Improved generalized prediction control method for work arm of underwater robot

Abstract

The invention discloses an improved generalized prediction control method for a work arm of an underwater robot. A T-S fuzzy identification model suitable for nonlinear system identification is established, a real-time CARIMA parameter model of the underwater arm is obtained by identification, a step factor is introduced to form a progressive relation, a progressive generalized prediction controlalgorithm is used to generate an optimal control variable via rolling prediction optimization, and a controller of a hydraulic driver is acted on, so that the underwater arm generates a correspondingmotion. According to the improved algorithm of the invention, it is not required to know the structure of a controlled object model, dynamic change of the system can be reflected timely, tedious matrix calculation processes are omitted, the computational complexity is reduced, and the algorithm is simpler and more practical.

Description

Technical field [0001] The invention relates to an improved generalized predictive control method of an underwater robot operating manipulator. Background technique [0002] An underwater robot is a device that moves freely underwater through remote control or autonomous control, has an environmental perception system, and uses robotic arms and other tools to replace or assist manual completion of underwater tasks. It is widely used in scientific research, social economy, Military and other fields are the most effective and potential high-end intelligent equipment for marine exploration and underwater operations, and a large number of specialized mature commercial products have emerged. [0003] The underwater manipulator is an indispensable operating tool for underwater robots. At present, hydraulically driven manipulators are mostly adopted by underwater robots. Such manipulators have a large power density ratio, strong load capacity, and are easy to achieve waterproof sealing. ...

AI Technical Summary

Problems solved by technology

However, the calculation amount of the GPC algorithm is relatively large, and the conventional identification algorithm used to obtain the model parameters is also poor in adaptability. Therefore, it is necessary to design a motion control algorithm suitable for underwater manipulators based on the GPC algorithm.

[0007] Chinese patent document CN 104865979A discloses an adaptive generalized predictive control method and system for sewage treatment process, which solves the problem that generalized predictive control cannot achieve stable control when dealing with large disturbances

However, this method uses fuzzy neural network model identification, and the generalized predictive control algorithm used is not simplified, resulting in long model identification time and control rate calculation time, which is not suitable for use in fast-moving fields such as robotic arms.

Images