Research And Realization Of Fast Filtering Techniques

Filtering technology plays a key role in modern signal processing. Filters are devices with frequency selectivity. That is, a filter allows signal component with frequency in the pass-band to pass by, and repress signal component with frequency in the stop-band. Filtering technology has undergone a long-term evolution and has developed mature theory and methodology. It is widely used in many fields of signal processing, such as communication systems, measurement systems and instruments. In the filter design, we hope the filters have good amplitude-frequency characteristic in the pass-band, and fast response, namely short rise time in the output signal. However, the conventional typical filters can hardly satisfy both of the requirements at the same time.In many fields of signal processing, the response speed is as important as the accuracy. Now some methods of fast-filtering techniques have already been developed. Fast filtering techniques are in great needs and are of great significances for the development of filter theory. The research of this paper is based on the realistic demand, and proposes fast-filtering techniques in analog domain and digital domain, respectively, aiming to provide techniques and theory of reference for practical implementations.The main work of this paper is as follows.(1) In analog domain a kind of fast-response low-drift active low-pass filter is proposed, including the circuit structure, the analysis of its theory and the simulation results. What’s more, the comparison in filtering property and drifting property with non-drift active low-pass filter and conventional active low-pass filter has been carried out. The obtained results show that the proposed fast-response low-drift active low-pass filter has both low-drift and fast response which is obtained at the cost of a slight increase in oscillations.(2) In digital domain a dynamic two-stage filtering algorithm is proposed, in which two digital filters are connected with a dynamic binding mechanism. The steps of the algorithm and its simulation results are given in details. In addition, the dynamic two-stage filtering algorithm is implemented in a high-precision digital weighting sensor signal processing system and achieves satisfactory results. The dynamic mechanism in this algorithm makes it possible to achieve a good combination of short rise time and good filtering output as compared with conventional digital filtering techniques.Simulations and practical applications show that the fast-filtering techniques proposed in this paper are effective and useful, and have great value in practical engineering implementations. The simulations in this paper are carried out in Pspice and Matlab.