Output-redefinition based flexible mechanical arm time-scale separation robust control method

Abstract

The invention discloses an output-redefinition based flexible mechanical arm time-scale separation robust control method. The method comprises the following steps: establishing a flexible mechanical arm dynamical model; regulating output of the flexible mechanical arm dynamical model to obtain an input-output sub system and an inner dynamic sub system; designing a robust control method; designinga state feedback control method; obtaining first control input according to the robust control method; obtaining second control input according to the state feedback control method; combining first control input with second control input to obtain combined data; inputting the combined data to the flexible mechanical arm dynamical model to obtain output data; separately comparing the output data with expected data and elastic vibration data of a joint angle of a flexible mechanical arm to obtain a compared difference value; and when the compared difference value is within a comparison thresholdvalue range, outputting a mark which represents that control is effective. The embodiment of the invention can realize tracing an expected instruction by the joint angle of the mechanical arm, also realizes suppression on an elastic modal, and reduces vibrations.

Description

technical field [0001] The invention relates to the technical field of robot control, in particular to a time-scale separation robust control method for a flexible manipulator based on output redefinition. Background technique [0002] Due to the strong uncertainty in the dynamic model of the flexible manipulator, it is not conducive to the design of the control algorithm. A variety of control theories have been successfully applied to the control of flexible manipulators. Among them, the proportional-derivative control algorithm is simple, but it has poor resistance to external disturbances. Intelligent learning control has a better learning effect on uncertain information, but online update parameters often lead to solution The speed is on the slow side. [0003] "Two Performance Enhanced Control of Flexible-Link Manipulator with System Uncertainty And Disturbances" (BIN XU, YUAN YUAN, "SCIENCE CHINAINFORMATION SCIENCE", Vol.60050202:1-050202:11) uses output redefinition ...

AI Technical Summary

Problems solved by technology

[0002] Due to the strong uncertainty in the dynamic model of the flexible manipulator, it is not conducive to the design of the control algorithm

A variety of control theories have been successfully applied to the control of flexible manipulators. Among them, the proportional-derivative control algorithm is simple, but it has poor resistance to external disturbances. Intelligent learning control has a better learning effect on uncertain information, but online update parameters often lead to solution Slow

[0003] "Two Performance Enhanced Control of Flexible-Link Manipulator with System Uncertainty And Disturbances" (BIN XU, YUAN YUAN, "SCIENCE CHINAINFORMATION SCIENCE", Vol.60050202:1-050202:11) uses output redefinition for the dynamic model of flexible manipulators After processing, the model avoids the non-minimum phase problem. For the uncertain items of the model, the paper uses the neural network for approximation learning. The control algorithm finally makes the joint angle effectively track the expected command. However, in order to obtain a good approximation effect, the neural network is online There are many update parameters, which makes the calculation speed of the algorithm slow

Images