Controls and Control Theory Commons^™

Power System Dynamic Control And Performance Improvement Based On Reinforcement Learning, Wei Gao

Electronic Theses and Dissertations

This dissertation investigates the feasibility and effectiveness of using Reinforcement Learning (RL) techniques for power system dynamic control, particularly voltage and frequency control. The conventional control strategies used in power systems are complex and time-consuming due to the complicated high-order nonlinearities of the system. RL, which is a type of neural network-based technique, has shown promise in solving these complex problems by fitting any nonlinear system with the proper network structure.

The proposed RL algorithm, called Guided Surrogate Gradient-based Evolution Strategy (GSES) determines the weights of the policy (which generates the action for our control reference signal) without back-propagation process …

Radar Detection, Tracking And Identification For Uav Sense And Avoid Applications, Erik George Moore

Electronic Theses and Dissertations

Advances in Unmanned Aerial Vehicle (UAV) technology have enabled wider access for the general public leading to more stringent flight regulations, such as the "line of sight" restriction, for hobbyists and commercial applications. Improving sensor technology for Sense And Avoid (SAA) systems is currently a major research area in the unmanned vehicle community. This thesis overviews efforts made to advance intelligent algorithms used to detect, track, and identify commercial UAV targets by enabling rapid prototyping of novel radar techniques such as micro-Doppler radar target identification or cognitive radar. To enable empirical radar signal processing evaluations, an S-Band and X-Band frequency …

Robust Distributed Stabilization Of Interconnected Multiagent Systems, Vahid Rezaei

Electronic Theses and Dissertations

Many large-scale systems can be modeled as groups of individual dynamics, e.g., multi-vehicle systems, as well as interconnected multiagent systems, power systems and biological networks as a few examples. Due to the high-dimension and complexity in configuration of these infrastructures, only a few internal variables of each agent might be measurable and the exact knowledge of the model might be unavailable for the control design purpose. The collective objectives may range from consensus to decoupling, stabilization, reference tracking, and global performance guarantees. Depending on the objectives, the designer may choose agent-level low-dimension or multiagent system-level high-dimension approaches to develop distributed …