Advanced machine learning for prediction of circumferential angle and lug spacing in hand tractor cage wheels using stacked ensemble machine learning

Abstract

Information on the performance of hand tractor cage wheels is crucial for implementing precision farming strategies in agricultural land tillage. Various approaches have been developed to predict cage wheel design, but previous efforts were limited to software, soil bins, and small scale testing. This study utilizes artificial intelligence with machine learning (ML) to optimize cage wheel design in hand tractors. We trained and evaluated multiple ML algorithms, including multiple linear regression, artificial neural network, k-nearest neighbor, adaptive boosting, extra trees regressor, and stacked ensemble machine learning (SEM), to predict circumferential angle and lug spacing. Models were assessed using the coefficient of determination (R²), and root mean square error (RMSE) based on features such as pull forces, lift forces, side forces, power, traction efficiency, and slip. SEM achieved the highest performance, with a perfect R² and an RMSE for circumferential angle prediction. SEM also showed strong performance for lug spacing, achieving an R² of 0.99 and an RMSE of 0.47. This study confirms the effectiveness of SEM in optimizing hand tractor wheel designs, supporting efficient solutions for agricultural engineering across varying field conditions.