Abstract
The numerical solution of ordinary differential equations has been widely used in many fields including wave propagation analysis. To represent a continuous function in terms of its discrete sampled values in a sequence, it should satisfy the sampling theorem. However, in conventional wave propagation analysis, the experiential finite difference technique has generally been used. In this paper, the sampling extension which converges more rapidly than in the case of classical cardinal series is proposed. The extension and aliasing errors including the truncation error are described specifically. The sampling extention is also generalized to include the sampled values of the derivative and integral of the signal.