Feasibility of implementation of intelligent simulation configurations based on data mining methodologies for prediction of tractor wheel slip

Abstract

This paper deals with implementation of intelligent simulation configurations for prediction of tractor wheel slip in tillage operations. The effects of numeral variables of forward speed (2, 4, and 6 km/h) and plowing depth (10, 20, and 30 cm), and nominal variable of tractor driving mode (two-wheel drive (2WD) and four-wheel drive (4WD)) on tractor rear wheel slip were intelligently simulated utilizing data mining methodologies of artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). Neuro-fuzzy potential of the ANFIS simulation framework against neural ability of the ANN simulation framework was apprised. Results confirmed higher efficiency of the best configuration of the ANFIS simulation framework with satisfactory statistical performance criteria of coefficient of determination (0.981), root mean square error (1.124%), mean absolute percentage error (1.515%), and mean of absolute values of prediction residual errors (1.135%) than that of the ANN simulation framework. Physical perception obtained from the ANFIS simulation results demonstrated that the wheel slip increased nonlinearly with increment of forward speed and plowing depth, while it decreased as tractor driving mode changed from the 2WD to 4WD. Therefore, the best configuration of the ANFIS based intelligent simulation framework implemented in this study can be used for further relevant studies of tractor rear wheel slip as a reference.