Skip to main content

Vision based integrated navigation for space exploration

Final Activity Report Summary - VINSE (Vision Based Integrated Navigation for Space Exploration)

This research project showed that the integration / fusion of visual information with inertial measurements can provide robust surface relative navigation suitable for planetary lander missions to many different celestial bodies (e.g. planets, asteroids and other small bodies). The objectives have been achieved through an international collaboration with the world's leading organisation on technology for planetary landers and space exploration, JPL (Jet Propulsion Laboratory) / NASA (National Aeronautics and Space Administration) and the UoD (University of Dundee), Europe's leading organisation on simulation for testing planetary landers.

An accurate, rapid and robust sensor / data integration / fusion method for planetary landers has been successfully developed, tested and validated. The results indicate that the new unscented Kalman filter to support vision based planetary lander navigation provides slightly more accurate navigation and attitude estimates for small initialisation errors and significantly better estimates for large initialisation errors than the existing GNC sensor data integration methods available (e.g. EKF). In addition, the results also indicate that the UKF is easier to tune, easier to implement and more robust than the equivalent EKF. The research results show that it is possible to significantly improve the performance (accuracy and robustness) of existing Guidance and Navigation systems for planetary landers using novel concepts, approaches or methods to perform GNC sensor / data integration / fusion. These results are particularly relevant for space explorations missions involving planetary landers. The novel vision based unscented Kalman filters developed, tested and validated can be used to navigate planetary landers during the entry, descent and landing phase.

A detailed knowledge of space exploration missions to small bodies (e.g. asteroids and comets) was achieved. The originally planned extension of GNC sensor integration techniques to small bodies was successfully performed through a modified system dynamics and model. An algorithm to develop a track management system for vision based planetary lander navigation was also developed.

The main relevance of the research work developed is for entry, descent and landing navigation of spacecraft, sample return missions, docking of spacecraft, navigation of space rovers and virtually in any space exploration mission involving planetary landers. It can be used in space exploration missions to planets, and especially in space exploration missions to small bodies where the dynamics is usually unknown until the spacecraft is in the vicinity of the target body. The robustness of the UKF compared to the EKF could play an important role in this kind of mission. Other possible applications include military systems and industrial systems where navigation is needed, such as aircraft, UAVs, etc.