Abstract
The Multiple Signal Classification (MUSIC) is a powerful algorithm for determining the Direction of Arrival (DOA). With the increase of the number of elements in a fixed array and the increasing bandwidth of the detection signal, the MUSIC algorithm is faced with a double challenge in complexity and computation It is urgent to improve the computational efficiency to realize in real-time in the passive sonar underwater acoustic signal processing This paper analyzes the computational characteristics of the broadband MUSIC algorithm in underwater acoustic array signal. Combined with large-scale array primitives and multi-frequency points, the GPU parallel optimization of the broadband underwater acoustic array signal MUSIC algorithm is implemented. And the experiments on CPU and GPU platform confirm that GPU parallel has a very good acceleration effect.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
