Alok Barua

Artificial neural-network model-based observers

Describes a pseudorandom testing scheme for fault diagnosis of analog integrated circuits. The goal is to implement a BIST technique with both a built-in pattern generator and a response analyzer for fault diagnosis. We have chosen a diagnostic framework for the analog ICs using a pseudorandom noise generator as the test-pattern generator and a model-based observer to detect and diagnose faults. The observer is implemented through a multilayer feedforward ANN trained with a back-error propagation (BEP) algorithm. Both the test-pattern generator and the model-based observer proposed in this article can be implemented either on- or offline depending on the need of the application and silicon area overhead.

A pseudo-random testing scheme for analog integrated circuits using artificial neural network model-based observers

This paper presents a simple and very efficient testing strategy for fault diagnosis of analog integrated circuits. The methodology is based on a technique of using a pseudorandom noise as the test pattern and a model-based observer for fast and robust testing and fault diagnosis. By incorporating device-level faults the efficiency is illustrated for stand-alone as well as embedded operational amplifiers as examples. The simulation results obtained are very encouraging. The technique can be viewed as a built-in self-test along with a design for testability scheme that dramatically improves the fault coverage and can be implemented for both on-line and off-line depending on the need of the application and silicon area overhead. Its main advantages are: a universal input stimulus (white noise) is used and thus test generation can be avoided, good and faulty signatures for high quality testing can be easily constructed and testing cost can be minimized, the technique is very efficient and robust, and the scheme can be well suited for built-in self-test implementation.

A Code Width Built-In-Self Test Circuit for 8-bit Pipelined ADC

This paper presents a novel built-in-self-test (BIST) scheme based on code-width and sample difference. The proposed BIST scheme is applied on 8-bit pipelined ADC (Analog to Digital Converter). An 8-bit pipelined ADC is designed. This pipelined ADC uses analog preprocessing to divide the input signal range into sub-intervals and amplification of a residue signal for further processing in the subsequent stages. The realization of the preprocessing stages has been implemented using switched-capacitor circuits. The proposed BIST scheme is verified by simulation of 8 bit pipelined ADC with arbitrary faults. The proposed method is alternative to histogram based analysis techniques to provide test time improvements. In addition to the measurement of DNL (Differential Non Linearity) and INL (Integral Non Linearity), non monotonic behavior and missing code fault have been detected.

TRANSELEX: A knowledge-based approach for the selection of transducers using a dynamic database

Abstract TRANSELEX (TRANsducer SELection EXpert) is a single expert system for selection of transducers in the areas of temperature, flow and pressure measurement to meet the requirements of the user. It allows numerical ranges to be inputted for greater accuracy of the selection process. At present the knowledge base of TRANSELEX has 44 sensors and 55 rules. This paper provides an overview of the expert program.

Feature Detection of an Object by Image Fusion

In this paper, we propose a novel method for feature detection of an object by fusion of range and intensity images. For this purpose, we have developed a data acquisition system with a laser source and camera interfaced with Silicon Graphics machine. 3-D mesh representation of the surface of the object is obtained from the acquired range images. Extraction of structural features from the range images has been performed by two methods i.e. coordinate thresholding technique and Laplacian of Gaussian (LoG) edge detector. Extraction of structural features from the intensity image of the object has been performed by the Hough transform technique and Canny edge detector. An approach using shape signatures has been proposed to detect corner points in the edge maps obtained using LoG detector as well as Canny detector. The extracted 3-D edge maps as well as the detected corner points have been mapped to 2-D plane. The methodology for manual fusion of edge maps with the help of affine transformation and the concept of automatic fusion of edge maps by affine transformation followed by iterative closest point (ICP) algorithm have been introduced in this work. The automated technique for fusion overcomes the drawbacks associated with manual fusion. The fusion algorithm provides a composite image with more accurate and reliable information about the important features of the object.

EXSHOF: An artificial intelligence approach to high-order filter synthesis

Abstract A two-tier expert-system-based synthesis of high-order filters is developed. Expert-system techniques have been used for choosing a particular biquad structure and for selection of the high-order filter structure as well. The data inputs will be either filter specifications or transfer functions and the output will be a detailed filter circuit with all its element values. EXSHOF ( Ex pert System based S ynthesis of H igh O rder F ilter) is totally menu-driven and highly interactive. It assumes little knowledge of computers or filters from the user. The complete package is implemented in Turbo PASCAL and Turbo PROLOG languages.

Feature level fusion of range and intensity images of an object

This paper presents feature level fusion of images of an object obtained from two different imaging techniques. A draft 3-D view of the object is generated by registering range images obtained from scanning the object by a laser beam. For that purpose, the method of optical triangulation is employed for data acquisition and feature matching technique for the registration of the images. Triangle-based linear interpolation is used to obtain a 3-D mesh representing the object surface. Structural features are extracted from the draft 3-D model of the object and subsequently mapped to 2-D plane by perspective transformation. Hough transform technique has been implemented for the extraction of structural features from the intensity image of the object obtained under illumination light. Fusion of both the extracted edge maps is accomplished by affine transformation after identifying a set of control points. This provides a high-level description of the object.

Novel enhanced pole selectivity bandpass filter with large dynamic range

A high pole selectivity (Q/sub p/) bandpass filter realizable in MOSFET-C topology is suggested. The op-amps are used in fully balanced topology which has good power supply noise rejection. The proposed filter has low op-amp count in addition to a wide dynamic range. The circuit has been compared with conventional enhanced pole selectivity bandpass filter. The circuit is simulated by SPICE and a dynamic range of 123.39 dB is obtained for Q/sub p/ value of 60.

EXSHOF-II: Active filter design, from approximation function to graphic display of the circuit

Abstract This paper presents an expert-system-based package for active filter synthesis, called EXSHOF-II. It starts with approximation function and ends with graphic display of the complete filter circuit. The package is an improved and enhanced version of its mother EXSHOF ( EX pert system based S ynthesis of H igh O rder F ilter). EXSHOF-II gives quantitative and qualitative help at various stages of filter design. Besides the Butterworth, Chebyshev and Elliptic functions which were considered in EXSHOF, the inverse Chebyshev function is also included in EXSHOF-II. For choosing particular approximation EXSHOF-II gives the order of the transfer function. The group delay characteristics of all four approximation functions are plotted on the screen by EXSHOF-II. This aids the user to make a proper choice of an approximation function. Like its mother, EXSHOF-II also considers four high-order structures. The filtering requirement in EXSHOF-II can be directly input on the attenuation characteristics, making the system more user-friendly. The circuit implementation has been done using nine different biquadratic active RC building blocks. The delay requirement can be easily satisfied by cascaded all-pass second-order networks.

CHOFIL: a knowledge based approach to active biquad synthesis

This paper describes an expert system based approach for synthesis of active biquad filter. Though expert systems have been designed for approximating filter functions, so far no effort has been made to encode the knowledge base available for the variety of biquad circuits in use. Motivated by this problem Choose Filter (CHOFIL), program in Turbo prolog is developed. The circuit element values of the particular filter structure are obtained by running corresponding synthesis programs. CHOFIL can be linked with synthesis programs which are written in FORTRAN 77.