Abstract
Abstract
This paper presents an experimental comparison of two widely adopted real-time scheduling algorithms—Rate Monotonic Scheduling (RMS) and Earliest Deadline First (EDF)—implemented on an ARM Cortex-M4 microcontroller (STM32F407) using FreeRTOS v10. Unlike prior studies that rely predominantly on simulation or analytical models, this study conducts empirical evaluation on physically embedded hardware under realistic periodic workloads. A set of five periodic tasks with varying periods and execution times were designed and deployed under both scheduling policies. Key performance metrics, including deadline miss ratio, CPU utilization, and task response time, were measured across 500 task instances per policy. Experimental results demonstrate that EDF achieves zero deadline misses and higher CPU utilization (92%) compared to RMS (2% deadline miss ratio; 85% CPU utilization), consistent with classical schedulability theory. However, RMS exhibits more predictable and deterministic response times, which may be preferable in safety-critical contexts. The findings provide practical guidance for embedded system designers selecting between fixed-priority and dynamic-priority scheduling strategies, and highlight the importance of hardware-based evaluation in validating theoretical scheduling models.
Keywords: Real-Time Scheduling, Rate Monotonic Scheduling, Earliest Deadline First, FreeRTOS, ARM Cortex-M, Embedded Systems, CPU Utilization, Deadline Miss Ratio, Task Response Time

National Library of Nigeria
Association of Nigerian Authors
Nigerian Library Association
EagleScan
Crossref