Xinkai Zhang Ph.D. defense this Friday

CSE Graduate Defenses
CSE Graduate Defenses

Ph.D. Oral Defense: Xinkai Zhang, Nimbus Lab
Friday, February 26, 2021
9:30 AM
Zoom Link: https://unl.zoom.us/my/justinbradley

Title: "End-to-end Design of Resource-Aware, Co-regulated Controllers for a Multicopter UAS"

Abstract: Modern Unmanned Aircraft Systems (UASs) are equipped with a complex suite of computational (cyber) and electromechanical (physical) systems. They are expected to provide both highly varied functionality and outstanding application performance within the available resources. System resources must be carefully allocated to achieve mission objectives. This raises new challenging questions in the context of real-time scheduling and control, especially when the computing resources are limited and/or shared by multiple tasks. Cyber-physical co-regulation is a novel resource-aware design framework proposed in [26] that augments sampled-data control systems by incorporating computing resource reallocation, sampling rate, alongside control performance in a single model.

Our research provides a complete end-to-end design of the co-regulation strategy on a multicopter UAS. Building on the co-regulation scheme, this work focuses on the detailed implementation, computational and physical controller design, system stability analysis, performance evaluation, and real-world experimental validation of a co-regulated UAS. The performances of the co-regulated UAS are compared with fixed-periodic, event-triggered, and self-triggered control strategies in various experimental scenarios. The advantages of the proposed co-regulated UAS design in control performance, computational efficiency, and robustness are analyzed and discussed based on simulation and experimental results. Our research provides a more intelligent design that makes autonomous vehicles capable of assessing and adjusting both performance and resource allocation to maximize holistic system performance -- a complete alternative strategy to traditional fixed-resource, worst-case vehicle autonomy design.

Committee Members: Justin Bradley (Chair), Carrick Detweiler, Jerry Hudgins, Qing Hui

Zoom Link: https://unl.zoom.us/my/justinbradley