부산대학교 도서관

With the continuous development of underwater acoustic signal processing technology, the demand for data volume and speed in signal processing is increasing. In response to the demand for large data volume, high transmission speed, and strong real-time performance in underwater acoustic real-time signal processing systems, heterogeneous collaborative processing is widely used in underwater acoustic operations. This article designs a heterogeneous master-slave operating system architecture scheme for multi-core DSP based on a real-time processing module for underwater acoustic signals. The 0 core of multicore DSP serves as the main core, running the ucLinux operating system for task monitoring and management, realtime data transmission and scheduling, and real-time data processing from the slave core; The 1-7 cores of the multi-core DSP serve as the slave core, running the SylixOS embedded real-time operating system, responsible for quickly responding to high-speed signals and completing real-time processing tasks for underwater acoustic signal algorithms. The software design of the master-slave architecture shifts the management of data transmission and reception to the master core, greatly reducing the consumption of data transmission and reception on the slave core. The slave core can focus more on underwater acoustic signal algorithm processing. Experimental verification shows that the multi-core DSP heterogeneous master-slave software architecture designed in this paper can effectively implement underwater acoustic target detection algorithms and meet the high-speed real-time signal processing requirements of underwater acoustic tasks.