model of neural-adaptive fuzzy inference system

Soil-wheel interaction is very attractive topic for agricultural researchers for its effect on the energy consumption and soil properties especially in agriculture. The purpose of this research is measuring soil-wheel contact area under the effect of independent variables of vertical load on the tire, tire inflation pressure and tire forward speed at controlled condition of soil bin also its prediction using adaptive neuro-fuzzy inference system. The tests were accomplished at two forward speeds (0.386 and 0.879 km/h), three inflation pressures (18, 25 and 32 psi), three vertical loads (150, 300 and 450 kg) plus 3 replications and totally 54 passes. All data analysis was done using Genstat software. Results showed that increase of vertical load on the wheel caused to increase of soil-wheel contact area and increment of inflation pressure caused to decrease of it. Low forward speed had not effect on the soil-wheel contact area. Furthermore, correlation coefficient (R2) of ANFIS models (0.9182) was very more than regression one (0.359). Thus, ANFIS models could predict the soil-wheel contact area with high accuracy using measured input variables included vertical load, tire inflation pressure and forward speed at soil bin.