ESA GNC Conference Papers Repository

Manned spacecraft rendezvous has a long, successful history. Unmanned, completely automated rendezvous has become state of the art. In both scenarios, rendezvous is carried out with a cooperative target, meaning that the target spacecraft is operational, communications is working properly and the targets location and status are known precisely. However, the immediate future holds new challenges for spaceflight in general, and rendezvous GNC systems in particular. Space Debris threatens future manned and unmanned spaceflight. The number of debris objects in orbit around Earth has increased considerably since the first missions into space. Collision avoidance manoeuvres are executed daily. The collision of Iridium 33 and Cosmos 2251 in 2009 has demonstrated impressively, that satellite collisions in orbit, despite thorough precautions, can happen, and will happen! It has shown the severe consequences of such a collision. Undeniably, dead satellites have to be either removed from or revived in orbit in an active manner. The former is known as Active Space Debris Removal (ADR), the latter as On-Orbit-Servicing (OOS). OOS has the potential to prolong satellites‘ operational lifetimes considerably, thereby launching a whole new business model. But such a mission is not easily accomplished. How can the target be grasped safely? Can an approach manoeuvre towards the totally passive and possibly tumbling target object take place fully autonomously or semi-autonomously? How can this be controlled and monitored from ground in hard real-time? And most importantly: How can all this be achieved with the highest reliability possible, avoiding another collision at all cost? To answer alike questions and to pave the way for such a mission, DLR‘s On-Orbit-Servicing End-to-End simulation (OOS-E2E) project is building the infrastructure to fly a simulated OOS or ADR mission from control centre consoles, including real sensors and robotics hardware. One of the project‘s key components is the rendezvous GNC system. It controls autonomously a precise and safe approach to a totally uncooperative target object. There is no communications with this target, no precise information about its position or status and no attitude stabilization. Rendezvous is possible only by using visual relative navigation based on optical sensors. However, there are no visual aids on the target‘s surface and illumination conditions may vary strongly, from total darkness, over sharp shadows, to full sunlight. Confronting these challenges, the rendezvous GNC system has to realize the approach with the highest reliability possible to prevent a collision at all cost. Therefore, the rendezvous GNC system includes three different sensors with distinct strengths and weaknesses: A classical CCD camera, a 3D camera (PMD) and a scanning LiDAR. The real-time data of these sensors are fed to sophisticated pose estimation algorithms, specific to each sensor, computing the relative position and attitude of the target object with respect to the chaser spacecraft in real-time. An innovative Extended Kalman filter fuses these pose estimates, considering different pose estimation frequencies and computation times, and produces a single high quality pose estimate. A highly advanced guidance function provides the approach trajectory and brings the chaser from hold point to hold point completely autonomously. The GNC system has been designed as a laboratory prototype of a complete satellite subsystem with standard conform telemetry and telecommand interfaces (Packet Utilization Standard, PUS). The European Proximity Operations Simulator (EPOS), a robotic Hardware-in-the-Loop simulator for the final 20 meters of rendezvous manoeuvres, serves as the simulation environment for the rendezvous GNC system. The target object, a true-to-scale satellite model (about 2.5 meters wide), with realistic surface materials, is mounted onto one of the industrial EPOS robots. The GNC sensors are mounted onto the other robot, representing the chaser spacecraft. A deep black theatre curtain behind the target serves as realistic space background. A 12 KW daylight spotlight realizes realistic illumination conditions. After outlining the rendezvous GNC system‘s role in the OOS-E2E project, this paper describes the rendezvous GNC system and its interacting components. This includes sensors, pose estimation algorithms, navigation filter, guidance and control functions. The system‘s operation within the EPOS simulation environment is shown. The paper concludes with the GNC system‘s current status and future work.