DSpace Collection:
http://hdl.handle.net/10203/25357
2024-03-03T06:33:55ZConceptual framework and extensible modeling method supporting ontology-based scenario specification
http://hdl.handle.net/10203/309241
Title: Conceptual framework and extensible modeling method supporting ontology-based scenario specification
Authors: Baek, Young-Min
Abstract: In recent decades, scenario-based techniques (scenario methods) for software/systems engineering have been actively employed to resolve intricate engineering problems in complex systems. However, despite the widespread use of scenarios, the lack of a well-established reference framework that systematically organizes key concepts and attributes of scenarios has made the application of such scenario methods more challenging. As a result, engineers who need to utilize scenarios and scenario methods must be assisted by more than systematic guidance at the method level. To address these problems, this study focuses on providing reference framework and modeling method for the development of scenario methods. In order to solidly build a conceptual basis, a literature review is conducted to analyze scenarios and scenario methods used in various engineering fields. Based on the data collected through the survey, a Conceptual Scenario Framework (CSF) is proposed, which defines core components of scenario methods and models. In addition, we develop an Extensible Scenario Modeling Method (ESMM) that supports engineers to perform scenario modeling and domain-specific extension based on the framework’s components. ESMM provides an Extensible Scenario Modeling Language (ESML) consisting of model types and classes that enable both scenario description and ontological analysis. ESML also supports language-level extensions that allow scenario engineers to design domain-specific scenario elements flexibly. This study evaluates the modeling method with existing scenario development methods, suggested in the automated driving system domain. It was confirmed that the language constructs of ESML have semantic expressiveness to serve as a reference framework by analyzing whether they can effectively express the scenario data and characteristics. In addition, the case study results validate that ESMM has the extensibility to be extended and specialized for developing a domain-specific scenario modeling language and adequately supports the ontological analysis of specific application domains.
Description: 학위논문(박사) - 한국과학기술원 : 전산학부, 2023.2,[v, 102 p. :]2023-01-01T00:00:00ZContext mining-based fault analysis of collaboration failures in cyber-physical system-of-systems
http://hdl.handle.net/10203/309244
Title: Context mining-based fault analysis of collaboration failures in cyber-physical system-of-systems
Authors: Hyun, Sangwon
Abstract: and (4) an absence of an end-to-end solution from pattern analysis to the fault identification. To overcome these limitations, we define a context model for CPSoS logs and propose an FII pattern mining algorithm covering the main features of the sequential analysis, an overlapping clustering technique for multiple pattern mining, and a pattern-based fault localization method. In experiments conducted on several CPSoS examples, we found that the proposed approach achieved the highest context mining accuracy and clustering precision. We also checked that the proposed localization method presented the highest fault localization efficacy. We newly detected undiscovered failure scenarios and bugs in this study. The findings of this study can facilitate the accurate analysis of collaboration failures.; A cyber-physical system-of-systems (CPSoS) tries to achieve prominent goals, such as increasing road capacity in platooning that groups driving vehicles in proximity, through interactions between constituent systems (CSs). However, during the collaboration of CSs, unintended interference in interactions causes collaboration failures that may lead to catastrophic damage, particularly for the safety-critical CPSoS. It is necessary to analyze the failure-inducing interactions (FII) during the collaboration and resolve the root causes of failures. Existing studies have utilized pattern-mining techniques to analyze system failures from logs. However, they have four limitations when applied to collaboration failures: (1) limited data model to handle discrete and continuous logs generated from CPSoS; (2) limited coverage of main features required to sequentially analyze the logs; (3) limitations on identifying multiple failure patterns in a single log
Description: 학위논문(박사) - 한국과학기술원 : 전산학부, 2023.2,[iv, 86 p. :]2023-01-01T00:00:00ZInfluence-directed policy generation using reinforcement learning for collaborations in system of systems
http://hdl.handle.net/10203/309242
Title: Influence-directed policy generation using reinforcement learning for collaborations in system of systems
Authors: Belay, Zelalem Mihret
Abstract: A system of systems (SoS) tries to utilize constituent systems' (CSs) capabilities to achieve its goals. This activity is quite challenging because CSs are themselves systems that are autonomous and decide when and how to activate their capabilities on their own account. For example, in the case of mass casualty incidents, response systems such as fire fighters, emergency vehicles, rescuer agents, \textit{etc.} have to commit their capabilities and collaborate to achieve the response goals. For such collaborations, it is not only synchronizing CSs' actions and decisions, but requires influencing the CSs' decision making behavior in order to achieve the common goals. In this research, we present the design and development of an influence-directed policy generation approach that aims on generating collaboration policies. The collaboration policies are generated by leveraging decision pattern analysis and reinforcement learning techniques.
The outcomes of a collaboration in a SoS can be significantly influenced by the CSs' autonomy and belongingness. The notion of autonomy describes CSs' managerial and operational independence to make choice about when and how their capabilities should be activated, while belongingness refers to the degree of relatedness (or overlap) between the SoS purpose and the CSs goals. As desirable as belongingness can be in achieving the SoS purpose, autonomy can be a limiting factor. We aim to utilize CSs' capabilities to perform the SoS specific tasks while maintaining the CSs' autonomy.
This research contributes to resolving two major collaboration concerns in SoS: how to enable SoS designers to compose CSs' capabilities, and CSs developers to design their systems to make efficient collaboration in a SoS context. To a certain extent, the proposed approach addresses guided emergency problems which is rated as one of the highly valued research endeavors in the SoS engineering domain.
Description: 학위논문(박사) - 한국과학기술원 : 전산학부, 2023.2,[iv, 77 p. :]2023-01-01T00:00:00ZData-driven environment model generation using imitation learning for efficient cyber-physical system goal verification
http://hdl.handle.net/10203/309243
Title: Data-driven environment model generation using imitation learning for efficient cyber-physical system goal verification
Authors: Shin, Yong-Jun
Abstract: Cyber-Physical Systems (CPS) continuously interact with their physical environment through software controllers that observe the environment and determine actions. Engineers can verify to what extent the software controller under analysis can achieve given goals by analyzing its Field Operational Test (FOT) logs. However, repeating many FOTs to obtain statistically significant results is expensive in practice. Simulation-based verification is an efficient alternative for reducing the FOT cost for CPS goal verification. However, it requires an accurate virtual environment model that can replace the real environment interacting with the CPS, and it is challenging to craft the environment model manually.
This dissertation proposes a novel data-driven approach that automatically generates the virtual environment model from a small amount of FOT logs. It generates an environment model that mimics the behavior of the real environment using Imitation Learning (IL). Specifically, this dissertation provides 1) a systematic and comprehensive survey on environment modeling, 2) a formal framework of CPS goal verification and a formal problem definition of environment model generation, 3) a data-driven environment model generation approach using IL, and 4) an empirical evaluation based on case studies of an autonomous driving system goal verification and reusable datasets. The evaluation results show that the approach can generate accurate virtual environment models for CPS goal verification with small FOT log data. Therefore, CPS software engineers can automatically obtain accurate virtual environment models and efficiently verify the controller based on the simulation.
Description: 학위논문(박사) - 한국과학기술원 : 전산학부, 2023.2,[vi, 100 p. :]2023-01-01T00:00:00Z