Demonstration of Tight Optical Integration (TOI) algorithm using field data

Recently a biologically inspired algorithm, called tight optical integration (TOI), was developed for tightly integrating optical sensor with GPS. The algorithm involves the integration of a standard camera along with GPS range (pseudorange or carrier phase) measurements to form position estimates. Initial simulations showed that TOI is capable of providing a position solution with an insufficient number of GPS satellites and a visible ldquomarkerrdquo at a known location, with an inertial unit to provide attitude information. This paper demonstrates how a marker is selected from a picture frame and tracked among consecutive frames. TOI has the potential to navigate with one known marker and two or three GPS satellites. In this work attitude information is derived from the GPS velocity estimates assuming zero roll for a terrestrial vehicle. Additionally, the same TOI algorithm can auto-locate unknown features when the position of the marker is not available, and navigate by these features when location is lost. The TOI algorithm is unique because it relies only on GPS range measurements and the pixel data from a camera. No ranging sources such as radar or LIDAR are required. It has particular application to scenarios involving a reduced constellation; such a reduced constellation may be due either to an urban canyon or a denied signal environment.