A Load Cell Test Compensation Method Based on Fuzzy Recognition

Abstract

The invention relates to a weighing method, in particular to a weighing sensor testing compensation method based on fuzzy recognition. The method comprises the steps of building a weighing system measured by multiple sensors; evaluating the influence degree of influence factors on data output by the weighing sensors, and building a weight set of interference factors; building a compensation coefficient evaluation set of an evaluation system; defining a single-factor membership fuzzy evaluation set, and according to data actually collected by the sensors, substituting the data into a matrix R,and performing comprehensive fuzzy evaluation. The method has the advantages that by means of fusion of the multiple sensors, the weighing measuring accuracy is improved; the method has the advantagesof being capable of performing adjustment according to different application environments and being good in expansibility; calculation is easy, on the condition that the number of the sensors is notlarge, not much calculation time is needed, and therefore no adverse effect is caused on the dynamic performance of the sensors.

Description

technical field [0001] The invention relates to a weighing method, in particular to a fuzzy recognition-based load cell test compensation method. Background technique [0002] When agricultural machinery is working in the field, it may be necessary to measure the weight of harvested agricultural products in real time. Due to the unstable working environment, agricultural machinery and vehicles may be subjected to impulses from all directions when working in the field, which will reduce the measurement accuracy of the load cell installed on the agricultural machinery; at the same time, the unevenness of the ground will also make the weighing The heavy sensor does not necessarily maintain the ideal horizontal position during measurement, which affects the measurement accuracy, resulting in large differences in yield estimates, which is not conducive to the industrialization of precision agriculture technology. [0003] Therefore, how to perform weight compensation and correct...

AI Technical Summary

Problems solved by technology

However, as more and more influencing factors are considered, the number of sensors required also increases, and the amount of sensor test data obtained is also large, and the correlation between them is difficult to quantitatively analyze, and it cannot be directly identified by a simple method analyze

The data-based neural network method is suitable for the establishment of such complex nonlinear relationships, but requires a large amount of training data, and sometimes the neural network is difficult to converge during training

In order to improve the convergence performance, the linear transfer function of neurons can be changed to a nonlinear transfer function, or the number of neurons and the number of hidden layers of the network can be increased, which will cause the neural network to be too complex and have a large amount of measurement, which is not conducive to practical applications. to promote

At present, how to use the test data of multiple (more than 3) sensors to compensate the load cell test, there is no simple, easy, and high compensation accuracy method

Images