Algorithmic approach leveraging a real-time task scheduler with fan-out strategy

Abstract

This study focuses on the communication challenges among processes in mobile drone systems, specifically addressing the dynamics and decentralization of their topology. An algorithmic approach for real-time systems is proposed, emphasizing its application in drones. The Fan-shaped Real-Time Task Scheduling Algorithm (APTTRA) serves as the cornerstone, distributing processes with deadline constraints in a fan-shaped manner to ensure timely completion. It introduces a metric that evaluates not only task compliance but also when and how long, providing a comprehensive insight into the system's effectiveness. To support performance evaluation, the use of a connected acyclic graph is proposed, offering a detailed understanding of performance across various process sections. The system's adaptability is highlighted through the incorporation of variables in real-time applications, providing a complete view in dynamic situations. Along with the use of Minix as a modular operating system, allows for testing APTTRA before implementation in real drones. The importance of real-time task scheduling for drones, especially helicopters and quadcopters, is emphasized, underscoring the need to tailor control algorithms. The evaluation focuses on implementation, successes in real flights, and the application of APTTRA in a genetic algorithm for calibration within the planning ranges.