Time-frequency representation (TFR) based onAdaptive Optimal Kernel (AOK) normally performs well onlyfor monocomponent signals and has poor noise robustness. Toovercome the shortcomings of AOK TFR mentioned above, anew TFR algorithm is proposed here by integrating nonlinearmode decomposition (NMD) with AOK TFR. NMD is used todecompose multicomponent signals into a bundle of meaningfuloscillations and then AOK is applied to compute the TFR ofindividual oscillations, finally all these TFRs are summedtogether to generate one TFR. Through quantitative comparisonwith other TFR methods to both simulated and real signals, thesuperiority of proposed TFR based on NMD and AOK onremoving noise and many other measurement index of TFR areshown.
展开▼