Self-learning capabilities for a mission oriented data quality and security assurance in military IoT systems
Abstract
The fast growth of IoT systems and their integration into sensor and communication networks has resulted in a significant growth in an amount of data collected, communicated, processed and used in the military operations. Also, this development has moved the point of possible data use away from data source locations, thus enhancing the importance of cybersecurity and safety factors consideration in military data collection design. As IoT devices have very limited resources commonly, many security mechanisms cannot be enforced on a device level. The lack of resources along with the strong possibility of malicious attacks against both cyber and physical components in the battlefield environment may result in a significant decrease of the quality of data delivered by IoT systems and networks. The project developed by the Rochester Institute of Technology (PI: Dr. Leon Reznik) aims at enhancing cybersecurity on the system level by developing and implementing the integrated procedures for data quality evaluation and assurance with the strong focus on cybersecurity aspects. Our system approach, which assumes that cybersecurity will not be isolated but evaluated as an important influencing factor on overall data quality, will enable integrating cybersecurity consideration into a military data collection and communication system design. We shall produce novel design procedures, which support a higher level of IoT system flexibility and security along with possible architectural polymorphism and optimization against multiple criteria such as an overall performance in a particular mission accomplishment as well as data quality and security (DQS) and quality of service (QoS). The goal of this project is to include intelligent self-learning capabilities into a proof-of-the-concept design prototype of the IoT data collection and security evaluation framework, that would enable structural and procedural autonomic re-adjustments of the data sources, data streams communication and fusion in order to assure the required DQS for a particular mission at the data point of use. Unlike conventional IoT data collection systems, the developed procedures will deliver data quality and security level estimates that would significantly simplify further data analysis on the battlefield. To make the procedures operational in real time and in a dynamic environment we will develop and employ artificial intelligence and machine learning techniques. We will target the framework prototype implementation on mobile devices. In order to achieve this goal, we need to accomplish the following tasks: Task 1. Further develop and advance a comprehensive DQS evaluation methodology integrating various metrics with a focus on cybersecurity aspects by developing new and adopting existing common data structures and data formats for DQS evaluation and calculus in IoT applications domain. For a prototype implementation in this Task, we will target Android mobile devices, in particular smartphones as they have multiple embedded sensors. Task 2. Develop and introduce self-learning capabilities based on artificial intelligence and machine learning techniques employment in the data collection and security evaluation procedures aiming at assuring a specified DQS level at the point of data use and enabling the fast system re-adjustment to a particular mission in a highly dynamic environment. In our research and development, we will employ methods and data structures based on graph search and genetic algorithms.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 22, 2020
- Source ID
- W911NF2010337
Entities
People
- Leon Reznik
Organizations
- Army Contracting Command
- Rochester Institute of Technology
- United States Army