published_or_final_version / Computer Science / Master / Master of Philosophy

No description available.

Thesis (M. Phil.)--University of Hong Kong, 1990.

Tracking Systems provide an important analysis technique that can be used in many different areas of science. A Tracking System can be defined as the estimation of the dynamic state of moving objects based on `inaccurate’ measurements taken by sensors. The area encompasses a wide range of subjects, although the two most essential elements are estimation and data association. Tracking systems are applicable to relatively simple as well as more complex applications. These include air traffic control, ocean surveillance and control sonar tracking, military surveillance, missile guidance, physics particle experiments, global positioning systems and aerospace. This thesis describes an investigation into state-of-the-art tracking algorithms and distributed memory architectures (Multiple Instructions Multiple Data systems - “MIMD”) for parallel processing of tracking systems. The first algorithm investigated is the Interacting Multiple Model (IMM) which has been shown recently to be one of the most cost-effective in its class. IMM scalability is investigated for tracking single targets in a clean environment. Next, the IMM is coupled with a well-established Bayesian data association technique known as Probabilistic Data Association (PDA) to permit the tracking of a target in different clutter environments (IMMPDA). As in the previous case, IMMPDA scalability is investigated for tracking a single target in different clutter environments. In order to evaluate the effectiveness of these new parallel techniques, standard languages and parallel software systems (to provide message-passing facilities) have been used. The main objective is to demonstrate how these complex algorithms can benefit in the general case from being implemented using parallel architectures.

A Virtual Reality is a computer model of an environment, actual or imagined, presented to a user in as realistic a fashion as possible. Stereo goggles may be used to provide the user with a view of the modelled environment from within the environment, while a data-glove is used to interact with the environment. To simulate reality on a computer, the machine has to produce realistic images rapidly. Such a requirement usually necessitates expensive equipment. This thesis presents an implementation of a virtual reality system on a transputer architecture. The system is general, and is intended to provide support for the development of various virtual environments. The three main components of the system are the output device drivers, the input device drivers, and the virtual world kernel. This last component is responsible for the simulation of the virtual world. The rendering system is described in detail. Various methods for implementing the components of the graphics pipeline are discussed. These are then generalised to make use of the facilities provided by the transputer processor for parallel processing. A number of different decomposition techniques are implemented and compared. The emphasis in this section is on the speed at which the world can be rendered, and the interaction latency involved. In the best case, where almost linear speedup is obtained, a world containing over 250 polygons is rendered at 32 frames/second. The bandwidth of the transputer links is the major factor limiting speedup. A description is given of an input device driver which makes use of a powerglove. Techniques for overcoming the limitations of this device, and for interacting with the virtual world, are discussed. The virtual world kernel is designed to make extensive use of the parallel processing facilities provided by transputers. It is capable of providing support for mUltiple worlds concurrently, and for multiple users interacting with these worlds. Two applications are described that were successfully implemented using this system. The design of the system is compared with other recently developed virtual reality systems. Features that are common or advantageous in each of the systems are discussed. The system described in this thesis compares favourably, particularly in its use of parallel processors. / KMBT_223

Two important problems namely a versatile, efficient communication system and allocation of processors to processes are analysed. An efficient communication system has been developed, in which a central controller, the bus-master, dynamically configures the point-to-point network formed by the links of the transputers. The links are used to form a point-to-point network. An identical kernel resides on each of the nodes. This kernel is responsible for all communications on behalf of the user processes. It makes ConnectLink and ReleaseLink requests to the central controller and when the connections are made it sends the the messages through the connected link to the destination node. If direct connection to the destination node cannot be made then the message is sent to an intermediate node, the message hops through intermediate nodes until it reaches the destination node. The communication system developed provides low latency communication facility, and the system can easily be expanded to include a large number of transputers without increasing interprocess communication overhead by great extent. Another problem, namely the Module Assignment Problem (MAP) is an important issue at the time of development of distributed systems. MAPs are computationally intractable, i.e. the computational requirement grows with power of the number of tasks to be assigned. The load of a distributed system depends on both module execution times, and intermodule communication cost (IMC). If assignment is not done with due consideration, a module assignment can cause computer saturation. Therefore a good assignment should balance the processing load among the processors and generate minimum inter-processor communication (IPC) ( communication between modules not residing on the same processor). Since meeting the deadline constraint is the most important performance measure for RTDPS, meeting the response time is the most important criteria for module assignment. Understanding this we have devised a scheme which assigns processes to processors such that both response time constraints and periodicity constraints are met. If such an assignment is not possible, assignment would fail and an error would be generated. Our assignment algorithm does not take into consideration factors such as load balancing. We believe that the most important factor for RTDPS is meeting the deadline constraints and that's what our algorithm accomplishes. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

No description available.

Thesis (M.S. in Engineering Acoustics and M.S. in Electrical Engineering)--Naval Postgraduate School, June 1990. / Thesis Advisor(s): Miller, James H. ; Chiu, Ching-Sang ; Yang, Chyan. "June 1990." Description based on title screen as viewed on October 20, 2009. DTIC Identifier(s): Acoustic Tomography, Theses, Range Kutta Fehlberg Method, C Programming Language, Macintosh 2 Computers, Transputers, Parallel Processors, Numerical Integration. Author(s) subject terms: Acoustic Tomography, Ray Tracing, Parallel Processing, Gaussian Beams, Transputers. Includes bibliographical references (p. 122-124). Also available online.

Thesis (M.S.)--Rochester Institute of Technology, 1990. / Includes bibliographical references.

With the evolution of parallel and distributed systems, users from diverse disciplines have looked to these systems as a solution to their ever increasing needs for computer processing resources. Because parallel processing systems currently require a high level of expertise to program, many researchers are investing effort into developing programming approaches which hide some of the difficulties of parallel programming from users. Linda, is one such parallel paradigm, which is intuitive to use, and which provides a high level decoupling between distributable components of parallel programs. In Linda, efficiency becomes a concern of the implementation rather than of the programmer. There is a substantial overhead in implementing Linda, an inherently shared memory model on a distributed system. This thesis describes the compile-time analysis of tuple space interactions which reduce the run-time matching costs, and permits the distributon of the tuple space data. A language independent module which partitions the tuple space data and suggests appropriate storage schemes for the partitions so as to optimise Linda operations is presented. The thesis also discusses hiding the network topology from the user by automatically allocating Linda processes and tuple space partitons to nodes in the network of transputers. This is done by introducing a fast placement algorithm developed for Linda. / KMBT_223