Autonomous optical flow-based initial orbit determination

The navigation algorithms proposed in the literature to achieve pinpoint landing for space exploration missions require initialization with state estimates from ground. This limits the missions that can be accomplished and significantly increases operation costs. This paper introduces an orbit determination (OD) algorithm, which can be used to autonomously initialize existing terrain absolute navigation (TAN) algorithms with pinpoint landing accuracy. Visual data is collected over the course of two orbit periods. This data is used to estimate the Keplerian orbit parameters of the trajectory. The estimated orbit parameters are used to initialize a crater-based TAN system independently from ground telemetry, thus providing a fully autonomous navigation system with pinpoint landing capability. An extensive test campaign was performed using both simulated measurements and rendered images. The algorithm was shown to successfully enable TAN initialization.