Seok-Won Lee

A goal-driven approach for adaptive service composition using planning

Abstract Due to the demands of complex services in our daily life, we must deal with various application devices in the dynamic and constantly changing ubiquitous computing environment with more adaptive and personalized application services. However, it becomes more difficult to realize users’ needs due to the ever increasing complexity of the task and the limitation of the capability of a single service alone. Creation of a new service on demand to carry out more complex tasks is possible through the composition of existing services. We have developed a preliminary prototype that supports dynamic service composition in a smart workspace environment. The system is able to achieve the desired user-tailored goals by composing the combination of available services in the smart office domain.

Probabilistic Risk Assessment for Security Requirements: A Preliminary Study

Risk assessment is a critical decision making process during the Security Certification and Accreditation (C&A) process. However, existing infrastructure-wide C&A processes in real world are challenged by the ever increasing complexity of information systems and their diverse socio-technical operational environments. The lack of an explicit model and the associated uncertainties of software behavior are two main reasons that directly impact the effectiveness of risk assessment as well as the subjective decisions made based on the different level of domain expertise. In this paper, we propose a method for a probabilistic model driven risk assessment on security requirements. The security requirements and their causal relationships are represented using MEBN (Multi-Entities Bayesian Networks) logic that constructs an explicit formal risk assessment model that supports evidence-driven arguments. The proposed approach is described by using real-world C&A scenarios to show not only its feasibility for security requirements risk assessment but also its effectiveness for the sensitivity analysis to identify critical influences among information entities in a complex and uncertain operational environment.

Evaluating the use of model-based requirements verification method: A feasibility study

Requirements engineering is one of the most important and critical phases in the software development life cycle, and should be carefully performed to build high quality and reliable software. However, requirements are typically gathered through various sources and represented in natural language (NL), making requirements engineering a difficult, fault prone, and a challenging task. To address this challenge, we propose a model-based requirements verification method called NLtoSTD, which transforms NL requirements into a state transition diagram (STD) that can be verified through automated reasoning. This paper analyzes the effect of NLtoSTD method in improving the quality of requirements. To do so, we conducted an empirical study at North Dakota State University in which the participants employed the NLtoSTD method during the inspection of requirement documents to identify the amibiguities and incompleteness of requirements. The experiment results show that the proposed method is capable of finding ambiguities and missing functionalities in a set of NL requirements, and provided us with insights and feedback to improve the method. The results are promising and have motivated the refinement of NLtoSTD method and future empirical evaluation.

User-Independent Activity Recognition via Three-Stage GA-Based Feature Selection

Advancement in wireless sensor networks gave birth to applications that can provide friendly and intelligent services based on the recognition of human activities. Although the technology supports monitoring activity patterns, enabling applications to recognize activities user-independently is still a main concern. Achieving this goal is tough for two reasons: firstly, different people exhibit different physical patterns for the same activity due to their different behavior. Secondly, different activities performed by the same person could have different underlying models. Therefore, it is unwise to recognize different activities using the same features. This work presents a solution to this problem. The proposed system uses simple time domain features with a single neural network and a three-stage genetic algorithm-based feature selection method for accurate user-independent activity recognition. System evaluation is carried out for six activities in a user-independent setting using 27 subjects. Recognition performance is also compared with well-known existing methods. Average accuracy of 93% in these experiments shows the feasibility of using our method for subject-independent human activity recognition.

Weed Image Classification using Wavelet Transform, Stepwise Linear Discriminant Analysis, and Support Vector Machines for an Automatic Spray Control System *

We tested and validated the accuracy of wavelet transform along with stepwise linear discriminant analysis (SWLDA) and support vector machines (SVMs) for crop/weed classification for real time selective herbicides systems. Unlike previous systems, the proposed algorithm involves a pre-processing step, which helps to eliminate lighting effects to ensure high accuracy in real-life scenarios. We tested a large group of wavelets (46) and decomposed them up to four levels to classify weed images into weeds with broad leaves versus weeds with narrow leaves classes. SWLDA was then employed to reduce the feature space by extracting only the most meaningful features. Finally, the features provided by SWLDA were fed to the SVMs for classification. The proposed method was tested on a database of 1200 samples, which is a much larger database size than that studied previously (200-400 samples). Using confusion matrices, the crop/ weed classification results obtained using different wavelets at different decomposition levels were compared, and this approach was also compared with existing techniques that use statistical and structural approaches. The overall classification accuracy obtained using the symlet wavelet family was 98.1%. These results represent an improvement of 14% in performance compared with existing techniques.

Model-based requirements verification method

ContextRequirements engineering is one of the most important and critical phases in the software development life cycle, and should be carefully performed to build high quality and reliable software. However, requirements are typically gathered through various sources and are represented in natural language (NL), making requirements engineering a difficult, fault prone, and a challenging task. ObjectiveTo ensure high-quality software, we need effective requirements verification methods that can clearly handle and address inherently ambiguous nature of NL specifications. The objective of this paper is to propose a method that can address the challenges with NL requirements verification and to evaluate our proposed method through controlled experiments. MethodWe propose a model-based requirements verification method, called NLtoSTD, which transforms NL requirements into a State Transition Diagram (STD) that can help to detect and to eliminate ambiguities and incompleteness. The paper describes the NLtoSTD method to detect requirement faults, thereby improving the quality of the requirements. To evaluate the NLtoSTD method, we conducted two controlled experiments at North Dakota State University in which the participants employed the NLtoSTD method and a traditional fault checklist during the inspection of requirement documents to identify the ambiguities and incompleteness of the requirements. ResultsTwo experiment results show that the NLtoSTD method can be more effective in exposing the missing functionality and, in some cases, more ambiguous information than the fault-checklist method. Our experiments also revealed areas of improvement that benefit the methods applicability in the future. ConclusionWe presented a new approach, NLtoSTD, to verify requirements documents and two controlled experiments assessing our approach. The results are promising and have motivated the refinement of the NLtoSTD method and future empirical evaluation.

Interactive requirements validation for reactive systems through virtual requirements prototype

Adequate requirements validation can prevent errors from propagating into later development phases, and eventually improve the quality of software systems. However, validating natural language requirements is often difficult and error-prone. An effective means of requirements validation for embedded software systems has been to build a working model of the requirements in the form of a physical prototype that stakeholders can interact with. However, physical prototyping can be costly, and time consuming, extending the time it takes to obtain and implement stakeholder feedback. We have developed a requirements validation technique, called Virtual Requirements Prototype (VRP), that reduces cost and stakeholder feedback time by allowing stakeholders to validate embedded software requirements through the interaction with a virtual prototype.

SQ^(2)E: An Approach to Requirements Validation with Scenario Question

Adequate requirements validation could prevent errors from propagating into later development phase, and eventually improve the quality of software systems. However, often validating textual requirements is difficult and error prone. We develop a feedback-based requirements validation methodology that provides an interactive and systematic way to validate a requirements model. Our approach is based on the notion of querying a model, which is built from a requirements specification, with scenario questions, in order to determine whether the models behavior satisfies the given requirements. To investigate feasibility of our approach, we implemented a Scenario Question Query Engine (SQ2E), which uses scenario questions to query a model, and performed a preliminary case study using a real-world application. The results show that the approach we proposed was effective in detecting both expected and unexpected behaviors in a model. We believe that our approach could improve the quality of requirements and ultimately the quality of software systems.

REASSURE: Requirements elicitation for adaptive socio-technical systems using repertory grid

Abstract Context Socio-technical systems are expected to understand the dynamics of the execution environment and behave accordingly. Significant work has been done on formalizing and modeling requirements of such adaptive systems. However, not enough attention is paid on eliciting requirements from users and introducing flexibility in the system behavior at an early phase of requirements engineering. Most of the work is based on an assumption that general users’ cognitive level would be able to support the inherent complexity of variability acquisition. Objective Our main focus is on providing help to the users with ordinary cognitive level to express their expectations from the complex system considering various contexts. This work also helps the designers to explore the design variability based on the general users’ preferences. Method We explore the idea of using a cognitive technique Repertory Grid (RG) to acquire knowledge from users and experts along multiple dimensions of problem and design space. We propose REASSURE methodology which guides requirements engineers to explore the intentional and design variability in an organized way. We also provide a tool support to analyze the knowledge captured in multiple repertory grid files and detect potential conflicts in the intentional variability. Finally, we evaluate the proposed idea by performing an empirical study using smart home system domain. Results The result of our study shows that a greater number of requirements can be elicited after applying our approach. With the help of the provided tool support, it is even possible to detect a greater number of conflicts in user’s requirements than the traditional practices. Conclusion We envision RG as a technique to filter design options based on the intentional variability in various contexts. The promising results of empirical study open up new research questions: “how to elicit requirements from multiple stakeholders and reach consensus for multi-dimensional problem domain”.

Incorporating Sustainability Design in Requirements Engineering Process: A Preliminary Study

Sustainability is usually treated as an after-thought in software engineering practice. The software engineers tend to focus more on the technical dimension rather than the entire sustainability dimension. Designing software sustainability is a big challenge in current software engineering practices due to the lack of well-defined guidelines that provide tangible decomposition of sustainability aspect. Thus, we propose a framework to analyze the sustainability dimension and structure it into software requirements. Moreover, the research goal of this paper is to develop a methodology that determine sustainability requirements. The proposed meta model integrates four sustainability dimensions with the other quality attributes such as performance and usability. The contribution of this work is to help the requirements engineer to incorporate sustainability concern into software system design.