Visual-INS Using a Human Operator and Converted Measurements

Abstract

A method for intermittently aiding an inertial navigation system (INS) is explored in which a ground feature of unknown position is optically tracked over a short measurement epoch—one then refers to Visual-INS. In contrast to current machine vision-based research trends, a human operator is entrusted with visually tracking the ground feature. Converted measurements of the feature position are generated from successive bearing measurements, and the estimated aircraft position. A linear regression algorithm is then applied to the converted measurements providing an estimate of the INS horizontal velocity components’ errors and also accelerometer biases. At the completion of the measurement epoch, the INS is corrected by subtracting out the estimated velocity errors and using the estimated accelerometer biases. 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 method include navigation in global positioning system denied environments and/or when the use of RF transmitting navigation aids is undesirable.