Artificial intelligent denoising spectrograms approach for enhanced chatter detection in robotic machining

Accurate chatter detection becomes vital to preventing chatter issues. Because of the complicated dynamics involved, chatter can be particularly difficult to identify and reduce in robotic machining, where robotic arms are used for material removal operations. In order to detect chatter in robotic machining, this paper provides an artificial intelligence (AI) approach called attention-based denoising and adaptive thresholding methods. A potent method for spectrogram denoising, which provides flexibility to adaptively focus on pertinent information while maintaining significant characteristics. By utilising an efficient attention mechanism that can extract pertinent information from the input data, it provides a high-resolution spectrogram in comparison to the conventional method. The efficiency of the proposed method was demonstrated by a series of experimental tests. The average entropy of the spectrograms generated by the conventional and proposed methods is found to be 12.66 and 10.01, respectively.