Domenico Porto

ANN and Non-Integer Order Modeling of ABS Solenoid Valves

This brief presents an electric model of an on/off solenoid valve used to regulate brake pressure in anti-lock braking control systems (ABS). The model has been derived with the purpose to test and validate some coil driver integrated circuits in a Hardware-in-the-loop approach. The valve model is composed by magneto-dynamic, mechanical and fluid dynamic subsystems. The global system is highly nonlinear and presents hysteresis phenomena due to the ferromagnetic effect. A non-integer order system has been adopted to simply represent the high order dynamics present into the system, while Duhem equations coupled with two different neural networks are used to model the hysteresis. Several experimental trials have been performed in order to validate the suitability of the proposed approach.

Analog implementation of non integer order integrator via Field Programmable Analog Array

Abstract Recently a renewed interest has been devoted to non integer, or fractional, order systems. This is due to the fact that they well model a lot of physical systems and can be usefully applied in the area of automatic control. On drawback, mainly faced in the area of control systems, is related to their practical realization, essentially due to their infinite dimension nature. In this paper it is proposed an analog implementation of a non integer order integrator by using Field Programmable Analog Array. The reported example shows the feasibility and reliability of the proposed approach.

A neuro-fuzzy approach for hardware modeling of nuclear fusion reactor plasma

In this paper a hardware design of a behaviour predictor of a nuclear fusion plasma is presented. It is based on a fuzzy neural network modelization of the nonlinear relation existing among the six forcing magnetic fields and the geometric displacement of the high-temperature plasma. This approach fits quite well with the intrinsic nonlinearity of this phenomena, providing a good degree of approximation. Each design step is described and the final board is suitably tested in order to validate the proposed approach.

Parameter Tuning and Hardware Implementation of a Non Integer Order PID Controller

In this paper a new type of PID controllers is introduced and some tuning properties are given. The novelty of the proposed controller consists in the extension of derivation and integration order from integer to non integer numbers. This approach provides a more flexible tuning strategy and therefore an easier achieving of control requirements with respects to classical one. Some examples are also reported to validate the introduced strategy. Moreover two approaches for controller implementation are proposed. The first one is based on digital circuits while the second proposes the analog approach

A new method for matrix description of genetic algorithms

In this paper, a new method providing an efficient simulation of genetic algorithms through an alternative description of them is introduced. It is based on the representation of each step of the algorithm by means of suitable matrix operations, where each matrix depends basically from population size and from initial parameters settings. The algorithm is therefore rewritten as an autonomous nonlinear discrete system

Logical foundations for constraints on fuzzy sets in soft computing: MV-partitions and refinement

Although many authors have proposed constraints on membership functions for automatic optimization processes, formal definitions of these constraints are usually not well connected with the theory of partitions in many-valued and fuzzy logic. Our aim here is to derive these constraints from the theory of partitions in MV-logic, and to analyze the concept of refinement of a partition.