The objective of this project is to perform exploratory investigation into accurate wireless node localization for wireless as hoc networks in realistic network environments, where signal obstruction due to multipath or shadowing is a major concern. To address such practical challenges, this project explores a new and reliable localization technique that mathematically fuses range estimates derived from Received Signal Strength (RSS) and Time Difference Of Arrival (TDOA) measurements.
Intellectual merit: The intellectual merit lies in the design and implementation of an efficient, hybrid algorithm integrating Dempster-Shafer theory with a state-of-the-art machine-learning tool referred to as Support Vector Machine (SVM). The hypothesis is that SVM-based signal propagation models will possibly help improve the accuracy of range measurements. The proposal entails high-risk as it is applied to real-time environments that are unpredictable and dynamic. If successful, there will be a high-reward in terms of accurately mapping RSS and TDOA signal information into distances, and offering precise node-position with no hardware upgrades whatsoever.
Broader impacts: The broader impacts are reflected in several aspects. This exploratory research will benefit US-based business enterprises that thrive on wireless network applications, particularly in such strategic areas as emergency services and wildlife habitat monitoring. The node localization challenge will be investigated from an interdisciplinary perspective combining the fields of Computer Science and Communications, Geomatics Engineering, and Applied Mathematics. Finally, the proposed research will lead to (i) Creation of new interdisciplinary graduate curriculum for students including minorities and underrepresented groups and (ii) Promotion of interdisciplinary NSF research.
