Novel Hybrid Scheduling Technique for Sensor Nodes with Mixed Criticality Tasks

Area of Science:

• Computer Science
• Embedded Systems
• Real-Time Systems

Background:

• Sensor networks are crucial for complex control applications.
• Safety- and time-critical domains require predictable, jitter-less task execution on sensor nodes.
• Existing scheduling mechanisms may not fully address the unique demands of sensor nodes.

Purpose of the Study:

• To propose an efficient scheduling solution for sensor nodes that ensures high execution predictability.
• To maintain a high Central Processing Unit (CPU) utilization factor within sensor nodes.
• To address the need for adapted task scheduling in safety- and time-critical sensor network applications.

Main Methods:

• Introduced Hybrid Hard Real-Time Scheduling (H²RTS), combining static, clock-driven, and dynamic, event-driven techniques.
• Developed detailed, integrated schedulability analysis for H²RTS.
• Introduced sufficiency tests based on processor demand and linear upper bound metrics.

Main Results:

• H²RTS provides high execution predictability.
• H²RTS achieves a high node CPU utilization factor.
• The scheduling technique was validated on a simulator and an ARM7 microcontroller-based sensor mote.

Conclusions:

• The proposed H²RTS is an efficient and effective solution for real-time task scheduling in sensor networks.
• The hybrid approach balances predictability with resource utilization.
• The validated performance confirms the suitability of H²RTS for critical sensor network applications.