Multi-dimensional evaluation and prediction of vibration comfort in electric loaders using ACO-Transformer

Engineering machinery plays a vital role in supporting modern economic development. The electric loader represents a key innovation driven by environmental protection and the pursuit of sustainable development. However, the absence of engine masking effects in electric machinery makes structural vibration and impact noise more pronounced. To address this issue, this study proposes an ant colony optimization-Transformer (ACO-Transformer) model that integrates the ant colony algorithm with the Transformer framework to accurately and efficiently evaluate the vibration comfort of electric engineering machinery. An improved objective evaluation method for vibration was employed to extract objective data from four measurement points, while 34 subjective scores were obtained through a structured subjective evaluation protocol. The combined analysis of subjective and objective data demonstrated the validity of incorporating additional vibration measurement points. Using these datasets, the ACO-Transformer model was developed to establish a mapping between multi-dimensional objective vibration parameters and subjective comfort ratings. Results indicate that the proposed model achieved high prediction accuracy (MAPE = 6.22%) and strong generalization performance (RMSE = 6.11). This study offers a novel approach for evaluating and predicting the vibration comfort of engineering machinery cabins.