Open Access. Powered by Scholars. Published by Universities.®
- Discipline
- Institution
- Publication
- Publication Type
Articles 1 - 4 of 4
Full-Text Articles in Engineering
Convexity Applications In Single And Multi-Agent Control, Olli Nikodeemus Jansson
Convexity Applications In Single And Multi-Agent Control, Olli Nikodeemus Jansson
All Graduate Theses and Dissertations, Spring 1920 to Summer 2023
The focus of this dissertation is in the application of convexity for control problems; specifically, single-agent problems with linear or nonlinear dynamics and multi-agent problems with linear dynamics. A mixture of convex and non-convex constraints for optimal control problems is also considered. The main contributions of this dissertation include: 1) a convexification of single-agent problems with linear dynamics and annular control constraint, 2) a technique for controlling bounded nonlinear single-agent systems, and 3) a technique for solving multi-agent pursuit-evasion games with linear dynamics and convex control and state constraints. The first result shows that for annularly constrained linear systems, controllability …
Extended Kalman Filter Based Resilient Formation Tracking Control Of Multiple Unmanned Vehicles Via Game-Theoretical Reinforcement Learning, Lei Xue, Bei Ma, Jian Liu, Chaoxu Mu, Donald C. Wunsch
Extended Kalman Filter Based Resilient Formation Tracking Control Of Multiple Unmanned Vehicles Via Game-Theoretical Reinforcement Learning, Lei Xue, Bei Ma, Jian Liu, Chaoxu Mu, Donald C. Wunsch
Electrical and Computer Engineering Faculty Research & Creative Works
In This Paper, We Discuss the Resilient Formation Tracking Control Problem of Multiple Unmanned Vehicles (MUV). a Dynamic Leader-Follower Distributed Control Structure is Utilized to Optimize the Performance of the Formation Tracking. for the Follower of the MUV, the Leader is a Cooperative Unmanned Vehicle, and the Target of Formation Tracking is a Non-Cooperative Unmanned Vehicle with a Nonlinear Trajectory. Therefore, an Extended Kalman Filter (EKF) Observer is Designed to Estimate the State of the Target. Then the Leader of the MUV is Adjusted Dynamically According to the State of the Target. in Order to Describe the Interactions between the …
Cooperative Deep $Q$ -Learning Framework For Environments Providing Image Feedback, Krishnan Raghavan, Vignesh Narayanan, Sarangapani Jagannathan
Cooperative Deep $Q$ -Learning Framework For Environments Providing Image Feedback, Krishnan Raghavan, Vignesh Narayanan, Sarangapani Jagannathan
Electrical and Computer Engineering Faculty Research & Creative Works
In This Article, We Address Two Key Challenges in Deep Reinforcement Learning (DRL) Setting, Sample Inefficiency and Slow Learning, with a Dual-Neural Network (NN)-Driven Learning Approach. in the Proposed Approach, We Use Two Deep NNs with Independent Initialization to Robustly Approximate the Action-Value Function in the Presence of Image Inputs. in Particular, We Develop a Temporal Difference (TD) Error-Driven Learning (EDL) Approach, Where We Introduce a Set of Linear Transformations of the TD Error to Directly Update the Parameters of Each Layer in the Deep NN. We Demonstrate Theoretically that the Cost Minimized by the EDL Regime is an Approximation …
Cooperative Deep Q -Learning Framework For Environments Providing Image Feedback, Krishnan Raghavan, Vignesh Narayanan, Sarangapani Jagannathan
Cooperative Deep Q -Learning Framework For Environments Providing Image Feedback, Krishnan Raghavan, Vignesh Narayanan, Sarangapani Jagannathan
Publications
In this article, we address two key challenges in deep reinforcement learning (DRL) setting, sample inefficiency, and slow learning, with a dual-neural network (NN)-driven learning approach. In the proposed approach, we use two deep NNs with independent initialization to robustly approximate the action-value function in the presence of image inputs. In particular, we develop a temporal difference (TD) error-driven learning (EDL) approach, where we introduce a set of linear transformations of the TD error to directly update the parameters of each layer in the deep NN. We demonstrate theoretically that the cost minimized by the EDL regime is an approximation …