Criticality, complexity, and expansiveness make the control and management of civil infrastructure systems a particularly challenging problem. A plethora of outdated techniques are used to control much of the legacy infrastructure that is unreliable and wastes energy. Upgrading these systems with advanced controllers which take advantage of ubiquitous computing and networking may present a win-win scenario with decreased costs and increases sustainability. My interests aim to confront this challenge with the application of networked control theory, embedded computing, and wireless communications.
Systems control theory, primarily developed in the fields of aerospace, electrical, and mechanical engineering, is a vital, yet oft overlooked discipline of infrastructure engineering. This intersection of civil engineering with control theory covers a breadth of challenging applications and philosophies. Systems such as heating ventilation and air-conditioning (HVAC) systems, feedback structural control, fluid distribution networks, and the electric grid all fall under the umbrella of civil engineered systems which may benefit from advancements in control, communication, and sciences.
Common amongst many of these systems are nonlinear dynamics, complex state and control constraints, and distribution of sensing, actuation, and computation. My research has brought together recent theoretical developments in bilinear dynamical systems (BDS), model predictive control (MPC), and agent-based control (ABC). Bilinear systems have a state that evolves as a linear function of the current state, plus a multiplicative term with respect to the control input and state. Such systems are linear under constant input, but nonlinear when control dynamics are considered. Additional complexity in civil systems can be found in force-limited actuators, maximum power constraints, and safe operating limits. Model predictive control is a well-developed control theory particularly well suited for such control systems with constraints and model disturbances.
Interdisciplinary research is fundamental to solving today’s engineering problems. To this end, my Ph.D. research aimed to increase the reliability and reduce energy consumption of bilinear civil systems through the development of a distributed model predictive control (DMPC) system constrained by limited wireless communication bandwidth. I lead an interdisciplinary intercollegiate team through the design of a wireless control platform capable of solving complex online open-loop control optimization problems required for DMPC. The civil application areas for these new devices all exhibit a large number of measurements and points of actuation which may overwhelm a centralized control device, but not the decentralized computation and collaboration inherent in a network of these wireless decision makers. With the vision of improving infrastructure system reliability and efficiency, my research made progress towards unifying bilinear MPC with the design of wireless control hardware and thorough validation on test beds.