Scheduling Techniques for Real-Time Systemsfor Defense Applications on Multicore Processors

The real-time scheduling of tasks of any application software is an important requirement in defense applications. This paper explains partitioning and scheduling techniques in the real-time implementation of any defense system on embedded multicore processors with Linux as the operating system. A typical sonar system is considered as the reference to elaborate partitioning and scheduling techniques. This can be extended to any defense applications like radars, missiles, etc. This paper describes the Partitioned Earliest Deadline First (PART-EDF) scheduling algorithm integrated into LITMUSRT real-time Linux kernel to enhance task schedulability. The proposed system partitions computationally intensive, periodic operations like filtering, beam forming, target classification, etc., and sporadic operations like input and control functions onto different clusters on the multi-core processors.