Parallel implementation of a model-based spectral estimator for Doppler blood flow instrumentation

This paper presents a parallel implementation of a model-based spectral estimation method, based on the modified covariance parametric algorithm. This method has been implemented in order to take advantage of its resolution benefits when it is applied to spectral estimation in Doppler blood flow instrumentation. As the algorithm is computationally intensive and must be executed in realtime, parallel processing techniques have been utilized. Two different approaches are presented, these being implemented on a transputer-based system. Performance analysis of these approaches is presented together with concluding remarks. The results are being applied in the development of a real-time spectrum analyzer for pulsed Doppler blood flow instrumentation