Data Analysis for Security and Privacy in Advanced Traffic Management Systems (ATMs)

Investigator(s):

• Dr. Yunpeng (Jack) Zhang, University of Houston, ORCID # 0000-0001-6208-9571  (PI)

Project Description:

Figure 1. The Framework of Data Analysis for Security and Privacy in ATMs

The project is driven by two main objectives: 1) Real-Time Intrusion Detection Algorithm (DL_TraSec): The first objective involves the creation of a cutting-edge intrusion detection algorithm that harnesses the power of deep learning. This algorithm, named DL_TraSec, is purpose-built to cater to the intricacies of ATMS. By analyzing real-time data streams, it will identify anomalies and patterns indicative of impending DDoS attacks. The aim is to proactively prevent such attacks before they disrupt the system. 2) Intrusion Detection System (ID_TraSec): Building upon the DL_TraSec algorithm, the second objective is to develop an integrated intrusion detection system, ID_TraSec. This system will not only accurately analyze the ongoing behaviors of potential DDoS attackers but also predict their actions. The ID_TraSec system will act as a comprehensive shield against a variety of DDoS flooding attacks targeting different components of the ATMS infrastructure.

The proposed system's efficacy spans multiple levels of the ATMS architecture: 1) Central Control System: The system will bolster the Central Control System, enhancing its resistance to DDoS attacks and ensuring continuous operation. 2) ATMS Sub-systems (Corridors or Areas): It will secure ATMS sub-systems by vigilant monitoring and prompt response to DDoS threats specific to these zones. 3) Intersection Advanced Transportation Controllers: The system will safeguard the critical intersection controllers, a key component of efficient traffic management. The project's methodology is dynamic and interdisciplinary, combining several elements to form a comprehensive solution. It integrates: 1) Literature Synthesis: A thorough review of existing literature on DDoS attacks, deep learning algorithms, and traffic management systems will inform the project's direction. 2) Conceptual Models: Building upon the literature, conceptual models will be developed to design and structure the DL_TraSec algorithm and the subsequent ID_TraSec system. 3) Real-Time Big Data: The collection and analysis of real-time big data from the ATMS environment will provide the foundation for refining and validating intrusion detection algorithms. 4) Algorithm Development: Novel algorithms will be crafted, leveraging deep learning techniques, to enable the accurate detection and prediction of DDoS attacks in real-time scenarios.