Autonomous Video-based Navigation for Airborne Platforms

The overall project goal is to develop a video-based autonomous navigation system (VANS) that provides a reliable, accurate backup to GPS for navigation and targeting of guided munitions and platforms. The system consists of video cameras and an inexpensive inertial measurement unit (IMU) along with image analysis algorithms running on a real-time computer. The approach is based on previous OKSI algorithms and expanded to include imagery from multiple cameras. During Phase I, a flight package was developed and used to collect over 100 GB of time stamped flight video and ancillary data (IMU, altimeter, AHRS, and GPS data for verification) at various altitudes and attitude changes. The data was successfully used to demonstrate the techniques with the determination of aircraft heading and position using a frame-by-frame analysis of low-resolution video. Under Phase II, the system will be enhanced for night service using thermal infrared cameras, and data will be collected day/night over different terrains to fully test and optimize the algorithms. In addition, the algorithms will be ported to a real-time computer and demonstrated in a captive flight. A brassboard test system will be delivered to the Army at the contract end for use in further development and testing.