A Signal Processing Algorithm Based on Multiple Microprocessors for an Underwater Acoustic Imaging System

摘要

An algorithm has been designed to provide near real-time location and classification information to an operator of an underwater acoustic imaging system. The overall system to detect objects buried up to five meters in unconsolidated marine sediments consists of a specialized projector co-located with a five meter receiving line array. This system is described in basic terms together with the principles of the signal processing that extracts information from the scattered acoustic signals. Using simulated objects, the angular resolution and imaging performance of a number of system designs were evaluated. To obtain near real-time execution, the algorithm was optimized and partitioned using parallel, pipeline, and double buffering techniques for independent but synchronized operation on multiple microprocessors. The feasibility of the design approach was experimentally demonstrated using four single board computers in a microcomputer development system. (Author)