一种低功耗水下目标识别系统设计

Design of Ultra-low Power System for Passive Underwater Target Recognition

针对新一代水下长期值更的被动目标识别系统的开发需求,研究并开发了一种低功耗、识别实时、准确的目标识别系统,该系统以低功耗芯片MSP430F5438与TMS320C5509A处理器为核心,配合大动态范围的模拟预处理系统,构建硬件平台,利用目标信号的小波变换域特征和轴频特征,实现了对舰船目标信号的识别和大小船分类;湖上和海上实验识别结果表明设计的系统平均正确检测概率达到83%,满足实际应用要求;由于采用周期性上电识别模式和超低功耗设计理论,设计的系统工作功耗低,识别算法全速运行时仍小于70mW,满足水下长期值更工作的特殊要求,有很好的开发应用前景。

In order to satisfy the needs for the new generation passive underwater target recognition system which must be of great ultra-low power for long-time working,real-time working,accuracy,the paper presents a design of hardware platform based on ultra-low power devices MSP430F5438,TMS320VC5509A and large dynamic range preprocessing module.Benefiting from the energy feature and the module maximum feature in some wavelet transform band and shaft frequency feature extracted by a fuzzy model,the whole system holds the capacity of high accuracy while the system is real time working.The results of experiments show that the designed system provides a good performance of low power consumption（less than 70mw when it works in a full speed operation mode）,real time computation,and accuracy.The detection and recognition Probability of the system reached 83%,which matches the factors that the new generation passive underwater target recognition system expects,and has a good application prospect.