Compound control method for electric loading system

Abstract

The traditional CMAC control algorithm and PD algorithm are combined and applied to torque control of a steering engine electric loading system, though the learning speed is fast, the problem of too frequent control output variation is introduced, is partially characterized by the facts that the output is not smooth enough, and can be integrally characterized by output divergence along with the passage of learning time and accumulation of errors and the like. The invention relates to a compound control method for an electric loading system, and discloses a novel CMAC control algorithm based on input vector nonuniform quantization and a Gaussian function. The compound control method optimizes a nonuniform quantization method according to input features, introduces the concept of quantitative distance to determine Gaussian weight and activated area of a node, can dynamically adjust CMAC generalization performance, adopts a stray mapping method and greatly saves storage resources. The CMAC and PD compound intelligent electric loading control method can effectively suppress the over-learning phenomenon of the control system, reduces memory usage rate, and improves tracking and control precision.

Description

technical field [0001] The present invention is a novel control method for the electric loading system, specifically, a new CMAC algorithm applying non-uniform quantization of Gaussian weights and traditional PD control combined to suppress excess torque of the electric loading system and improve control accuracy and stability. Background technique [0002] In the load simulation test of the UAV steering gear, the use of electric loading can effectively reproduce various loads on the UAV steering gear in the air, and can predict the technical performance indicators of the steering gear system in the laboratory environment The automatic test can save the development cost of the UAV, shorten the development cycle of the UAV, and effectively improve its flight reliability and success rate. [0003] The electric loading system is a passive torque control system. In the movement of torque loading following the steering gear system, redundant torque will inevitably be introduced, ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to effectively suppress the redundant torque in the electric loading process, and at the same time solve the over-learning phenomenon of the traditional CMAC control method, and provide a new type of Gaussian weight non-uniform quantization CMAC control method. Optimized the non-uniform quantization method, introduced the concept of quantization distance to determine the Gaussian weight and CMAC activation area, dynamically adjusted the generalization performance of CMAC, and used the remainder method to map the weight of the concept space to the physical space, reducing the waste of storage space and speeding up learning efficiency , with better real-time control and control precision

Images