Catálogo de publicaciones - libros

This paper considers reinforcement learning control with the self-organizing map. Reinforcement learning uses the observable states of objective system and signals from interaction of the system and environment as input data. For fast learning in neural network training, it is necessary to reduce learning data. In this paper, we use the self-organizing map to partition the observable states. Partitioning states reduces the number of learning data which is used for training neural networks. And neural dynamic programming design method is used for the controller. For evaluating the designed reinforcement learning controller, a double linked inverted pendulum on the cart system is simulated. The designed controller is composed of serial connection of self-organizing map and two Multi-layer Feed-Forward Neural Networks.

In this paper a neural network mapping of magnet based position sensing system for an autonomous robotic vehicle. The position sensing system using magnetic markers embedded under the surface of roadway pavement. An important role of magnetic position sensing is identification of vehicle’s location. The magnetic sensor measures lateral distance when the vehicle passes over the magnetic marker. California PATH has developed a table-look-up as an inverse map. But it’s requires too many memories to store the vast magnetic field data. Thus we propose the magnetic guidance system with simple mapping using neural network.

This paper considers the problem of regulating the output voltage of an asymmetric half-bridge (AHB) DC/DC converter via fuzzy integral control. First, we model the dynamic characteristics of the AHB DC/DC converter with the state-space averaging method, and after introducing an additional integral state of the output regulation error, we obtain the Takagi-Sugeno (TS) fuzzy model for the augmented system. Second, the concept of the parallel distributed compensation is applied to the design of the TS fuzzy integral controller, in which the state feedback gains are obtained by solving the linear matrix inequalities (LMIs). Finally, numerical simulations are performed for the considered application.

An application of the Adaptive Tabu Search (ATS), an intelligent search method in industrial control domain, is presented. The ATS is used to search for the optimum controller’s parameters denoted as proportional, integral, and derivative gains. The obtained controllers are tested against some hard-to-be-controlled plants. The results obtained are very satisfactory.