Intelligent Prediction of Ergonomics Evaluation Metrics in Human-Ai Collaborative Systems Using Multiple Machine Learning Models

Abstract

In human-AI collaboration systems, ergonomic evaluations have traditionally relied on subjective questionnaires or expert assessments, which are time-consuming, labor-intensive, and prone to bias. Moreover, the complexity of formulas and input factors further complicates the process, posing challenges to the automated and intelligent computation of ergonomic evaluation metrics. To address this issue, this paper proposes a machine learning-based framework for intelligently predicting ergonomics evaluation metrics. First, a human-AI collaboration experimental method is presented for ergonomics evaluation metric data collection. Then, four regression models are trained, namely, linear neural network, nonlinear deep neural network, random forest regressor with data augmentation, and random forest with k-fold cross-validation. These models are designed to predict ergonomics evaluation metrics based on influencing factors.  Data augmentation techniques are employed to expand the dataset and enhance the model's generalization capability. Finally, the models' accuracy and reliability in predicting these metrics are comprehensively evaluated using relative absolute error, mean relative absolute error, root mean squared error, relative root mean squared error, and the coefficient of determination (\(R^2\)), ensuring their validity. The results show that the proposed machine learning framework achieves high prediction accuracy. This enables effective automated ergonomic evaluations within human-AI collaboration systems and thus provides guidance for designing fluent and efficient human-AI teams.

Keywords: Human-AI Collaboration Ergonomic Evaluation Machine Learning Regression Prediction Data augmentation