FAST PARTIAL FREQUENCY SPECTRUM COMPUTATION FOR REAL-TIME INFORMATION ACQUISITION SYSTEMS.

摘要

In radar and sonar systems, it is important to get arnfrequency spectrum on real time. To implement a highrnperformance system, the update interval should be asrnshort as possible, but this requirement is not easilyrnachievable due to the limited hardware resource.rnAlthough some fast Fourier transform (FFT) algorithmsrnand pruning techniques may be able to reducerncomputational complexity, the use of them does notrnguarantee the best solution. To overcome this problem,rnwe propose a pruned generalized FFT (PGSFFT)rncombining the GSFFT and transform decomposition (TD).rnSince the PGSFFT takes advantages of both GSFFT andrnTD, it is possible to reduce the computational complexityrnof the system. The enhanced computational efficiencyrnleads to the improvement of the system performance. Tornverify the performance of the proposed method, wernanalyze the required operations for the PGSFFT andrnperform simulations. Simulation results show that thernproposed PGSFFT has better performance than otherrnFFTs or pruning algorithms.