Research On Neural Dynamics Of Variable Gain Periodic Rhythm And Its Application In Inverse Kinematics Of Surgical Robot Redundant Manipulator

In order to overcome the problems of low accuracy of traditional surgery,excessively long operating time,which may cause fatigue and error,and hand jitter,surgical robots need to be real-time,stable,efficient,and accurate.Since its function execution is mainly completed by a controlled robotic arm,the robotic arm is required to achieve real-time high-precision control.However,the joint angle may drift when the robotic arm moves repeatedly,resulting in errors.In order to solve the repetitive motion problem of surgical robot manipulator,this paper abstracts the problem as a time-varying matrix equation model.Related researches on neural dynamics are carried out and applied to the time-varying matrix equation with noise,and combined with swarm intelligence algorithm to solve the biconvex optimization problem and the repetitive motion problem of the redundant manipulator arm of surgical robot.First of all,in order to solve the redundant manipulator motion problem model with limited intensity noise,this paper constructs a time-varying matrix equation model for the peoblem,and proposes a variable gain periodic rhythm neural network.Four types of activation functions and four types of noise are used for test verification,and the theory proves that the neural network has good convergence.After that,the network is compared with the traditional cognitive rhythm network and the zeroing neural network through simulation experiments.The results show that the variable gain periodic rhythm neural network has better convergence and anti-noise capabilities.Then,combining the particle swarm optimization method and the variable gain periodic rhythm neural network,a new collaborative neurodynamic network for solving biconvex problems is proposed.The network has time-varying characteristics and can realize the cooperative work of two neural networks on a two-layer time scale.The theory proves that the network is stable and converges with the probability of one.The simulation experiment shows that the cooperative variable gain periodic rhythm neural network has faster convergence efficiency and certain anti-noise ability.Finally,the variable gain periodic rhythm neural network solver is used to solve the inverse kinematics problem of the redundant manipulator of the surgical robot,that is,the inverse kinematics problem is transformed into a quadratic programming problem solved in the velocity layer.In addition,a collaborative variable gain periodic rhythm neural network is used to solve the inverse kinematics problem of the manipulator.In order to verify the effectiveness of the above-mentioned network,the experimental simulation comparison between the variable gain periodic rhythm neural network and the zeroing neural network solver shows that the variable-gain periodic rhythm neural network is effective and the effect is better than the zeroing neural network method.At the same time,the results also verify that the collaborative variable gain periodic rhythm neural network can achieve high accuracy and low error when solving this inverse kinematics problem.