Relative Optical Navigation for a Lunar Lander Mission

This work explores the problem of providing relative velocity navigation for an autonomous precision landing approach on the moon without the use of telemetry or known points of support. An error-state Unscented Kalman Filter for the fusion of inertial and optical imaging sensors is presented. These sensors include a star tracker, a monocular surface camera and a laser altimeter. The filter estimates position, velocity and attitude, which, together with an initial position based on crater matching, allows for trajectory following to the surface. A main difficulty is the scale ambiguity in optical flow. The laser altimeter has been included to resolve this ambiguity and allow for velocity and altitude estimation. The scenario of a lunar landing from parking orbit was chosen to test and verify the developed navigation method in simulation using a high resolution surface model of the moon.