Dr. Dallas H Snider

This paper presents an on-going research on machine learning (ML) systems for connected vehicle security. It proposes the application of various ML techniques that are applied to Basic Safety Message (BSM) test datasets, both on normal operation and anomalous behavior. The BSM test datasets conform with the SAE J2735 Standard on message sets that support vehicle-to-everything (V2X) communications systems. The purpose of the study is to determine the suitability of ML systems in identifying and classifying normal and anomalous BSM messages in a network of connected vehicles and the V2X systems.

This paper presents a subcomponent of an on-going research on connected vehicle security. It proposes techniques on the generation of synthetic Basic Safety Message (BSM) test datasets, both on normal operation and anomalous behavior. The synthetic test data conform with the SAE J2735 Standard on message sets that support vehicle-to-everything (V2X) communications systems. The purpose of such datasets is for the derivation of machine learning systems that can be deployed in a V2X operating environment.

Online analytical processing databases allow for the efficient analysis of vast amounts of data. In this paper, we describe the design of an online analytical processing data cube structure for use in the analysis of multiple measures of environmental water quality data. The measures, also known as facts, will be the quantitative values returned by various water quality tests while the dimensions will be the attributes that describe the what, when and where of the water quality test measures. The data model and build process presented here will allow for varying types of tests and for the insertion of new data as it becomes available. The overall goal of this design is to provide business intelligence capabilities to water quality decision makers.

In this paper, transient stability assessment is performed on a power system using a classification approach and data mining algorithms. As a first step, offline training data was collected by conducting load flow studies under normal operating conditions and faulty operating conditions at buses, at three different locations at lines and at different load levels. Twenty-three features were chosen to represent the training data for each load flow simulation. A support vector machine model was built and trained using the training data as well as a Naive Bayes model and Decision Tree model. Then an online testing model was developed and real-time data was used to test the validity of the model developed. The results indicate a higher accuracy and less time consumed by the core vector machine model compared to previous models available in literature. The IEEE 14 bus system was used for training data and for verifying the speed and accuracy of the proposed data mining algorithm.

Effective data visualization theory and technique as applied to network security data can provide efficient and meaningful insight to massive, obtuse sets of data otherwise imperceptible to network security analysts. Given this potential impact, interest in understanding insightful data visualization in cybersecurity continues to gain momentum and draw researchers to this domain. In this paper, the authors present an overview of research in visual data analysis as applied to cybersecurity, including a discussion of challenges related specifically to visualization in network security, a synthesis of visualization techniques, and a brief conversation and critique of specific visualizations utilized in modern tools. In conclusion, the authors present a brief and open-ended proposal for further development of a flexible data visualization project to integrate these findings into a product to illustrate the power of effective data visualization to network security data.

The traditional view of facility management (FM) previously included tasks which were more of a physical nature in terms of the responsibility of the building. Newer FM roles now include maintaining critical operations through the use of information technology (IT) and automated controls. Therefore, these changing operations now include new concerns for facilities personnel, who today typically must work in conjunction with subcontractors and IT staff for installation as well as operations. Facilities personnel are now also becoming aware of the needs to secure university (or any campus setting) administrative and facilities data. This awareness has created an additional layer of concern and responsibility for facilities management personnel. Although publications often address cyber-threats to campus operations, rarely have these topics included what the specific concerns might be for facility managers regarding the potential threats. Additionally, rarely has the discussion addressed the role of the stakeholders with regards to the mitigation of those concerns.

In this paper, we describe a process that has been developed to transfer network intrusion data captured by Fail2ban to an adaptive enterprise intrusion detection and prevention system. The process involves software agents that we have created that are interconnected to a central behavior analysis database service where each software agent records attack meta-information collected during previous intrusion attempts. These distributed agents are the first phase of an overall plan to create a smarter network defense system through the collection and analysis of network signatures generated by real security threats. The central database to which the agents report warehouses and analyzes the meta-information collected by the interconnected agents. The agents can then utilize both instantaneous and historical data by integrating rules derived from the data collection and analysis process into intrusion prevention policies. The final result will be a modular and scalable network defense system that should be more responsive and adaptable to imminent threats.

Software defined networks and network function virtualization are providing much needed agility to network and system administrators attempting to meet ever changing demands from their stakeholders. However, there exists a need for tools to assist in the configuration, deployment, testing and knowledge transfer of these networks and their components. In this paper, we present our plans to develop a set of open source tools to assist with the aforementioned tasks. The goal is to apply existing disparate pieces of technology so they may work together to assist with the management of virtual network environments.

Commercial off the shelf biosensors have become ubiquitous in the fields of exercise physiology and have been utilized in a variety of extreme environments. Recently, two COTS biosensors have been validated with regards to their accelerometry compared to the F/A-18 jet aircraft. Furthermore, they have shown promise in monitoring inflight physiologic data during high performance maneuvers. This paper describes the knowledge discovery processes implemented during this study to evaluate the performance of wearable sensors in F/A-18s during practice flights conducted by the U.S. Navy's Flight Demonstration Squadron, The Blue Angels. Challenges in data integration, data transformation and noise handling will be presented along with future plans to expand the study to other aircraft.