfort.