A sliding mode strategy for adaptive learning in multilayer feedforward neural networks with a scalar output

The learning process of artificial neural networks is an important and complex task in the supervised learning field. The main difficulty of training a neural network is the process of fine-tuning the best set of control parameters in terms of weight and bias. This paper presents a new training method based on hybrid particle swarm optimization with Multi-Verse Optimization (PMVO) to train the feedforward neural networks. The hybrid algorithm is utilized to search better in solution space which proves its efficiency in reducing the problems of trapping in local minima. The performance of the proposed approach was compared with five evolutionary techniques and the standard momentum backpropagation and adaptive learning rate. The comparison was benchmarked and evaluated using six bio-medical datasets. The results of the comparative study show that PMVO outperformed other training methods in most datasets and can be an alternative to other training methods.

Download Full-text

Dual feedforward neural networks based synchronized sliding mode controller for cooperative manipulator system under variable load and uncertainties

Proceedings of the Institution of Mechanical Engineers Part C Journal of Mechanical Engineering Science ◽

10.1177/0954406220916484 ◽

2020 ◽

Vol 234 (19) ◽

pp. 3859-3872

Author(s):

Jin Wang ◽

Anbang Zhai ◽

Fan Xu ◽

Haiyun Zhang ◽

Guodong Lu

Keyword(s):

Neural Network ◽

Neural Networks ◽

Radial Basis Function ◽

Basis Function ◽

Sliding Mode ◽

Feedforward Neural Networks ◽

Internal Force ◽

Variable Load ◽

Sliding Mode Controller ◽

Radial Basis

The problem of simultaneous position and internal force control is discussed with cooperative manipulators system under variable load and dynamic uncertainties in this study. A position synchronized sliding mode controller is proposed in the presence of variable load, as well as modeling uncertainties, joint friction, and external disturbances. To deal with the complex situation brought by variable load, virtual synchronization coupled errors are introduced for internal force tracking control and joint synchronization in the meantime. Dual feedforward neural networks are adopted, where a radial basis function-neural network based dynamic compensator and a radial basis function-neural network based internal force estimator are established, respectively, so that precise dynamic knowledge and force measurement are out of demand through their cooperation. Together with simulation studies and analysis, the position and internal force errors are shown to converge asymptotically to zero. Using Lyapunov stability approach, the proposed controller is proven to be robust in face of variable external load and the aforementioned uncertainties.

Download Full-text

Dynamic Neural Sliding Mode Control for Uncertain Nonlinear Systems

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.898.701 ◽

2014 ◽

Vol 898 ◽

pp. 701-704

Author(s):

Hong Yu Gao ◽

Jun Liu ◽

Yuan Gao ◽

Ke Yong Shao

Keyword(s):

Neural Networks ◽

Nonlinear System ◽

Asymptotic Stability ◽

Theoretical Analysis ◽

Adaptive Learning ◽

Upper Bound ◽

Sliding Mode ◽

Uncertain Nonlinear Systems ◽

System A ◽

Upper Boundary

For uncertainty nonlinear system, a dynamic neural sliding mode (SM) controller is designed in this paper. A union upper bound of uncertainty is constructed. The union upper bound is from the uncertainty and disturbance of the system, and it is unknown. The designed dynamic neural SM controller can ensure the system asymptotic stability. Using Radial Basis Function Neural Networks (RBFNN) to learn adaptively the union upper boundary of the uncertainty and verifying its validity by theoretical analysis and simulation examples. The design scheme of the adaptive learning upper bound reduces the condition of theoretical analysis of SMC, effectively suppresses the chattering.

Download Full-text