Performance Analysis and Modelling of Impact-based Sensor in Yield Monitor System

In order to accurately acquire the spatial distribution of grain yield, the impact-based yield monitor system of grain combine harvester was independently developed. The yield monitor system consists of a flow sensor module, a data acquisition module, a GNSS module, and a yield management terminal In this paper, the designed system was used for collecting yield data. The yield prediction model for different areas was established by the voltage, then it was applied to predict regional yield. The universality of model was analyzed. Based on the characteristics of spatial field variability, the preprocessing method of threshold filtering and the local average interpolation was used before establishing the relational model between actual yield and voltage. Field experiments included data acquisition part and model establishment part. The Data acquisition experiments were carried out in two fields, which were respectively defined as Fl and F2. The experiment in F1 were repeat 3 times, which were represented as group A1 similar to A3. The experiment in F2 were repeat 6 times, which were B1 similar to B6. The relational model was established between weight and voltage of each group. The mutual prediction verification was performed to demonstrate model universality. As a result, F2 yield was predicted by predictive model of Fl indicated that the relative error was 20.06%, which were not universal. The intra-group prediction results of F1 showed that the lowest relative error was 6.36%, the accuracy of the model need to improve furtherly. When B1 and B2 groups were mutually modeling and verification in the F2, the relative error of predicted yield was less than 5%. Modeling and verification accuracy R-2 were both above 0.9, which proved predictive models of B1 and B2 were highly accurate. However, it was not suitable for other groups forecast. The same result also appeared in B5 and B6 in the F2. The results showed that the system can correctly judge the grain yield changes. The plot of yield map in the plane-coordinate system can provide reference for fine farming and harvesting in the next quarter. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.