Unified Geolocation Using Heterogeneous Measurements: Mathematical Framework and Software Architecture, 16-R9743Printer Friendly Version
Inclusive Dates: 07/01/07 06/30/08
Background - SwRI's Signal Exploitation and Geolocation Division develops many unique and specialized geolocation systems. While most of these systems are built on a common conceptual foundation, the types of signals used in the geolocation process and the details of the implementations vary tremendously between systems. The effort required to design and construct these systems, as well as the effort required to optimize the systems for the best possible performance, can be extensive. This project derived a rigorous, general mathematical formulation of the geolocation problem and developed a specialized programming language for constructing these systems. The objective of the effort was to develop tools that enable the rapid construction and deployment of robust, high-performance geolocation systems that may be easily reused, modified and extended.
Approach - This project pursued four major research tasks. The first was the derivation of a general mathematical formulation of the geolocation problem. This effort formalized the mathematical elements that are common to all geolocation problems and developed optimal solutions to the problem for any geolocation architecture and any combination of geolocation signal measurements. The second task was the formal specification of the executable elements of a practical geolocation system, which enabled the complete representation of the components of deployable geolocation systems. The third task developed a notation that allows geolocation system designers to concisely specify a system using concepts and relationships that are native to the geolocation application domain. The final task consisted of the development of a compiler technology to process formal descriptions of geolocation systems and produce fully executable systems for a variety of platforms.
Accomplishments - A mathematical framework for geolocation has been derived that is capable of representing a wide variety of geolocation system architectures and supports the efficient integration of heterogeneous signal measurement types. Optimal solution strategies and quantitative predictions of system performance can be automatically derived within the framework. A new domain specific programming language for geolocation has been created that allows geolocation systems to be described succinctly, using the natural concepts of the geolocation problem. The language is closely tied to the mathematical framework, such that any statement made in the programming language has an unambiguous meaning within the formal mathematical framework. Compilers have been developed to transform programs written in the new language into executable geolocation components for multiple execution environments. Extensive testing of the new geolocation programming language was conducted by reproducing realistic, existing geolocation systems using the new design methodology. These tests demonstrated equivalent performance of the generated geolocation systems as well as reduced development time for the systems and reduced effort required to modify and extend those systems.