糖心原创

Ph.D. Committee:聽 Drs. Junjie Zhang (advisor), Michelle Cheatham, Bin Wang, and Phu H. Phung (University of Dayton)

ABSTRACT:

Security聽has聽become of the聽most significant concerns聽for聽our cyberspace.聽聽Securing聽the cyberspace, however, becomes increasingly challenging.聽 This can be attributed to the聽rapidly growing diversities and complexity of the modern cyberspace.聽 Specifically, it is聽not聽any聽more聽dominated by connected personal聽computers (PCs);聽instead, it is greatly聽characterized by cyber-physical systems (CPS), embedded systems, dynamic services, and聽human-computer interactions [28]. Securing modern cyberspace therefore calls for a multi-faceted approach capable of systematically integrating these emerging characteristics. This聽dissertation presents our novel and significant solutions towards this direction.聽聽 Specifically, we have devised automated, systematic security solutions to three critical aspects of聽our modern cyberspace including i) cyber-physical systems, ii) dynamic web services, and聽iii) social networks.聽 This dissertation makes the following contributions.聽 First, we have聽conducted systematic vulnerability assessment for a real-world, complex CPS, namely Unmanned Systems Autonomy Services (UxAS). Our assessment has identified a set of exploitable vulnerabilities. Second, we have designed an adaptive traffic morphing algorithm聽to conceal CPS communications into background network traffic. Third, we have designed聽a CPS self-destruct model and studied the security-and-performance trade-off using probabilistic model checking.聽聽 Finally, we have designed a novel detection system to detect聽suspicious behaviors in an online emotional support system.