Digital filter sampling using non-uniform sampling
Krishna Busawon, Wai Pang Ng · Northumbria Research Link (Northumbria University) · 2007
In this paper, a non-uniform sampling algorithm is proposed for one-dimensional bandlimited time varying signals. The main feature of the proposed sampling scheme is that the sampling steps are inversely proportional to signal gradient or slope of the signal. As a result, a smaller sampling step is obtained whenever the gradient is high and vice-versa. Thus, a better representation of the signal is obtained when it is linearly interpolated between the samples. The proposed sampling algorithm is then applied to the design of digital filters using the well-known impulse invariance method. It is shown that the nonuniform sampling allows to design a FIR filter whose length is 14 times shorter than what is obtained with uniform sampling; with both filters having similar characteristiscs. Simulations are provided to show the performance and characteristics of the sampling sequence.