Access Type

Open Access Dissertation

Date of Award

January 2025

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Computer Science

First Advisor

Nathan Fisher

Abstract

This work adds novel co-design, demand characterization, and utilization bounding tools to the real-time community toolbox for the effective deployment of real-time, safety-critical cyber-physical systems. Specifically, this work exploits system dynamics for improved real-time analysis in: software-based short circuit detection, intelligent power distribution systems (IPDSes), robotic arm motion, and internal combustion engines (ICEs). In short detection, inductor sub-saturation back-EMF is exploited to link circuit size and task utilization. In the IPDS, sub-maximal loading is used to bound system utilization for variable-frequency current monitoring tasks. For the robotic arm, repeated motion and error-dependent worst-case execution time are leveraged for faster schedulability analysis. In the ICE, kinematic limitations (max speed, max acceleration) are exploited to reduce the search space for maximum demand and a corresponding fully polynomial time approximation scheme and parallelization regime for faster demand calculation. The techniques enumerate varied approaches on different cyber Physical Systems for improving analysis of real-time tasks with complex parameters.

Share

COinS