To extend vision-aided capabilities to navigation of manned and unmanned aerial platforms over water, QuNav proposes to develop a technological approach of navigation and feature-state tracking (NavTrack). To reduce drift in inertial navigation outputs, NavTrack derives relative position updates by utilizing navigation-related features that are extracted from video images. To accommodate water movements, estimation of feature motion is explicitly included into the navigation mechanization. The proposed approach (i) builds on our mature feature-aided navigation mechanization in unknown environments; and, (ii) extends it to non-stationary features by augmenting the state-space formulation with modeling of water disturbances. The algorithmic approach is formulated as a joint estimation of INS error states, feature locations and their displacement over time. Main goals of the Phase I effort are set forth as follows: 1) Fully develop the navigation mechanization of NavTrack; 2) Initially validate the system functionality using simulations; 3) Collect initial experimental data with UAV flying over the water in order to assess behavior of vision features; 4) Demonstrate the feasibility of navigation mechanization with experimental results.
Benefit: The technological approach of NavTrack has significant commercialization potential for both military and civilian applications. For military applications it will be directly developed for GPS-denied missions over water. For the private sector, the largest commercialization potential will be realized for UAV inspection applications such as autonomous monitoring of bridges. For this application, NavTrack will enable seamless positioning between open sky and under-the-bridge segments where GPS positioning capabilities are degraded or denied. Phase I development will create a foundation for prototyping and transitioning of NavTrack. Successful accomplishment of Phase I tasks will enable us to (i) develop and verify all major algorithmic components of the NavTrack vision-aided navigation mechanization and (ii) demonstrate its technical feasibility through extensive simulations and initial experimental validations.
Keywords: feature-tracking, feature-tracking, GPS-denied navigation, motion modeling, vision-aided inertial, Navigation Over Water