(147i) Advances in Next Generation Cyber-Physical Systems: Cybersecurity, Quantum Computing, and Smart Material Design

The advent of IIoT and Industry 4.0 has led to a revolution that favors a cyber-physical setup involving greater integration of processes with computers to improve transparency into the processes and increase process efficiency. However, this integration introduces obstacles such as increased system complexity and increased vulnerability to cyberattacks with devastating consequences ranging from reduced profitability to loss of life as control actions are affected by faulty data and no longer guarantee process safety.

In response to these concerns, my doctoral thesis is focused on the development of cyber-secure control strategies with theoretical guarantees of safety and stability while accounting for practical challenges such as handling process disturbances, noise and changing process dynamics. The thesis addresses the control of process systems with nonlinear dynamics, which represent the majority of real-world processes, using a modified control algorithm based on an optimization-based model predictive control algorithm to also detect for cyberattacks. The research conducted thus far analyzes multiple detection strategies such as comparing state predictions/estimates to measurements or applying random updates to the control algorithm to detect cyberattacks on control elements such as process sensors and actuators. Initial methods have seen limited success in terms of the duration for which theoretical guarantees hold, especially when the dynamics of the nonlinear process change with time. To address this, a random string of binary numbers that alter the control input law applied at any sampling time during process operation is utilized to make the prediction of control actions challenging, consequently making it more difficult to develop stealthy cyberattacks. This formulation gives us an opportunity to develop theoretical guarantees of safety until a cyberattack is detected.

The various cyberattack detection strategies are computationally expensive and motivated the consideration of quantum computing to implement control algorithms. The two major obstacles encountered with quantum computations include nondeterminism introduced by quantum noise inherent to quantum devices and the determination of a control algorithm that performs better than its classical counterparts. The first obstacle has been the focus of our research thus far where a detailed analysis was executed using simple control algorithms to study the impacts of quantum noise in the determination of stabilizing control inputs in various scenarios. This was achieved using both a quantum simulator provided by IBM's quantum experience, Qiskit. The long-term goal of exploring this topic is to build a foundation to identify scenarios where quantum computing can be taken advantage of for applications in the field of process engineering and control. Potential applications of quantum computing in the future include the simulation of smart materials that respond to external stimuli. In order to achieve this, an analysis into the fields of molecular modeling and behavior analysis such as Density Functional Theory (DFT) and Molecular Dynamics (MD) has been initiated with the objective of integrating background knowledge of process systems engineering to develop an optimization-based material design to control the behavior of such stimuli-responsive materials.

Research Interests:

The topics that I am interested in working on include:

1. Dynamic system modeling and advanced process control algorithm development
2. Development of machine learning and AI algorithms (Currently working on a project to develop some background in this field)
3. Cybersecurity
4. Quantum Computing
5. Optimization-based materials design