The Air Force Research Laboratory (AFRL) recently awarded Northrop Grumman [NOC] $1 million for the second phase of the Maintain Accurate Geo-registration via Image-nav Compensation (MAGIC) program, according to a company statement.
The goal of MAGIC is to develop and demonstrate advanced, real-time geo-registration and navigation algorithms using a combination of cameras, an inertial measurement unit and Global Positioning System (GPS) information, when available. The program aims to capitalize on recent advances in the availability of low-size, -weight, -power and -cost camera systems that make the inclusion of camera information in navigation and geo-registration systems for airborne vehicles a significant opportunity.
 |
| Photo: Air Force. |
MAGIC is to continue improving geo-registration accuracy for positioning and pointing applications, even in GPS-denied conditions. Northrop Grumman, in MAGIC’s first phase, integrated geo-registration algorithms in a vision-aided inertial navigation system that can even operate in GPS-denied conditions. Having successfully demonstrated a prototype system in phase one, Northrop Grumman will flight-test the integrated system in phase two as well as incorporate additional improvements, such as highly-detailed 3D map generation in the algorithm.
Northrop Grumman spokeswoman Yolanda Murphy said the last two phases of the program are optional and contingent upon satisfactory progress and funding.
Northrop Grumman in phase one developed a non-real time, but working, prototype system, which was demonstrated using government-furnished data, Murphy said. In phase two, the company will implement the prototype as a real-time system and test fly it on an unmanned aerial vehicle (UAV). In phases three and four, Murphy said Northrop Grumman will continue to refine the system and test fly it under progressively challenging situations and concepts of operations.
Geo-registration of data is critical for accurate interaction between systems, such as located targets and handing off coordinates to other aircraft, according to Northrop Grumman. Geo-registration of images involves pairing unreferenced images with the physical locations or exact coordinates of depicted items. This allows aircraft to create accurate maps by stitching together photos and correlating them with their world-based locations, which is useful for intelligence gathering and targeting.
Northrop Grumman, partnering with Toyon Research Corp., is building on its experience in vision-aided inertial navigation under past programs such as Collaborative Robust Integrated Sensor Positioning (CRISP), which matched image features and processed visual motion estimations for precise navigation without relying on GPS.