Based On Rough Set-Neural Network For Autoloading Detection And Diagnosis

Automatic control system is filled artillery shells for the input device is one very important sub-system, its structure is very complex and a higher incidence of failure, the system performance will affect the overall operation of the gun. Therefore, how to improve the reliability of artillery to reduce the incidence of failure in the modern intelligence war has very important significance. With smart technology, many scholars using neural network technology for fault diagnosis, and the rough set theory as a pre-sample data pre-processing system, but also applications in many areas has been confirmed. Therefore, this article has designed a self-loading control system, and on this basis, the use of rough sets-neural network algorithm, to solve the automatic loading system fault diagnosis.This article first analyzes the loading system to complete the basic tasks and control requirements, combined with practical application, the design of the overall control scheme. LPC2294 chip used as the master control system, additional features of other units, completed a loading control system hardware design and software development. Display interface design using VS2008, using Serial Port complete PC and control systems communication.Use the loading system as the research object, the control key parts of the process action status, focus on the rough set and neural network for fault diagnosis algorithm on this system. First, data collection, the formation of the sample data table, and then use the Pawlak rough set theory to complete the attribute importance algorithm of data reduction, removal of redundant attributes;Then, use the reduction data as the sample data, Radial Basis Function network training, at the training process, by constantly adjustment the expansion of Radial Basis Function constant, optimal output of the network, and finally a set of test the network, verify that the pros and cons of network diagnostics.Experiments show that this system has a high diagnostic efficiency, greatly improving the reliability of the automatic loading system.This article complete hardware and software system debugging for automatic loading control system; completed the rough set and artificial neural networks fault diagnosis algorithm, diagnosis is accurate, verified the feasibility of the program, meet the design requirements of the system.