Mr Daniel Fernandez 991672 Mr Michael Still 964076 Mr Blake Swadling 982087 Ms Kristy Van Der Vlist 983118 Mr Nicholas Wheatstone 983131 fs 1.2 07 April 2002 Modified title to include "proposal" 1.1 05 April 2002 Spelling and grammar corrections. 1.0 04 April 2002 Created. The aim of this project is to develop a computer software package to assist the University of Canberra survey Laboratory and other relevant institutions with the analysis of their GPS, VLBI and SLR data relating to tectonic movement. Such an analysis involves using mathematical techniques to model the data itself, as well as any anomalies that may occur, such as white noise and random walk, in both the time and frequency domains. Once the data has been transformed it will be displayed in a graphical format, thereby allowing the user to gain an accurate understanding of what the data represents. The analysis of GPS, VLBI and SLR data is required by many areas of natural science such as geology and meteorology. Consequently, the efficient and precise analysis of such data sets is crucial. This, in turn, will allow far greater accuracy in modeling and understanding many natural phenomena including:- The ability to correctly assess and understand continental drift. The ability to pin point stress lines within a given land mass, which often leads to the discovery of mineral deposits. A clearer, more precise picture of rising sea levels, which are widely believed to be a consequence of global warming. A greater understanding of events that occur prior to natural disasters such as earthquakes, volcanic eruptions, floods and droughts, which may, in turn, give us the ability to accurately predict them. There are a number of errors and measurement accuracy issues to consider during the processing of these data sets. In addition, some data sets are more susceptible to certain types of error, such as white noise, than others. By allowing the accurate modeling of such data sets and their errors, the suitability of each individual data set for some types of analysis can be more readily assessed. Although the effects of white noise can be seen from a time series of reasonable length (about 3-5 years), to accurately see the effects of random walk, at least 6 years of continuous time series data is required. Consequently, the ability to process large data sets (at least 10 years) will be critical in determining the effects of random walk. In order to achieve the stated aims of the Tectonic Information Translation System ('the System') the following development methodology will be adopted and implemented. During this design phase extensive research will be carried out in order to gain a better understanding of the mathematics involved in the translation of tectonic data sets. This includes weighted least squares approximations, Fourier transforms, windowing functions, and the interpolation of data sets. Such research will also serve to enhance the knowledge of the team with respect to the purpose and relevance of such data analysis. In addition, at this stage, any potential customers will be identified and consulted to determine the desired functionality of the System. Once this is carried out, an analysis of all customer viewpoints will be undertaken in order to establish the functional and/or non-functional requirements of the system. A formal specification document, detailing all functional and/or non-functional requirements of the System. A conceptual design for the System. Documentation outlining the project aims, significance, methodology and the budget for the System. This stage is for the purpose of research and requirements collection only. Whilst a conceptual design for the System will be produced, it will not be implemented at this time. Once all requirements of the System have been specified a prototype will be developed. It is proposed that this prototype be produced using the MATLAB mathematics package. The prototype itself will incorporate all basic functionality determined in Stage 1. Once the prototype has been produced it will be used to validate and verify the customer requirements. In addition, this prototype may also be employed at a later stage to verify that the System's output is correct. The completion of this stage is considered to be the minimum requirement to achieve a pass level in the Computer Engineering Project course. A MATLAB throw-away prototype of the System. No part of the prototype produced in this stage will be incorporated into the final System. In the system development stage the conceptual design produced in Stage 1 will be implemented. It is anticipated that the system will be developed using the C++ programming language and be capable of execution on all Unix (i.e. both System V and BSD) style platforms. In addition, the system shall be as modular, scalable and multi-threaded as possible in order for it to take full advantage of Symmetric Multi Processor (SMP) machines. It is intended, at this time, that the System contain the following modules A core math engine whose function is to perform all required mathematical operations on the data set(s) provided A graphing module that is capable of graphing the transformed data sets produced by the math engine A file input/output module that will read the required data from the specified file(s) and transform it into the correct format before passing it to the math engine A Graphical User Interface module to allow the user to specify data set(s), perform selected mathematical operations and see the results displayed To produce a working Tectonic Information Translation System Documentation for all modules produced at this stage Upon conclusion of this stage a working system shall be produced. Although this system will be developed in accordance with the theoretical model produced in Stage 1, it may, however, be necessary to make minor revisions of this model at this time. In this phase the System will be thoroughly tested in accordance with a rigorous testing framework to ensure that it performs in accordance with its specifications. It is intended that tools such as XUnit testing modules be utilised to aid this process. In addition, an independent audit of all source code and a performance evaluation of the System will be undertaken at this stage. A document detailing the System verification and validation process(es) employed as well as the results of all test(s), audit(s) and performance evaluation(s) carried out in this stage At this time a fully working, thoroughly tested system shall be produced and implemented. Any necessary modifications required to the system designed in Stage 3 shall be carried out in order to ensure it meets all specifications. In this final phase all required documentation for the system will be produced. The System will also be demonstrated to the appropriate University of Canberra body at a time to be determined. The final report of the System The demonstration of a working, fully implemented Tectonic Information Translation System N/A The minimum requirement for the System during the prototyping stage is 24-hour 7-day access to 1 Linux machine with MATLAB installed on it. Although much of the prototyping code can be written in a Windows MATLAB environment, as the System itself will be produced for a Unix based platform it would be preferable to maintain consistency in Operating Systems ('OS') for the duration of the project. In order to complete the project the minimum hardware requirements are considered to be a high end Machine. Essential attributes for an Intel based machine are; a CPU running at approximately 1.5 GHz. at least 512 Mb of RAM. a Unix based operating system. If a non-Intel based machine is to be used then performance should be comparable with this system. It should be noted, however, that if restricted to the minimum requirements, the desired response time of the system (approximately 60 seconds to process the data and produce a result) may not be achievable. Desirable hardware requirements would be a high end, multi-processor machine with in excess of 1 Gb RAM Group members will also require 24-hour 7-day access to the target machine. All members of the group will also require any necessary privileges to said machine to enable the development of the System. For maximum performance, the target machine needs to be known, preferable in the early stages of the project. Writing the code for the system with a single processor machine in mind and then running it on a multiprocessor machine and vice versa will not maximize the efficiency of the System, and may, in fact detrimentally impact on performance. Mao, Harrison & Dickson, 1999 AilinMao Christopher G. A.Harrison Timothy H.Dickson 1999 Journal of Geophysical Research Journal of Geophysical Research 104 B2 2797-2816 Zhang et al, 1997 JieZhang YehudaBock HadleyJohnson PerigFang SimonWilliams JoachimGenrich ShimonWdowinski JeffBehr 1997 Journal of Geophysical Research Journal of Geophysical Research Error analysis of daily position estimates and site velocities 102 B8 18035-18055