Transcale fusion and estimation of 2-D multisensor dynamic systems

This paper presents a transcale data fusion scheme for a class of discrete two-dimensional (2-D) dynamic systems, which are measured by several different-scale sensors. The 2-D Haar wavelet transform is used to link the state notes at each of the scales within a 2-D time block and the standard 2-D state space model of the dynamic systems is established. This model satisfies the requirements of the 2-D Kalman filtering, which offers an optimal estimation algorithm for the 2-D dynamic systems. An example is included to illustrate the obtained results.