Test Data Proceeding Method Of Agricultural Machinery Centroid Position Measuring System Based On Gray Neural Network Algorithm

Small agricultural machinery has been widely used in the greenhouses and hilly areas. Because of its characteristics, it can do lots work such as soil cultivation, plowing, weeding in greenhouses and hilly areas. The small agricultural machinery has become the most important machine in hilly areas. In order to design the ideal farm machinery for these areas, the parameters of small agricultural machinery must be optimal, and the centroid location is one of the most important parameters.Adopting the single fulcrum drive inclined platform method to measure the centroid location of small agricultural machinery. In the course of testing, the error is inevitable. The error will influence the accuracy of centroid location. According to the error in the measurement, the method of combining of grey forecasting model and BP neural network, and the error analysis theory are all used to process the measure dada of small agricultural machinery. The main work and achievements of this thesis are listed as followed:1. Introduce the measurement method and the measurement principle of the centroid location of small agricultural machinery. Adopting the formula of centroid to calculate the centroid location of small agricultural machinery. According to the inadequate of the measurement platform, the existing measurement unit and measurement method are improved to get precision data.2. Based on principles of mathematics and statistics, measurement dada is processed. Excluding gross error in the measurement dada by the Bessel equation and smoothing the measurement dada to reduce the error.3. Based on grey theory and BP neural network theory, grey BP neural network forecasting model is established. The model is used to process the measurement dada to reduce the error. The optimization of background value and initial of grey forecasting model is proposed. The structure of BP neural network is optimized.4. According to the characteristics of dada process and small agricultural machinery, the experiment is designed to test and verify the grey BP neural network forecasting model. The processing technology and assembly precision are improved to reduce the error exist in the process. Angle displacement sensor is replaced by the linear displacement sensor to improve the testing accuracy. And the system and platform of the measurement of centroid location are established. The theoretical analysis and the processing of measurement dada are tested and verified through the experiment.