A assessment method for ergonomic risk based on fennec fox optimization algorithm and generalized regression neural network

Abstract

The Rapid Upper Limb Assessment method depends mainly on the subjective perception of the assessor, resulting in inconsistent results and a low sensitivity to changes in input variables. In this study, a new scoring system is developed using the Fennec Fox Optimization Algorithm and the Generalized Regression Neural Network approach to overcome the drawbacks of traditional method. First, the deep convolutional neural network was used to identify the keypoints of the human working posture in an image and calculate the joint angle. Second, the new model was used to improve the traditional method, and the prediction results for different postural risk scores were output. The proposed network was trained and tested, and the data were analyzed for comparison. Finally, the correlation between the top 15 predictions in the dataset and the scores was verified. The comparison results show that the proposed method performed better than the other methods in terms of the mean absolute error, mean square error, root-mean-square error, mean absolute percentage error, coefficient of determination, runtime, and spatial complexity. Additionally, the proposed method is more sensitive to small variations in inputs, reducing the likelihood of obtaining the same assessment scores for different postures. This increased sensitivity makes the scoring method more conservative, resulting in a more accurate risk assessment, minimizing potential oversights, and effectively reducing occupational risk. These results underscore the effectiveness of the proposed method in improving the traditional assessment.