A. Liberatore

SAPEC-a personal computer program for the symbolic analysis of electric circuits

A personal computer program for the symbolic analysis of lumped, linear, time-invariant networks is presented. The program, written in Microsoft Lisp, utilizes methods of symbolic manipulation of algebraic expressions. The proposed approach is based on modified nodal analysis and an algorithm developed by the authors for the completely symbolic solution of the related linear systems. A useful property of the program is that it allows the user to select the circuit components which he wishes to consider as symbolic parameters. In order words, with SAPEC, it is possible to obtain the network functions in complete or partial symbolic form.<<ETX>>

Analog network testability measurement: a symbolic formulation approach

A symbolic formulation approach is applied to the problem of computing testability features of analog linear networks. The program, SAPTES, obtained by following this approach is presented. The program can be a very useful tool for designers and researchers in the field of linear analog circuits. SAPTES, which is written in LISP and runs on MS-DOS personal computers, is able to compute the testability of linear circuits of rather high complexity (composed of tens of components and nodes). Computational times range from a few tens of seconds to some tens of minutes on medium speed computers. The program is easily transportable to workstations or a mainframe, and, for the mainframe, program performance will considerably increase. >

On absolute stability of neural networks

The aim of this paper is to discuss the role of Absolute Stability (ABST) in the design of neural optimization solvers and to find necessary and sufficient conditions for ABST for some classes of neural networks of applicative interest. By ABST it is meant that there is a unique equilibrium point attracting all trajectories of motion and that this property is valid for all neuron activation functions belonging to a specified class of nonlinear mappings and for all constant neural network inputs. ABST neural networks are best suited for solving optimization problems being devoid of spurious suboptimal responses for every choice of the activation function and of the input vector. A necessary and sufficient condition for ABST has been found for symmetric neural networks of the Hopfield type. In this paper, we show that the concept of ABST can be applied also to special classes of nonsymmetric Hopfield neural networks and to neural models different from the Hopfield one. It is shown in particular that necessary and sufficient conditions for ABST can be found for two interesting classes of nonsymmetric networks, namely, cooperative Hopfield-type networks and composite neural networks with variable and constraint neurons used for solving linear and quadratic programming problems in real time.<<ETX>>

Multifrequency measurement of testability with application to large linear analog systems

With increasing electronic circuit complexity, assessing the testability features becomes a necessity during the design, implementation, and operational or maintenance phases of an analog system. A quantitative measure of testability, based on several multifrequency stimuli, is adopted which is able to handle multiple faults and may provide information also on the degree of complexity encountered in a specific test. An efficient and practical algorithm is proposed which is associated with the result of Sen and Saeks and has a well-defined interpretation even with a large number of circuit parameters liable to failure. The described technique is a basis for optimizing the number and allocation of the selected test points; furthermore, it may serve as an aid in functional partitioning of the same system to facilitate testing and/or reduce computational complexity. An application to a classical active filter is also given.

A new efficient method for analog circuit testability measurement

Testing an electronic circuit is a vital part of its overall design and fabrication process. This problem is even going to be more critical with technology improvements and with the coexistence on a chip of analog and digital components. In fact the testing of mixed signal microsystems is very difficult compared to digital only devices. They do not lend themselves to earlier and simpler test routines. The entire mixed signal segment is hampered by the lack of design for testability methodologies and tools. In this field the concept of analog circuit testability is of fundamental importance. The aim of this work is to present a new symbolic approach for testability measurement of analog networks. The new method presents noteworthy advantages from a computational point of view with respect to previous techniques. In fact it does not require the computation of the sensitivities of the network functions but it is based only on the study of the network function symbolic coefficients. The new approach allows also the formulation of simple necessary conditions for maximum testability based only on the order of the network functions.<<ETX>>

A new symbolic program package for the interactive design of analog circuits

A new program package which constitutes an environment for the interactive exploration and improvement of analog circuit topologies is presented in this paper. The environment is provided with functionalities which permit the graphical schematic entry of the circuit, the symbolic analysis, the approximation of the symbolic results, the use of an external numerical simulator and the graphical postprocessing of both the symbolic and numerical simulation results. These functionalities allow us to immediately evaluate the influence of both topology and component value changes on the circuit behavior. The result is useful for educational/training purposes and for the interactive synthesis of new high-performance analog circuits.

Network sensitivity analysis via symbolic formulation

A personal computer program for the sensitivity analysis of linear networks is presented. The program, which is part of an integrated program package for the symbolic analysis of electric circuits, is based on artificial intelligence techniques and, in particular, on symbolic manipulation of algebraic expressions. The program obtains the network sensitivity functions in completely symbolic form with respect to the complex frequency and in completely (or partially, if requested) symbolic form with respect to the circuit components.<<ETX>>

Automatic test point selection for linear analog network fault diagnosis

A program for the selection of test points in linear analog networks is presented. The program, written in PROLOG, is based on a knowledge base constituted by simple rules derived from experience and heuristic reasoning. It utilizes a program previously developed by the authors which is able to compute network testability, having specified a set of test points. Illustrative examples are given.<<ETX>>

A symbolic approach to the fault location in analog circuits

The increased complexity of electronic circuits due to technological improvement has caused the need of always more sophisticated testing and fault diagnosis methodologies. However, while for digital systems these methodologies have now reached full automation, for analog and mixed signal systems there is a lack of efficient and simple methods in this field. The aim of this work is to present a completely new methodology for the parametric fault diagnosis of linear analog circuits. The new method, which is based on the k-fault diagnosis hypothesis and takes into account tolerances and measurement errors, has been fully automated. The automatic fault diagnosis system has been developed by exploiting symbolic techniques, which permit a significant reduction in the computational complexity.

Global asymptotic stability for a class of nonsymmetric neural networks

The authors show that the property of global asymptotic stability is guaranteed for a class of neural circuits with a special form of nonsymmetric interconnection matrix. They also show that neural networks used to solve typical optimization problems such as linear and quadratic programming problems fall into the class of circuits studied here and are characterized by a unique globally asymptotically stable equilibrium. The results are proved by means of the Lyapunov method and by finding suitable Lyapunov functions that are valid for special classes of neural networks with nonsymmetric interconnection matrices as described.<<ETX>>