Three-domain fuzzy wavelet broad learning system for tremor estimation

摘要

This paper proposes a new three-domain fuzzy wavelet broad learning system (TDFW-BLS) for tremor estimation in tele-operation. In the feature layer of our novel method, feature nodes can be mapped by a three-domain fuzzy wavelet sub-systems (3DFWs) and the k-means method is applied to determine the parameters in the TDFWs. The architecture of the new proposed system maps the input of different dimensions to different groups feature nodes by 3DFWS. In the enhancement layer, feature nodes are mapped to enhancement nodes. Moreover, feature nodes and increment nodes can be concatenated into a matrix to map the total output of the novel system by the full connection layer. Finally, the semi-physical simulation experiment is designed to demonstrate the effectiveness of the novel TDFW-BLS. Meanwhile, it is compared with some existing methods and the results have shown superior performance.