No description available.

One of the aspects of Computer Integrated Manufacturing is the ability to provide current work instructions to the operator at their workstation on a terminal or graphic monitor. The problem is, today, paperless planning is displayed to the operator one page at a time, making the operator report the completion or not completion of that task before showing the next page. This allows the operator no freedom of choice in how to do the job, which leads to reduced productivity and quality. A possible solution is presented by structuring the planning instructions and enhancing the workstation. The operator can be allowed to do the work the way he or she feels the most comfortable and yet, ensure the work is done according to design requirements.

A low variance pulse time-of-arrival (TOA) estimator is considered. It should be insensitive to all non TOA parameters and operate without a priori information about any pulse parameters assuming they are constrained to a wide range. The estimator is driven by a wide-band input diode detector with a narrow filtered base-band output. The double differentiator TOA estimator is selected. It marks inflection points of pulses and noise. An event qualifier is necessary to distinguish the pulse-only TOAs and is synthesized using the unique polarities of pulses and their derivative by detecting level crossings on both. It experiences errors of mis-detection and false alarm with noise present. Error performance is established at constant false alarm rate (FAR), set by choice of threshold level pairs. Detection probability is maximized by simulation over the locus of constant FAR levels. Design information for operating a qualifier is provided.

A system of self maintenance for fault detection and correction in a highly automated system is presented in this thesis. The self maintenance scheme consists of three parts: a watchdog monitor for fault detection, a diagnoser to pinpoint the exact component failure, and a method of repair. The second part of the scheme, the diagnoser, is developed in detail and applied to the teleoperated THAW telescope. The THAW telescope is an Earth-bound prerunner of the Astrometric Telescope Facility (ATF), an approved payload of the forthcoming International Space Station Freedom (SSF). The simulation of the ATF is modified to produce permanent hardware failures. A diagnoser is created using shallow reasoning and a rule base which pinpoints the most likely failure(s) with reasonable success. Implementation of the watchdog monitor and the repair system is left for future work.

This thesis presents a method to reconfigure objects in an articulated robotic hand, where reconfiguration is defined as moving an object from one orientation in the hand to another orientation. Several methods using object configuration rely on a stationary work area. These methods assume that the object is stably grasped by the hand, and is moved around the work area in an assembly process. This thesis assumes that the hand is the primary obstacle. The object now reacts to the hand instead of the work area. Reconfiguration of the object within the hand is the goal. Models were developed that portrayed the mechanics and the contacts for frictionless manipulation in a plane. Contact formation trees used in conjunction with the simulated mechanics is a unique method for solving this problem; the solution we attain is an optimal reconfiguration plan. (Abstract shortened with permission of author.)

Flexible manipulators provide significant advantages over the commonly-used rigid robots due to their lightweight properties, but an accurate control of these manipulators is more difficult to attain, and it is especially demanding in task executions involving changing payloads. This thesis addresses the problem of payload adaptive control of flexible manipulators. The nonlinear model describing the manipulator dynamics is completely derived and is then used for an accurate computer simulation of the flexible manipulator motions. Payload identification is implemented by using a novel neural network approach to identify distinct payload classes from tip deflection patterns which result from different payloads. The identification procedure is then used to select a controller which best meets the control objectives specifying hub speed and maximum tip deflection. Two distinct controller synthesis procedures, one using a pole-placement design and one employing a variable structure technique, are developed. The merits of payload adaptive control are shown by several simulation experiments.

This thesis introduces the problem of knowledge acquisition in developing a Trouble Shooting Guide (TSG) for equipment used in integrated circuit manufacturing. TSG is considered as a first step in developing an Expert Diagnostic System (EDS). The research is focused on the acquisition and refinement of actual knowledge from the manufacturing domain, and a Hierarchical Data Collection (HDC) system is introduced to solve the problem of bottleneck in developing EDS. An integrated circuit manufacturing environment is introduced, and issues relating to the collection and assessment of knowledge concerning the performance of the machine park are discussed. Raw data about equipment used in manufacturing environment is studied and results are discussed. A systematic classification of symptoms, failures, and repair activities is presented.

Stochastic activity networks (SANs) have been used in the modeling of computer systems because of their suitability in representing distributed systems. SANs may be solved by analysis or by simulation. When simulation is used, future events list management is very time consuming, as with other simulation techniques. New methods that take advantage of the SAN model structure are presented which significantly reduce the cost of future events list management. Multiple future events lists are used to reduce operations required upon each state change.

<p> In a large-scale outdoor, airborne, hazardous materials (HAZMAT) incident, such as ruptured chlorine rail cars during a train derailment, the local Incident Commanders and HAZMAT emergency responders must obtain accurate information quickly to assess the situation and act promptly and appropriately. HAZMAT responders must have a clear understanding of key information and how to integrate it into timely and effective decisions for action planning. This study examined the use of HAZMAT plume modeling as a decision support tool during incident action planning in this type of extreme HAZMAT incident. The concept of Situation Awareness as presented by Endsley's dynamic situation awareness model contains three levels: perception, comprehension and projection. It was used to examine the actions of incident managers related to adequate data acquisition, current situational understanding, and accurate situation projection. Scientists and engineers have created software to simulate and predict HAZMAT plume behavior, the projected hazard impact areas, and the associated health effects. Incorporating the use of HAZMAT plume projection modeling into an Incident Action Plan may be a complex process. The present analysis employed a mixed qualitative and quantitative methodological approach and examined the use and limitations of a "HAZMAT Plume Modeling Cycle" process that can be integrated into the incident action planning cycle. HAZMAT response experts were interviewed using a computer-based simulation. One of the research conclusions indicated the "HAZMAT Plume Modeling Cycle" is a critical function so that an individual/team can be tasked with continually updating the hazard plume model with evolving data, promoting more accurate situation awareness.</p>

Markovian and self-similar data traffic models are currently being proposed as accurate characterizations of the way data traffic behaves on an asynchronous transfer mode (ATM) network. Because of the dearth of traffic data from real ATM networks, researchers in this area have relied on simulated data and data from Ethernet networks. Ethernet and ATM are both packet technologies and share some similarities. There are, however, important differences between the two technologies which largely preclude the use of Ethernet traffic data in studies of ATM network traffic behavior. We propose a model for ATM data traffic based on analysis of traffic traces from two real ATM networks. This analysis reveals that the interarrival time distribution for data traffic is bimodal, in a similar fashion to Ethernet, but is a mostly lognormal mixture distribution with a very small self-similar or pseudo self-similar portion in the right tail / acase@tulane.edu