ESA GNC Conference Papers Repository

PILOT (Precise and Intelligent Landing using Onboard Technologies) is a suite of advanced landing technologies aiming at providing global access to the lunar surface, precisely and safely. It includes Visual Navigation (VN) and Hazard Detection & Avoidance (HDA) functions to be integrated onboard the spacecraft. The PILOT subsystem consists of landing cameras, a 3D-imaging Lidar and a Landing Processing Unit (LPU) implementing the VN and HDA software functions. A first flight opportunity has been identified in the frame of the ESA-ROSCOSMOS cooperation on lunar exploration targeting South Pole regions. The main requirement of the HDA function is to detect slope, roughness and shadow hazards on the surface during the descent based on Lidar and camera measurements and to designate the coordinates of the safest landing sites. These position coordinates are then fed to the Lander platform managing the guidance and control of the descent manoeuvres. The HDA trajectory scenario is composed of two consecutive surface mapping and trajectory divert manoeuvres (referred to as scan #1/#2 and divert #1/#2). The first divert manoeuvre aims at avoiding slope and boulder hazards over a radius of 150 m on the surface at ground resolution of 1.5 m, while the second divert manoeuvre aims at refining the designation of the safe landing site and avoid slope and boulder hazards with finer ground resolution of 0.15 m over a radius of 30 m on the surface. The possible HDA trajectories can vary substantially depending on the guidance and control accuracy of the Lander platform affecting the possible altitude range of start of scan #1 and scan #2. Therefore, the HDA functions need to be able to operate under a relatively wide range of operational altitudes (1200 m to 500 m for scan #1 and 275 m to 200 m for scan #2). The HDA software is composed of functions dedicated to command the Lidar, process the raw Lidar measurements, perform motion compensation, reconstruct the surface topography, process the raw camera images, ground reference the resulting Lidar and camera images, compute hazard maps and then designate the safe landing site coordinates on the surface. These functions need to be executed within a restrictive time duration (e.g. within 15 s). This paper will first introduce the design of the PILOT HDA functions and the associated operational scenario. It will then describe the implementation of the functions in a high-fidelity simulation environment, real-time test bench environment and hardware-in-the-loop robot-on-rail laboratory environment. The paper will finally report and analyse the validation campaign results including Monte Carlo simulations, software execution and memory profiling on flight-representative processing unit and hardware-in-the-loop tests using flight-representative sensor unit in a dynamic laboratory environment.