Visual-Aided INS Using Converted Measurements
A new linearization method using converted measurements is developed for aiding a navigation grade Inertial Navigation System (INS), which is endowed with a monocular Electro-Optical (E/O) sensor—We refer to Visual-INS (V-INS). Using the new linearization method, a ground feature of unknown position is optically tracked over a short, 10-second measurement epoch by a human operator/pilot while the aircraft is on autopilot. Although the E/O sensor-provided bearing measurements are inherently non-linear, they are converted to a linear form allowing the application of a standard linear regression algorithm for nav. state estimation. The only linearization error introduced is in the short measurement error terms. Successive bearing-only measurements taken over time allow good estimation of the INS horizontal velocity error and horizontal accelerometer biases. At the completion of the measurement epoch, the INS is corrected by subtracting out the estimated errors. Aiding the INS in this manner provides a significant improvement in the accuracy of the INS-provided aircraft navigation state estimates when compared to those of a free/unaided INS. Applications for this autonomous navigation approach include navigation in Global Positioning System (GPS) denied environments and/or when RF emitting/receiving sensors are undesirable.
Abstract © IFAC
Montgomery, T. J., & Pachter, M. (2016). Visual-Aided INS Using Converted Measurements. IFAC-PapersOnLine, 49(17), 112–117. https://doi.org/10.1016/j.ifacol.2016.09.020