Date of Award

5-2017

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Engineering and Sciences

First Advisor

Marco Carvalho

Second Advisor

Thomas Eskridge

Third Advisor

Adrian Peter

Abstract

While looking for a high-level adaptive traffic generation tool, we came to realize that no such tool exists that can be used for rapid development while being platform agnostic. Having reviewed a wide array of tools to either implement user models or simulate traffic, we were unable to find a tool with the right capabilities while maintaining complexity, portability and extensibility. To overcome these issues, we introduce a new adaptive user-modelling framework for the specific use case of cyber activity emulation. Our framework supports the creation of high-level user models that can react to changes in their environments and vary the way they emulate cyber activity based on those changes. We review the problems with the current tools and show how our behaviour tree based solution can be used to achieve our goals in an illustrative scenario showcasing the framework’s adaptability – a key feature most other tools are lacking. Furthermore, we show that our framework is also extensible, portable, and more conducive to rapid development than other user modelling tools currently available.

Comments

Copyright held by author

Download

Share

COinS