DEVELOPMENT OF A TRACTOSENSOR DEVICE FOR EFFECTIVE FARM MACHINERY UTILIZATION

The traditional methods of estimating farm machinery performance parameters are tedious and time consuming. This has hindered the development of investment sustainability and growth of agricultural mechanization in Nigeria. The use of sensors and microcontroller will enhance the level of estimation accuracy. This study was conducted at Tada-Shonga irrigation scheme. ESP-32 microcontroller was used for automated real-time data collection and estimation of specific fuel consumption, depth of cut and forward speed during farm operations. Sensors were strategically mounted on an MF 375 tractor to capture these parameters. An artificial neural network (ANN) model was developed using Neural Designer software XL package to predict the specific fuel consumption, depth of cut and forward speed of operation. The developed ANN model has topology structures of (9-10-5-5) of Multilayer perception. The input-output variables relationship show that an appropriate value of 20.71 cm, 3.74 km/h and 11.4 L/kW/h for depth of cut, forward speed and specific fuel consumption are good for tillage operation at Tada-Shonga irrigation scheme. The coefficient of determination (R2) of 0.933 and 0.918 were obtained for specific fuel consumption and capacity utilization index respectively. The analysis shows that the input variables have a positive impact on the output variables. The Mean Absolute Percentage Error for the utilization index, mechanization index, capacity utilization index, specific fuel consumption and profitability index are 3.26%, 11.91%, 2.34%, 2.15% and 2.00% respectively. This research work has proven the effectiveness of sensors to accurately monitor simultaneously and in real-time the critical parameters influencing farm machinery performance on experimental fields.