Mechanical arm high-precision tracking control method with high robustness

Abstract

The invention discloses a mechanical arm high-precision tracking control method with high robustness, and belongs to the technical field of power supplies. The method comprises the following steps: establishing a multi-disturbance n-degree-of-freedom mechanical arm complex dynamic model by adopting a Lagrange method, converting the multi-disturbance n-degree-of-freedom mechanical arm complex dynamic model into a state space equation form, and converting system uncertainty into an unknown nonlinear function; designing an adaptive recurrent neural network to carry out online approximation on the system uncertainty in a mechanical arm system; designing a second-order nonlinear disturbance observer to estimate unknown disturbance suffered by the mechanical arm system, and integrating the designed adaptive recurrent neural network and the second-order nonlinear disturbance observer to comprehensively estimate performance items; and based on the adaptive recurrent neural network approximation value and the second-order nonlinear disturbance observer estimation value, designing an adaptive recurrent neural network tracking controller. The technical purposes of improving the anti-interference capacity of the mechanical arm and achieving high-precision trajectory tracking control are achieved, the anti-interference capacity of the system is improved, and the robust performance of the system is enhanced.

Description

technical field [0001] The invention belongs to the technical field of manipulator control, and relates to a high-precision tracking control method of a manipulator with strong robustness. Background technique [0002] Due to the advantages of flexibility, high efficiency, intelligence and adaptability to harsh working conditions, robotic arms are widely used in industrial production and manufacturing, and gradually replace human labor, which not only improves production efficiency, but also saves production costs. Especially in the dangerous working conditions with severe environmental pollution, the harm to the staff is reduced. In order to ensure that the manipulator can complete the established trajectory tasks with high precision, the performance requirements of the manipulator control system are getting higher and higher. The manipulator is a multiple-input multiple-output system with strong nonlinearity, strong coupling and multiple time-varying characteristics. In ...

AI Technical Summary

Problems solved by technology

The control law of the PID control method is simple and easy to implement, and does not require accurate dynamic model parameters of the manipulator, but its control accuracy is poor and its robustness is poor; The speed is fast, but there is a phenomenon of "chattering" in the control process, which affects certain tracking accuracy and wears out the equipment; the calculation torque control method has better control accuracy, but it requires accurate model parameter support

However, in actual situations, it is difficult to guarantee the requirements of accurate model parameters; the adaptive control method has better system adaptability and learning ability, and can self-adjust by continuously learning the changes of system model parameters, but the learning process is complex and computationally intensive. Large and complicated to implement, it is only suitable for objects with simple structures; the robust control method achieves the effect of stable control by setting the maximum upper bound of the disturbance, which is easy to implement, but it needs to determine the maximum upper bound of the disturbance based on the experience and subjective judgment of engineers range, without a certain learning ability and adaptability; the neural network control method has a good universal approximation effect, and it can approximate the unknown nonlinear function of the system without modulus parameters

However, it does not consider the unknown external disturbance of the system, and a robust term needs to be introduced for compensation

Images