In many pattern-classification and recognition problems, separation of different swarms of class representatives is necessary. As well, in function-approximation problems, neurons with a local area of influence have demonstrated measurable success. In our previous work, we have shown how intrinsic quadratic characteristics of traditional metal-oxide-semiconductor (MOS) devices can be used to implement hyperspherical discriminating surfaces in hardware-implemented neurons. In this work, we further extend the concept from quadratic forms to more-arbitrary closed-boundary shapes. Accordingly, we demonstrate how intrinsic characteristics of submicron MOS devices can be utilized to implement efficient pattern discriminators for various applications and, through representative simulations, show their success in some typical function-approximation problems. Further, we offer two mathematical interpretations of possible roles for these networks: Geometrically, we show that our networks employ closed hypercone shapes as their discriminating surfaces; analytically, we show that a set of these synapses connected to a common integrating body calculates the distance between their inputs and weight vectors using a power norm. The feasibility of the idea is practically investigated by design, implementation, and test of a three-dimensional (3-D) closed-boundary pattern classifier, fabricated in 0.35-microm complimentary MOS, whose results are reflected in this work.

In this paper a new family of low-power dynamic logic called data-driven dynamic logic (D/sup 3/L) is introduced. In this logic family, the synchronization clock has been eliminated, and correct sequencing is maintained by appropriate use of data instances. Then, it is shown that replacement of the clock with input data implies less power dissipation without speed degradation compared to conventional dynamic logic.

An electrooptical proximity sensor capable of measuring the distance and two-dimensional orientation of an object's surface is presented. The robustness of the sensor, targeted for utilization in robotic active sensing, is achieved via the development of a novel amplitude-modulated-based electrooptical transducer, an electronic-interface circuit that provides very good noise immunity and a wide dynamic operating range, and an effective multi-region calibration process that significantly improves pose-estimations at near proximities. An experimental setup was designed and implemented for the development and verification of the proposed proximity sensor in a simulated robotic environment. Experimental results using a variety of calibrated surfaces and materials are presented and discussed. It is shown that average accuracies of 0.01 mm and 0.03/spl deg/ can be achieved. The robustness of the proximity sensor is also verified for potential use in grasping objects with a priori noncalibrated surfaces.

A heuristically derived stabilization strategy for an unstable and unintuitive plant by fuzzy control is described. It is shown that the often used classical fuzzy controller, which is both static and time invariant, is incapable of stabilizing such types of plants. However, a simple modification to the classical fuzzy controller architecture that separates the measurement and control phases, together with a hierarchical control strategy, enable the unstable and unintuitive plant to be stabilized. The fuzzy control strategy, as well as the new fuzzy controller architecture, are based on the consideration of "what a human subject would do when dealing with a physical plant which is both unstable and unintuitive". The stabilization strategy is then generalized to other mathematically similar systems. While the rules for the stabilization of the plant are heuristically defined, the membership functions associated with the rules are tuned by a simulated annealing procedure.

Because of their nondeterministic nature of behaviour, the supervisory control of manufacturing systems must be carried out in closed loop. This trait greatly increases the size and complexity of the discrete-event system (DES)-based supervisory-controllers of manufacturing systems. In an attempt to cope with this complexity, a hybrid supervisory controller (HSC) is proposed. It splits operations between its three elements: (i) a DES supervisory controller, which contains the nominal control strategy, (ii) a diagnostic system, which monitors the workcell, and (iii) an alternate strategy driver (ASD), that asserts control by revising the nominal strategy of the DES supervisory controller, whenever events diverge from the states of the DES supervisory controller. This modified approach to DES control provides a more efficient controller than could reasonably be attained using solely a DES supervisory controller.>

An approach to uncertainty compensation using a multilayer feedforward neural network in multi-manipulator system control is proposed. The proposed approach is developed by formulating the dynamics of the multi-manipulator system in the constrained motion framework. The error-backpropagation algorithm is employed for neural network learning. The teaching signal for neural network learning is derived by analyzing the stability of the closed-loop system. It is shown that if the neural network learns to generate the proper compensating signal, then the constrained motion of the multi-manipulator system tracks the desired motion asymptotically; as a consequence, the desired forces can be achieved. Computer simulations are conducted to verify the proposed approach.>

A new 3-D object recognition method for robotic assembly workcells is presented. The proposed method is focused on two basic concepts, i.e., active vision and object preconditioning. The main aspects of the proposed system are presented.>

The development of a robust fiber-optic proximity sensor is presented. A method is proposed by which to solve the problem of robustness of amplitude- and phase-modulated electrooptical proximity sensors to variations in surface-reflection characteristics, and to reduce the influence of distance and orientation measurements on one other. The method comprises three different methodologies, which are combined for better efficiency: integration of distance and orientation sensors, a novel polarization-based optical-filtering approach, and active sensing. Emphasis is given to unique solutions for background-noise interference and dynamic-range-limitation problems.

In this paper, the dynamic behavior of a three-layer feedforward neural network as a uncertainty compensator for robotic control is investigated. The investigation is conducted in the context of the robot trajectory tracking problem, where the neural network (with the error-backpropagation algorithm) is used as a uncertainty compensator in conjunction with the feedback linearization control (i.e. computed torque) and a PD control. Through computer simulation, it is verified that the dynamics of the neural network has a specific pattern when the learning rate is sufficiently small, and that such a specific pattern of weight variation in the neural network represents a sufficient condition for closed-loop system performance improvement.

Estimation of 3D information from 2D-image coordinates is a fundamental problem in both machine vision and computer vision. A closed-form analytical solution is presented to the problem of 3D-location estimation of quadratic-curved features based on two images. Compared to previous solution methods, the proposed method has the advantages that it is a closed-form solution method, it is mathematically simpler, and a priori knowledge of the size and shape of the feature is not required. The method is demonstrated for an elliptical feature, using a simulated experimental set-up.>

In this paper, a new 2D-Shape-encoding scheme is introduced which is based on the idea of the Angle-Of-Sight (AOS). Using this scheme, a shape can be efficiently transformed into a 1D signature by recording the AOS-vs-distance of each boundary point with'respect to a shape-specific Chord-Of-Sight (COS). The COS is selected by using an extension of the notion of shape boundary, to the idea of shape-specific points and the characteristic ellipse (CE). The AOS signature has many important properties including: It is information-preserving, and thus unique; It does not require boundary smoothing; It has its own select- able smoothing property; It can provide a set of multi-scale representations by means of a simple operation; It is transformation-invariant; It is defined at all points; It preserves symmetries. As well, for matching purposes, a two-level matching process is proposed using a global meas- ure (the eccentricity of the CE of a shape) and a dissimilarity measure based on the AOS signature. The encoding and matching techniques developed have been tested %with 35 manufactured objects. The results obtained show that the AOS signature and the two-level-matching technique are quite effective and reliable for the recognition of 2D-shapes of typical manufactured objects.

The authors address the design and performance analysis of partial-multiple-bus interconnection networks. They are bus architectures that have evolved from the multiple-bus structure by dividing buses into groups and reducing bus connections. Their effect is to reduce cost and alleviate arbitration and drive requirements without degrading performance significantly. One such structure, called processor-oriented partial-multiple-bus (PPMB), is proposed. It serves as an alternative to the conventional structure called memory-oriented partial-multiple-bus (MPMB) and is aimed at higher system performance at less or equal system cost. It has been shown, both analytically and by simulation, that a substantial increase in system bandwidth (up to 20%) is achieved by the PPMB structure over the MPMB structure. With very large systems, the results also imply a significantly improved cost-effectiveness over the conventional multiple-bus architecture.>

An imager with built-in image-velocity computation capability is described. The image-velocity computation technique is based on signals propagating on delay lines. Silicon implementation using 3- mu m-CMOS technology is described. Experimental results show that a computational error of less than 20% can be achieved using available fabrication technology. This figure can be reduced by using larger arrays and better implementations of delay gates.>

A technique for image-velocity computation is described. Using one set of intersecting delay lines that are physically perpendicular to each other, the two perpendicular components of image velocity can be computed separately if the image area is known. To compute the image velocity without knowledge of the image area, two sets of intersecting delay lines are required. This technique is suitable for analog VLSI implementation.>

An architecture for the efficient and high-speed realization of morphological filters is presented. Since morphological filtering can be described in terms of erosion and dilation, two basic building units performing these functions are required for the realization of any morphological filter. Dual architectures for erosion and dilation are proposed and their operations are described. Their structure, similar to the systolic array architecture as used in the implementation of linear digital filters, is highly modular and suitable for efficient very-large-scale integration (VLSI) implementation. A decomposition scheme is proposed to facilitate the implementation of two-dimensional morphological filters based on one-dimensional structuring elements constructed using the dual architectures. The proposed architectures, which also allow the processing of gray-scale images, are appropriate for applications where speed, size, and cost are of critical significance.>

A methodology based on mathematical morphology is proposed for efficient recognition of two-dimensional (2D) objects or shapes. It is based on the introduction a shape descriptor called the morphological autocorrelation transform (MAT). The MAT of an image is composed of a family of geometrical correlation functions (GCFs) which define its morphological covariance in a specific direction. The MAT is translation-, scale-, and rotation-invariant. It is shown that in most situations, a small subset of the MAT suffices for image representation. The characteristics and performance of a shape recognition system based on the MAT are investigated and analyzed. Computational complexity of the proposed morphological-based recognition system is examined. It is shown that shape properties, such as area, perimeter, and orientation, are readily derived from the MAT representation, and that the proposed system is well suited for shape representation and classification.

A closed-form analytical solution to the problem of 3D estimation of circular-feature location is presented. Two different cases are considered: 3D orientation and position estimation when the radius is known and when it is not known. Extension of the method to general 3D quadratic features is also addressed. Simulated experimental results obtained for all three cases verified the analytical method. In the case of real experiments, a set of circles located on a calibration plate, whose locations were known with respect to a reference frame, were used for camera calibration as well as for the application of the method. A sequential compensation procedure was applied to the input gray-level image to compensate for distortion. These results also showed the validity of the process and the applicability of the analytical method.>

The problem of accurate parameter estimation of elliptical shapes is addressed. Three methods based on elliptical Fourier descriptors, area moments, and perimeter moments of 2-D elliptical shapes are presented. The mathematics of these methods is discussed and, for area moments, a set of new formulas based on Gauss's theorem is given. To implement a comparative performance study, a method developed by R. Safaee-Rad et al. (1990), the weighted minimum-square-error (MSE) function, is included, and an objective and independent measure of goodness of fit is used. The performance of these methods is compared for two different cases, that of a perfect simulated ellipse and an imperfect imaged ellipse.>

A new 2D shape-encoding scheme is introduced which is based on the idea of the angle-of-sight. The shape can be efficiently transformed into a 1D signature by recording the angle-of-sight vs. distance of each boundary point with respect to a shape-specific chord-of-sight. The chord-of-sight is selected by using an extension of the motion of shape to geometrical entities, which are not part of the shape boundary-to the idea of shape-specific points and the characteristic ellipse. The angle-of-sight signature is a unique, information-preserving, transformation-invariant shape representation. It is demonstrated that the angle-of-sight shape-encoding scheme is capable of deriving well-behaved noise-tolerant signals, convenient for shape matching through alignment of signatures.>

A new descriptor for representing two-dimensional continuous or discrete signals is introduced. The proposed shape descriptor, which is called the geometrical correlation function (GCF), is based on the principle of mathematical morphology. The properties of this shape descriptor are examined. It is shown that the family of GCFs associated with different orientations of a particular shape is translation, scale, and rotation invariant. Geometrical properties such as the area and perimeter of the shape can be derived from the GCF family. The utilization of the GCFs for shape recognition is considered. The GCF family can be computed using the associated morphological correlator which is composed of m parallel computation units and a feature-function selection unit where a small subset of the GCF family is selected for classification. It is shown that with a suitable criterion for selecting the feature function, promising results for successful classification are obtained.>

The main sources of errors that cause distortion in the shape of an ellipse are presented, and methods for compensation of these distortion factors are formulated. A method for estimation of ellipse center coordinates is proposed. A method for the determination of an optimal elliptical fit to a set of input-edge-points is presented. An error function, based on a new geometrical interpretation of an existing error function, for estimation of all five parameters of an ellipse is defined, and an objective and independent measure for the goodness of various fits is given. The proposed error function and two other previously developed error functions are applied to six different situations. The results of the comparative study show that when all the edge points are used for the fit, the performance of the proposed error function is better than the other two.>

Two premarking methods are proposed for a new 3D object recognition system under development at the University of Toronto. In this system, an object is modeled using only a small number of 2D distinct perspective views (standard views) predefined wit the help of markers placed on the object. During the recognition process, a standard view is acquired by first determining its surface normal (standard-view axis), and then aligning the camera's optical axis with it. Standard-view axes are obtained by analyzing the images of the markers. A morphological skeleton transform (MST) is used for the extraction of required marker features. This work presents the analytical solution for the two proposed premarking schemes, based on circular markers, that can be used in acquiring standard views of objects. Specific issues addressed include: the determination of the perspective distortion and its relative importance, the determination of the transformation parameters required for camera alignment, and the use of a class of MST, pseudo-Euclidean skeletons, for feature extraction.>

In a previous paper, we proposed a special environment and techniques for the implementation of highly parallel processing. A new VLSI supercomputer architecture, DYPP (DYnamically Programmable multi-Processor), was introduced, with the unique property of being able to embed, and execute directly, program graphs both statically and dynamically. Either asynchronous (dataflow) or synchronous implementations of program graphs can be realized in DYPP. In the present paper, we provide further insight in its operation.

The design and performance analysis of partial-multiple-bus interconnection networks is described. One such structure, called processor-oriented partial-multiple-bus (or PPMB), is proposed. It serves as an alternative to the conventional structure called memory-oriented partial-multiple-bus (or MPMB) and is aimed at higher system performance at less or equal system cost. PPMB's structural feature, which distinguishes itself from the conventional, is to provide every memory module with B paths to processors (where B is the total number of buses). This, in contrast to the B/g paths provided in the conventional MPMB structure (where g is the number of groups), suggests a potential for higher system bandwidth. This potential is fully fulfilled by the load-balancing arbitration mechanism suggested, which in turn highlights the advantages of the proposed structure. As a result, it has been shown, both analytically and by simulation, that a substantial increase in system bandwidth (up to 20%) is achieved by the PPMB structure over the MPMB structure. In addition to the fact that the cost of PPMB is less than, or equal to, that of MPMB, its reliability is shown to be slightly increased.>

In this correspondence, a new scheme is proposed in which ternary clocking signals are used to replace binary clocking signals in VLSI scan-testing designs. This scheme has the same advantage of high testability as the binary scan method [1], but it eliminates the mode- selecting signal line. Since this mode-selecting line must be routed to each flip-flop in the binary scan scheme, the saving is significant in reducing the circuit interconnection complexity and chip area. This correspondence describes the new ternary scheme in detail, and also suggests appropriate circuit designs using CMOS technology. Furthermore, comparisons are made between ternary scan and binary scan [3] and between ternary scan and a scan scheme using binary with a local decoder [2].

A prototype touch-sensitive tablet is presented. The tablet's main innovation is that it is capable of sensing more than one point of contact at a time. In addition to being able to provide position coordinates, the tablet also gives a measure of degree of contact, independently for each point of contact. In order to enable multi-touch sensing, the tablet surface is divided into a grid of discrete points. The points are scanned using a recursive area subdivision algorithm. In order to minimize the resolution lost due to the discrete nature of the grid, a novel interpolation scheme has been developed. Finally, the paper briefly discusses how multi-touch sensing, interpolation, and degree of contact sensing can be combined to expand our vocabulary in human-computer interaction.

A UIMS developed at the University of Toronto is presented. The system has two main components. The first is a set of tools to support the design and implementation of interactive graphics programs. The second is a run-time support package which handles interactions between the system and the user (things such as hit detection, event detection, screen updates, and procedure invocation), and provides facilities for logging user interactions for later protocol analysis. The design / implementation tool is a preprocessor, called MENULAY, which permits the applications programmer to use interactive graphics techniques to design graphics menus and their functionality. The output of this preprocessor is high- level code which can be compiled with application-specific routines. User interactions with the resulting executable module are then handled by the run-time support package. The presentation works through an example from design to execution in a step-by-step manner

This paper describes a versatile human-machine interface designed to achieve an important balance between human factors and flexibility. It incorporates attributes of keypads, tablets and 'mice' and possesses freedom of keyshape, size and function. The system uses a TV camera focused on a 'keyboard' area to generate electrical signals in response to optical inputs. Additional optical and electronic feature extraction result in an input device capable of processing discrete and continuous input simultaneously.

Advances in multiple-valued logic (MVL) have been inspired, in large part, by advances in integrated circuit technology. Multiple-valued logic has matured to the point where four-valued logic is now part of commercially available VLSI IC's. Besides reduction in chip area, MVL offers other benefits such as the potential for circuit test. This paper describes the historical and technical background of MVL, and areas of present and future application. It is intended, as well, to serve as a tutorial for the nonspecialist.

This correspondence describes an application of COS/MOS integrated circuits in the design of ternary rate-multipliers. The logical implementation of two types of ternary rate-multiplier is described. The first type produces an unevenly spaced output pulse train, while in the second one the output pulses are uniformly spaced. Advantages and disadvantages of the two types are presented.

RAP—a Relational Associative Processor—is a back-end or peripheral device to augment a general purpose computer for implementing a database management system (DBMS) Its architecture is based on the fact that database operations are inherently set-oriented and that data base addressing is best accomplished through associative reference to achieve high data independence. RAP utilizes these characteristics by combining the features of associative and array processors. Previous publications on RAP have dealt separately with the details of the first version of its architecture [1]-[4] language interface [5],[6] and performance evaluation [7]-[9]. This paper provides details on a recently evolved, faster, and more flexible architecture for RAP called RAP. 2 [17].

RAP - a Relational Associative Processor - is a back-end or peripheral device to augment a general purpose computer for implementing a data base management system (DBMS). Its architecture is based on the fact that data base operations are inherently set-oriented and that data base addressing is best accomplished through associative reference to achieve high data independence. RAP utilizes these characteristics by combining the features of associative and array processors. Previous publications on RAP have dealt separately with the details of the first version of its architecture [1,2,3,4] language interface [5,6] and performance evaluation [7,8,9,]. This paper provides details on a recently evolved, faster, and more flexible architecture for RAP.

The organization and functional design of a parallel radix-4 fast Fourier transform (FFT) computer for real-time signal processing of wide-band signals is introduced.

Many-valued switching systems have been of considerable academic interest, despite the apparent inability to use them in practical applications. The main drawbacks, as pointed out by a number of authors, are the difficulties associated with the implementation of the functional basic set and the lack of adequate simplification techniques.