Innovative Approach for Enhanced Navigation and Seeker Exploitation

  • Published
  • By Dr. Jimmy E Touma
  • Munitions Directorate
EGLIN AIR FORCE BASE, Fla. --  Researchers from AFRL and Northrop Grumman went on the OFFENSE, developing an innovative approach based on the use of optical flow for [achieving] enhanced navigation and seeker exploitation. The method works to adaptively fuse, in real time, all available navigation data--including that derived from inertial measurement unit/Global Positioning System inputs, altimeters, star tracker devices, passive imaging sensors, and digital elevation databases. Accordingly, the novel approach permits continuous navigation throughout non-GPS environments, while yielding improved exploitation in the presence of GPS. The technique also reduces target location error and provides moving target indication.

A critical need exists for a fully autonomous and robust navigation capability in GPS jamming or signal interruption environments. While GPS plays an important role in the guidance and navigation of small munitions and unmanned air vehicles by providing positional updates--which, in turn, bound the drift of onboard inertial navigation systems--it is nonetheless susceptible to jamming and is also unreliable (or unavailable) in urban canyons and indoors. The U.S. Air Force, Navy, and Army are thus keenly interested in a capability that reduces the vulnerability of GPS navigation to interruption--whether deliberate or unintentional.

Scientists dedicated to the pursuit of computer-enhanced vision technologies have witnessed many approaches for computing "ego-motion," which refers to the (sensor or biological organism) observer's movement. AFRL's newly devised approach to navigation exploitation computes ego-motion using optical flow measurements to estimate observer movement. Optical flow, defined as the apparent motion of brightness patterns, is characterized by a field of two-dimensional (2-D) velocity vectors. These 2-D vectors are, in fact, projections of the surface points' three-dimensional (3-D) velocity vectors onto the image plane. The 2-D velocity vectors are derived from image sequences that can then be used for inferring the 3-D velocity of the imaging platform. Since optical flow is a projection of a 3-D velocity vector onto a 2-D image plane, there is inherent ambiguity in inferring 3-D velocity. The observed optical flow must therefore be fused with other measurements--depth, for example--to remove ambiguity and provide an estimate regarding the 3-D velocity vector of the imaging platform. Knowing the observer velocity vector and the last (most recent) GPS position estimate will, in theory, enable navigation in non-GPS environments.