| Popis: |
Space missions are becoming more and more challenging, requiring fast and adaptive reactions to external stimuli that can be obtained with higher level of on-board autonomy. Because of such requirements, Guidance Navigation and Control technologies are increasing in complexity and, as a consequence, demand a more structured validation plan, to be started already during the preliminary stages of the mission. GMV standard for the Design, Development, Validation and Verification of GNC technologies is an incremental validation that starts with unitary testing of functions and modes up to flight code run inside representative processors and with engineering/qualification sensor models in close loop. The critical review of requirements allows for the derivation of GNC subsystem specification and for a preliminary design of GNC modes and functions necessary to achieve mission objectives. First step is functions unitary testing, but using previous experience and heritage, it is possible to test the algorithms in a representative closed loop GNC prototype already in phase A-B1. Model-In-the-Loop simulations demonstrate the feasibility of the preliminary design and the robustness of the selected solutions using Monte Carlo test campaigns. Being able to perform such tests during the preliminary stages of the development allows for efficient iterations at system level, giving valuable contributions for trade-offs that involve other subsystems. A fast prototyping based on autocoding is performed as following step, verifying the GNC code in Software-In-the-Loop tests. To demonstrate the feasibility of the design, the Processor-In-the-Loop step is then a key point of the incremental validation in order to verify the behaviour of the GNC code in a representative Hardware and to identify the computational resources required through code profiling. At this stage it is possible that specific function/algorithm (e.g. Image Processing) turns out to be too demanding for a space qualified processor like a LEON. In this case, a HW-SW co-design is performed in order to get advantage of the HW acceleration that an FPGA can offer. An estimation of the required resources and HW units is an essential input for Spacecraft system in order to perform budget activities and interfaces between subsystems. Furthermore, this high level of simulations fidelity, allows to take into account also operations constraints, defining a detailed time line of the different mission phases. In the final stage of the incremental validation Hardware-In-the-Loop is performed, including representative sensors in the loop (e.g. camera in a vision-based GNC). Two environments are foreseen for the HIL tests: an optical test bench where the images are generated on a high resolution screen and only a picture of it is taken; a robotic test bench (platform-art(c)) that re-creates a space-like scenario in terms of illumination conditions and simulated gravity using a dedicated laboratory with robotic arms. The platform-art(c) tests are the most complex to set-up and to execute, but there is a clear advantage in terms of representativeness because no artificial image is generated and the camera directly takes a picture of the target mock-up. This paper will include the details on GMV DDVV approach, which allows to perform end-to-end GNC simulations at very early phases of the mission to bring into considerations aspects normally not really considered in detail till later mission phases. GMV standard allows (through very fast iteration loops and full algorithms/SW coherence approach) algorithms updates/iteration till phases where they should be nominally frozen (thus, gaining flexibility and introduction of last techniques in the late phases of the mission. Copyright 2019 by by GMV aerospace and defence. Published by the EUCASS association with permission. A. Pellacani, M. Graziano |
