Research Interests

My current research centers around systems and controls with application to transportation cyber-physical-human systems. We aim to leverage emerging technologies in sensing, communication, and vehicle automation to build safe, sustainable and resilient cities of the future. I am particularly interested in the following open challenges:

Theory: 

Applications:

Overall, I hope to enlighten the creation of emerging technologies that promote the well-being of everyone and the planet in general. Some of my recent works are briefly summarized below.

Smoothing traffic via control of automated vehicles 

Stop-and-go waves are easily caused by unstable traffic due to the collective behavior of human drivers, resulting in higher vehicle fuel consumption and emissions. We aim to smooth and stabilize unstable traffic flow via intelligent control of automated vehicles (AVs). Using Pontryagin's minimum principle we have designed optimal additive AV controllers for smoothing nonlinear mixed traffic flow. Moreover, leveraging Barbalat's Lemma we have synthesized a class of provably safe AV controllers effective for smoothing and stabilizing unstable traffic, with demonstrations on car-following models for commercially available adaptive cruise control (ACC) vehicles. Speed perturbations of the unstable traffic are significantly reduced due to AVs employed with the synthesized controllers, resulting in lower fuel consumption and emissions. One of the distinctive features of the AV controllers synthesized is the elimination of the requirement on vehicle connectivity commonly seen in the literature. In other words, only local traffic information like spacing and relative speed to the preceding vehicle is required for controller synthesis, which makes the additive controllers readily implementable for ACC vehicles due to onborad radar sensors monitoring the road ahead.


Related publications

Transportation and the environment

With an increased level of connectivity and automation, connected and automated vehicles (CAVs) are expected to be able to proactively adjust their driving strategies subject to constraints imposed by the predicted future traffic, resulting in potential benefits such as improved energy efficiency, enhanced traffic safety, among others. Aimed at achieving fuel benefits for CAVs, we design optimal CAV control laws with co-optimization of vehicle speed and gear position leveraging real-time traffic prediction. The traffic prediction is conducted using an unscented Kalman filter in a rolling horizon fashion based on a modified Payne-Whitham (PW) model capable of handling mixed-autonomy traffic. Following real-world speed profiles collected on TH-55 in Minnesota, it is shown that energy benefits achieved by a 10 vehicle platoon range from 2% to 16%, with a 1%~5% reduction in travel time observed for legacy vehicles (LVs) behind CAVs at different penetration rates of CAVs. In addition, it is observed that CAVs using the proposed eco-driving approach could have a positive impact on the LVs behind in terms of energy consumption, regardless of the driving styles of the LVs ahead.

Related publications

Figure: Javed, Hamida, & Znaidi (2016)


Transportation safety and cybersecurity

Transportation cyber-physical-human systems (T-CPHS) are enabled by the increased feedback-based interactions among sensing, computation, and transportation, including research on AVs and how they will influence traffic flow. We are primarily interested in vehicle-based T-CPHS and vehicle-infrastructure coordinated T-CPHS, where safety and security are key elements. Any failure in T-CPHS, such as malfunctions of vehicle communication, could result in a staggering financial loss and even catastrophic loss of human lives. To address the fundamental limitations of current technologies for developing T-CPHS can not only help maintain and improve the economic competitiveness of a nation, but also contribute to protecting and extending human life. 

For physical systems, we design provably safe control laws for AVs with real-time synthesis using control barrier functions (CBFs). Current approaches to ensuring safety for T-CPHS rely on large-scale simulations and field testing, suffering from two fundamental challenges: cost and coverage. We work on leveraging CBFs for synthesizing provably safe AV controllers to fundamentally transform the conventional trial-and-error paradigm and improve safety in T-CPHS, while providing useful coverage at an acceptable cost. For cyber systems, we draw on concepts from differential games to model and characterize the interactive dynamics between T-CPHS and malicious attacks, where robust control, such as the min-max control technique, is employed for designing appropriate operational strategies for T-CPHS. This could result in substantial reduction in the cost of exhaustive simulations and field testing.

Related publications

Traffic operations and control

Transportation systems are inherently complex and full of stochasticity. We are interested in modeling and characterizing the stochastic arrival of vehicles and passengers at intersections and bus stops respectively to support robust and intelligent traffic operations. One of the distinctive features of our studies is the incorporating of non-homogeneous Poisson processes that mathematically characterize the aforementioned arrival process. Consequently, we apply dynamic programming to determine the optimal signal timing at intersections for minimizing traffic delays and maximizing throughput. Moreover, we design programs for optimally allocating a limited number of buses to transit routes with the objective of minimizing passenger waiting times, where the solution is obtained using nonlinear integer programming. 

We are also interested in extending ramp metering control to mixed autonomy traffic flow with varying degrees of automation. Specifically, we formulate an analytical fundamental diagram for mixed autonomy traffic, which is dependent on the market penetration rates of automated vehicles. Further, we model and simulate the composite traffic flow with a mixed autonomy macroscopic traffic flow model, and modify a standard ramp metering control strategy to optimize operations under the new flow conditions. 

Related publications

Dynamical systems and control theory

We study the fundamental challenges and open questions in nonlinear systems and control theory, such as feedback control, relaxed control, impulsive control, among others. Moreover, we have studied extensively on measure-driven control systems, extending the results of dynamical systems driven by regular controls (measurable functions).

We have applied systems and control theory to a wide range of applications, such as intelligent transportation systems, building maintenance units stabilization, predator-prey systems, among others. For example, we design optimal feedback incentive programs for accelerating the adoption of AVs into the auto market, where desired market penetrations can be achieved in a prespecified planning horizon given sufficient resources. We also work to further the control of mixed-autonomy traffic from local to city scale, using network modeling approaches as well as systems and control theory. Specifically, we develop mathematical frameworks to model the hybrid dynamics of Mobility-as-a-Service systems using impulsive stochastic differential equations. Based on such framework capturing the uncertain demand, impulsive control theory can be applied for determining the optimal AV dispatch policy to reduce customer waiting times. 

Related publications