Adaptive RBF Neural Network Research Articles

Some computational principles derived from cortical neuroanatomy

In robotic cooperation manufacturing occasions, like grinding, assembling, welding, etc., the position-force synchronization tracking control for robotic cooperative manipulators is critical to improve the comprehensive manufacturing quality with high-precision and high-adaptability. In terms of these problems, this paper proposes an adaptive neural synchronized impedance controller (ANSIC) for cooperative manipulators processing. The proposed method includes two non-parallel control loops of the cooperative system to achieve and guarantee the desired movement trajectory and manufacturing force of the cooperation task. In the inner position tracking loop, an adaptive RBF-neural network based synchronization sliding controller is designed to simultaneously estimate the uncertain dynamic parameters of the robotic manipulators and improve the cooperative position tracking precision. In the outer force tracking loop, another RBF-neural network is applied to reform the impedance control model automatically and compensate the position and stiffness errors of the uncertain workpiece environment. Mathematical proof and experiments under various conditions are conducted. The results demonstrate the effective convergences of both the cooperative processing trajectory and force despite the uncertain environments.

Adaptive neural network control for coordinated motion of a dual-arm space robot system with uncertain parameters

Control of coordinated motion between the base attitude and the arm joints of a free-floating dual-arm space robot with uncertain parameters is discussed. By combining the relation of system linear momentum conversation with the Lagrangian approach, the dynamic equation of a robot is established. Based on the above results, the free-floating dual-arm space robot system is modeled with RBF neural networks, the GL matrix and its product operator. With all uncertain inertial system parameters, an adaptive RBF neural network control scheme is developed for coordinated motion between the base attitude and the arm joints. The proposed scheme does not need linear parameterization of the dynamic equation of the system and any accurate prior-knowledge of the actual inertial parameters. Also it does not need to train the neural network offline so that it would present real-time and online applications. A planar free-floating dual-arm space robot is simulated to show feasibility of the proposed scheme.

A new method for fault prediction of model-unknown nonlinear system

In this paper, the subspace identification based robust fault prediction method which combines optimal track control with adaptive neural network compensation is presented for prediction the fault of unknown nonlinear system . At first, the local approximate linear model based on input–output of unknown system is obtained by subspace identification. The optimal track control is adopted for the approximate model with some unknown uncertainties and external disturbances . An adaptive RBF neural network is added to the track control in order to guarantee the robust tracking ability of the observation system. The effect of the system nonlinearity and the error caused by subspace modeling can be overcome by adaptive tuning of the weights of the RBF neural network online without any requisition of constraint or matching conditions. The stability of the designed closed-loop system is thus proved. A density function estimation method based on state forecasting is then used to judge the fault. The proposed method is applied to fault prediction of model-unknown fighter F-8II of China airforce and the simulation results show that the proposed method can not only predict the fault, but has strong robustness against uncertainties and external disturbances.

Robust H∞ control of multiple time-delay uncertain nonlinear system using fuzzy model and adaptive neural network

In this paper, a robust control method which combines fuzzy model-based control with adaptive neural network control is presented for a class of uncertain nonlinear system with multiple time-delay. The fuzzy T-S model with multiple time-delay is adopted for the approximate modeling of the nonlinear system with some unknown uncertainties, and fuzzy-model-based H ∞ control law is designed by means of LMI method. A full adaptive RBF neural network is added to the fuzzy H ∞ control in order to guarantee the robust stability of the controlled system. The effect of the unknown uncertainties and the error caused by fuzzy modeling can be overcome by adaptive tuning of the weights, centers and widths of the RBF neural network on line, and no constraint or matching conditions are required. The stability of the designed closed-loop system is thus proved. The proposed method is applied to a multiple time-delay nonlinear chaotic system and the simulation results show that the proposed method cannot only stabilize the chaos systems, but has strong robustness against uncertainties and external disturbance.

DYPOF: Dynamically Adaptive RBF Neural Network with Potential Functions

We present a method for data classification, which performs recognition based on a set of potential fields synthesized over the domain on input space by a number of potential function units. Proposed is DYPOF (DYnamic POtential Functions) neural network (NN), based on radial basis functions with a two-stage training procedure. A fundamental component in building DYPOF is a potential function entity (PFE), designed to generate a respective decision potential function. The desirable shape of the potential field characterizing the distribution of training set is synthesized by adjusting the weights as well as the parameter vectors of cumulative potential functions generated by the PFEs. The automatic adjustment of the minimum necessary number of hidden units-learning adjustment units (LAU)-for a given set of teaching patterns provides the network with a capability of performing dynamic adaptation and self-organization. We investigate the dependence of our method on these parameters and apply it to several data sets. The results indicate the power of the PFEs in generating classification solutions for various shapes of teaching patterns that are robust with respect to noise in the data.