Cyber-Resilient Road Networks

Connected vehicles herald much needed benefits to road transportation, including improving safety through reducing accidents, reduced traffic congestion, and smoother overall driving experience. However, internet connectivity is literally opening up new doors, that can be exploited by hackers. Nearly half of all the vehicles in the U.S. can connect to the internet. Even if you are not using in-car wifi, you might be using an entertainment system, or your car might be updating apple car play software. All of these need the internet. And even if you are not actively using these features, you are at risk!

There have been multiple demonstrations showing how easy it is to hack vehicles, and even worse: take control (or in some cases totally mess with) critical safety functions including starting and stopping a vehicle. All this is pretty scary as in the case of an actual cyber-attack on connected vehicles, lives are at risk. Alarmingly, the consequences of such a cyber-attack are largely unknown. Recent events such as the Colonial pipeline ransomware attach have shown how easy it is for cyber-attacks to cascade into the physical layer of societies in which we live in.

Using a combination of complex adaptive systems and multiple data sets, we ask: what is the potential for a targeted hack on transportation? In answering this question, we develop network based metrics and a framework for detecting cyber attacks on transportation through multiple data sources. Our work has been highlighted by scientific journals and news outlets.

Strategic Cyber Effects In Air Transportation Networks

Recent IT outages and cyber-attacks on air transportation networks illustrate just how vulnerable U.S. air passenger networks are to cyber-attacks. This work is in collaboration with Dr. Charles Harry, in the school of public policy at the University of Maryland. We develop methods to quantify the impacts of cyber-attacks on passenger transportation, specifically modeling the loss of capacity and propagation of delays. We do this through detailed air flight data from the Bureau of Transportation Statistics, and agent based modeling applied to air transportation networks. The video above visualizes cascading flight delays in a hypothetical cyber-attack scenario that disrupts the ATL Hartsfield-Jackson airport for the duration of one hour.

We find that targeted attacks on vendors that perform seemingly non-essential functions such as weight checks, for multiple airlines, are nontheless particularly concerning weak points of air transport networks, reminiscent of recent vendor attacks in other critical infrastructure systems, such as the SolarWinds breach. Our work was recently published at the International Conference on Cyber Conflict, Cycon 2021.

  • Strategic Cyber Effects in Complex Systems: Understanding the US Air Transportation Sector, Published: Cycon 2021.

Learning From Disasters Using Data

Texas electrical power reliance | ERCOT

Winston Churchill famously said after the second world war - “Never let a good crisis go to waste.” We are using multidimensional data sources to learn lessons from recent critical infrastructure disruptions, in order to mitigate future consequences and build societal resilience. One such disaster happened in February 2021.

The day after Valentine’s day, Texas was less than 5 minutes from a black start — an indefinite blackout that could have lasted months. Thankfully this did not happen, but the 2021 Texas power crisis was a disaster, resulting in 70 people dead, and almost $200 billion in damages, making it the costliest disaster in Texas history. So what really happened?

In the case of the Texas 2021 winter storm, coupled societal infrastructures led to a power crisis becoming a water crisis, and a food crisis later in the week. Impacts from other recent disruptions such as hurricanes and wildfires illustrate how little we understand about the disparate effects from disasters. In collaboration with Dr. Ryan Qi Wang in the department of civil and environmental engineering at Northeastern University, and Dr. Marta Gonzalez in the department of city and regional planning at UC Berkeley, we are gathering multidimensional data sources, and combining cross-disciplinary approaches to understand the chain of events that led to these intercoupled disruptions.

PhD Work

During my PhD under Dr. Eric Weeks at Emory University, I studied the statistical physics of two-dimensional (2D) amorphous materials. 2D materials are very different from their 3D counterparts. Take graphene for example. Phase transitions like crystallization are also known to be very different in 2D and 3D. In fact the recipients of the 2016 Nobel prize in physics (Kosterlitz and Thouless) did pioneering work in 2D phase transitions.​

In our work we looked at the affect of dimensionality on the tranistion from liquid to amorphous materials, also known as the glass transition. In of itself, the transition from liquid to amorphous solid is something of a mystery. In fact, nobel prize winner Philip W. Anderson has famously stated that the glass transition problem is probably the deepest and most interesting unsolved problem in condensed matter physics. Until recently, there has been a loose consensus that the glass transition is similar in 2D and 3D.

Surprisingly, we found the same differences between 2D and 3D crystallization also appear between 2D and 3D glass transitions. Through experiments and the development of novel image analysis algorithms, we observed that long-wavelength fluctuations similar to those that cause differences in 2D and 3D crystals, also affect dynamics in 2D glasses. However, unlike in crystallization, we did not observe a fundamental distinction between 2D and 3D glass transitions. Below, you can find some of our representative publications, as well as popular news articles:

Subscribe to my newsletter.