Computer aided process planning for tape automated bonding

This paper describes the design and development of a knowledge based Computer Aided Process Planning (CAPP) system for Printed Circuit Board assembly. The domain considered uses both Surface Mount Technology (SMT) and Tape Automated Bonding (TAB) components. The structure and design of the CAPP system, and its inputs and outputs are discussed.     

Use of captions and other collateral text in understanding photographs

This research explores the interaction of textual and photographic information in image understanding. Specifically, it presents a computational model whereby textual captions are used as collateral information in the interpretation of the corresponding photographs. The final understanding of the picture and caption reflects a consolidation of the information obtained from each of the two sources and can thus be used in intelligent information retrieval tasks. The problem of building a general-purpose computer vision system withouta priori knowledge is very difficult at best. The concept of using collateral information in scene understanding has been explored in systems that use general scene context in the task of object identification. The work described here extends this notion by incorporating picture specific information. A multi-stage systemPICTION which uses captions to identify humans in an accompanying photograph is described. This provides a computationally less expensive alternative to traditional methods of face recognition. A key component of the system is the utilisation of spatial and characteristic constraints (derived from the caption) in labeling face candidates (generated by a face locator).This work was supported in part by ARPA Contract 93-F148900-000. I would like to thank William Rapaport for serving as my advisor in my doctoral work; Venu Govindaraju for his work on the face locator; and more recently, Rajiv Chopra, Debra Burhans and Toshio Morita for their work in the new implementation of PICTION as well as valuable feedback.     

Knowledge based decision support for PCB assembly using SMT

Knowledge based Decision Support Systems (DSSs) in manufacturing help enhance productivity and quality. The need for DSS use in the Printed Circuit Board (PCB) assembly using Surface Mount Technology (SMT) domain is discussed. Process and manufacturing systems related DSSs for the PCB assembly domain are delineated, and specific systems are described.     

Pyramid representations for solids

A pyramid of a solid is a sequence of representations such that each succeeding representation is of higher resolution. The paper defines methods for deriving pyramids from 3-d arrays that represent volumetric information in 3-d scenes. The methods are based on the oct-tree and binary-tree representations of solids. Algorithms for defining abstractions of a scene and for adding detail to a previous representation are presented. The concepts are illustrated by showing how a solid can be modeled at varying levels of precision.     

A knowledge-based thermal profile identification advisor for surface mount PCB assembly

Reflow soldering of solder paste is a critical process in the surface mount assembly of printed circuit boards (PCBs). Infrared (IR) radiation and/or forced convection are often used in an oven to heat the PCB assembly and the deposited solder paste to cause the solder to reflow and subsequently form the interconnection between the component terminations and the attachment pads. Control of the reflow soldering process requires the understanding and regulation of several complex variables. Understanding the interrelationships among process control parameters is important in achieving a low defect rate and high yields in reflow soldering. When a variety of components are soldered onto a printed circuit board (PCB) simultaneously in an IR/convection oven through reflow soldering, the process control variables that need to be considered include the characteristics of solder paste used, the size, type, and quantity of components placed on the PCB, the substrate used, the mass and the thermal area of the workpiece, and the conveyor speed.A prerequisite for the achievement of high process yields in the reflow soldering process is the identification of the optimal temperature profile and the range between which the profile is acceptable for the specific assembly. This information is subsequently used by the process engineer to arrive at the relevant temperature of power (wattage) settings for the oven. The in-depth process knowledge required to study the process variables and deduce thermal profiles is not commonly available. Consequently, the objective of this research was to use an expert system approach to help a manufacturing engineer with thermal profile identification in surface mount PCB assembly.The system developed considers the major process control parameters and the interrelationships among them to deduce the optimal temperature profile and a range in which the profile is considered acceptable. The profile identification expert system was developed using PROLOG on a personal computer. Results are displayed using a combination of graphical and textual output. The system was designed to function in a stand-alone capacity. The profile identification expert system's architecture integrates an inference mechanism with the knowledge (facts and rules) provided by the domain expert. It allows the user to update or delete the facts and rules stored in the knowledge base. Consequently, it provides an avenue to enhance the capacity of the system to incorporate domain specific changes.     

Online and off-line handwriting recognition: a comprehensive survey

Handwriting has continued to persist as a means of communication and recording information in day-to-day life even with the introduction of new technologies. Given its ubiquity in human transactions, machine recognition of handwriting has practical significance, as in reading handwritten notes in a PDA, in postal addresses on envelopes, in amounts in bank checks, in handwritten fields in forms, etc. This overview describes the nature of handwritten language, how it is transduced into electronic data, and the basic concepts behind written language recognition algorithms. Both the online case (which pertains to the availability of trajectory data during writing) and the off-line case (which pertains to scanned images) are considered. Algorithms for preprocessing, character and word recognition, and performance with practical systems are indicated. Other fields of application, like signature verification, writer authentification, handwriting learning tools are also considered     

Working Digital Money into a Cash Economy: The Collaborative Work of Loan Payment

This paper examines how different forms of money, specifically digital versus cash, impact on the work of an organisation and its customers. In doing so it contributes to the body of literature exploring how the social meanings of money impact on practice. We describe the findings of an ethnographic study examining loan collection workflows, where bank loans given to auto-rickshaw drivers to purchase their auto-rickshaw are overseen and managed by an intermediary. We found that making the mobile money service usable for the drivers took considerable work and was largely achieved because it was embedded in the wider, trusted, loan payment ecosystem. Although Airtel Money promises anytime, anywhere payments, payment remains time and place bound for the drivers. It is tempting to take a transactional approach to payments, and indeed we initially approached the problem of enabling frequent payments as one of payment mechanism. However, in practice payments are embedded in sets of social relations and a socio-technical ecosystem. It takes considerable collaborative work, and a fair amount of flexibility, to enable these financially vulnerable drivers to pay off their loans, and reducing the issue to one of payment mechanism alone does not tell the full story.     

The dependence condition graph: Precise conditions for dependence between program points

The problem of identifying the conditions under which semantic or behavioural dependences arise between different program statements has interesting applications in various areas such as program understanding, software maintenance, software audits and software testing. We present an extension to the program dependence graph (PDG), called the dependence condition graph (DCG), that enables identifying the conditions for dependence between program points. We show that these conditions are not only correct with respect to the program's semantics, but also more precise than identified by other known techniques. We also present evidence that the DCG is a practical representation and can be built for large programs, and sketch many different applications of the DCG.     

Recursive implementation of a two-step nonparametric decision rule

The two-step approach to nonparametric discrimination is that of estimating class-conditional densities and deriving the Bayes decision rule as if the estimates were true. Direct implementation of such a decision rule ecounters two computational problems. Complexity increases with sample size, and finite precision limits the decision rule domain. Here a recursive algorithm to reduce the expected number of operations and word-length limitations below that of the direct approach is developed. A special case of the formulation reduces to the weighted k-nearest-neighbor rule.