Distributed filtering techniques of multiresolution multisensor based on wavelet transform

Wavelet transform is very suitable for processing multi-scale signals, and it can be used in multi-resolution filtering. In this paper, an algorithm for dynamic distributed filtering of multi-resolution multi-sensor is proposed, based on quick-lifting wavelet transform. By avoiding big matrix computation and eliminating the assumption of white noises for wavelet coefficients, it is efficient and highly parallelizable. Simulation results also show that this algorithm improves the performance of multi-scale data filtering. Moreover the quick-lifting wavelet transform is used as a signal-linking bridge between multi-resolution levels. The algorithm can thus be used for dynamic multi-resolution multi-sensor data fusion.