End-to-End Surface Roughness Prediction Method Driven by Multi-Source Information Fusion

To address the limitations of traditional data-driven surface roughness prediction approaches that rely on a single information source and require manual feature engineering, a deep learning-based surface roughness prediction model was proposed. The model used static processing parameters and raw signals from multi-sensors as inputs. Parallel convolutional neural networks were employed to extract multi-scale features from the signals, and a hybrid attention mechanism was introduced for feature enhancement to improve prediction accuracy. Experimental results demonstrate that the proposed method which integrates multi-source sensor signals may predict surface roughness effectively, outperforming other methods which are based on a single sensor signal or traditional machine learning models. It provides an effective approach to predict the surface quality of products in intelligent manufacturing rapidly and accurately.