The Global Positioning System (GPS) provides absolute positioning with meter-scale accuracy anywhere on Earth. However, centimeter-scale precision can only be obtained using expensive and specialized equipment, such as differential GPS. The key observation of the project is that many applications of wireless nodes, such as sensor arrays, require accurate relative node locations only and not precise absolute coordinates. The proposed effort utilizes information fusion in a network of GPS-equipped mobile nodes (sensors, smartphones, ground or aerial robots, etc.) to derive relative geographic location information that is more accurate than what traditional GPS is able to provide. In this project, the nodes share their raw GPS satellite measurements with each other and solve for their pairwise relative location vectors as opposed to individual absolute positions. The proposed work is clearly very challenging. GPS measurements correspond to ranges thousands of kilometers long obtained by utilizing radio signals traveling at the speed of light. All the measurement errors traditional GPS must handle will need to be addressed by our approach as well. The proposed unique combination of known methods from literature and novel techniques of our own shows great promise in achieving centimeter scale accuracy in GPS-based relative localization.
If successful, the project will have a truly broad impact through the numerous applications it will enable that are either not economical or simply not possible today. These include tight formation flying of UAVs, an urban guidance system for the visually impaired, collision avoidance and platoon control in autonomous vehicles, consumer grade surveying equipment, and mobile distributed microphone (and other sensor) arrays. Moreover, the project will focus on involving undergraduate students in our research. The Vanderbilt Mobile Application Team (VMAT) is a young volunteer self-organizing undergraduate student club with the primary purpose of teaching/involving students in the field of mobile phone application development. The PIs will encourage VMAT students to develop apps using our precise relative localization feature by providing them hardware and a library of our prototype system early on, and will place special emphasis on involving students from traditionally under-represented groups in VMAT in general and in this research project in particular.
