Impact case study database

PANGU supports the testing and benchmarking of guidance systems for space exploration by providing a real-time, virtual test environment for spacecraft lander navigation and guidance that is cost-effective and flexible [E1].

Major missions

ESA have provided a letter of support [E2] that highlights recent and current major space missions and projects where PANGU is being used. For example, ExoMars is a major ESA/Roscosmos space programme with an initial budget of over €1.3 billion. PANGU was used by Airbus Defence and Space (UK) to test the critical ExoMars rover perception system with simulated PANGU images and corresponding disparity maps, enabling the accuracy of the perception system to be quantified [E3].

PANGU is cited by a senior ESA Engineer as “…, an instrumental tool in robotic exploration for ESA, being used in different areas such as object recognition to perform early analysis in Mars Sample Return" [E4].

ESA highlighted the impact of PANGU in April 2018, by selecting PANGU as one of the technologies that are "Shaping the Future", promoting the most successful and interesting activities within the ESA Technology Research and General Support Technology programmes [E5]. This is further corroborated by the section head of the ESA Guidance Navigation and Control Section at ESTEC (TEC-SAG), “ *PANGU has been an essential tool throughout the years for a number of tasks at the core of the work performed in the Guidance Navigation and Control Section of ESA. Both in developing new technologies in the area of vision-based navigation and in supporting project developments relying on this technology (notable examples in the last few years are ExoMars, Mars Sample Return, Hera, PILOT, Heracles/EL3, the CleanSpace initiative), the image rendering capability of PANGU has been used to design, develop and test the image processing algorithms enabling vision-based navigation application.” [E6].*

Cost-effective solution

PANGU can simulate a wider variety of scenarios and terrain types than is possible with alternative physical testing such as laboratory cameras mounted on robotic arms taking images of physical mock-ups, or drone/helicopter flight-testing (which is limited to Earth terrain, lighting and atmospheric conditions). By increasing the range of possible testing scenarios to verify navigation and landing technologies required for future space missions, PANGU significantly reduces the cost of generating test images and can be integrated with other spacecraft test systems. According to the ESTEC TEC-SAG section head “ Space images of celestial objects, especially those with the right resolution and radiometric properties to allow their use in spacecraft navigation and those of little-known objects, are a rare resource, hence the flexibility of PANGU to artificially generate high-fidelity images provides a key benefit in reducing the cost of generating test images and enabling rapid prototyping of vision-based navigation solutions in real-time closed-loop test facilities.” [E6].

Benchmarking and validation

PANGU is an officially validated tool (by ESA) and is being used to benchmark and validate new technical solutions for autonomous guidance “ The availability of a validated tool such as PANGU allowed, on one hand, ESA contractors to focus their efforts on algorithm development, rather than on image generation, and, on the other hand, allowed ESA to have common benchmark to evaluate and compare technical solutions developed by different contractors” [E6].

According to a senior ESA engineer on the ESA Lunar Resource Lander Team (HRE-PL): “ PANGU is used extensively in the frame of the PILOT project (Precise and Intelligent Landing using Onboard Technologies)… PANGU is used for the generation of realistic landing images and 3D models, which are used, in simulation and in test on representative HW, to validate the PILOT functions” [E7].

PANGU is also being used to prove test and develop guidance systems. A unique, resolution-enhanced model of Phobos, a moon of Mars, was created with PANGU to support the ESA Phootprint activity (Phobos Sample Return) to prove the technologies for a Phobos sample-return mission by generating higher resolution, representative images of Phobos than would otherwise be possible. GMV Innovating Solutions (a large, private, international technology company) then used PANGU with this model to generate simulated images of Phobos for their high-fidelity simulator to validate their feature tracking navigation algorithms by emulating the behaviour of the spacecraft sensors [E8].

Proving and validating future space technology

PANGU is having significant impact in developing and verifying future generic space navigation technologies. For example, GMV (UK) have also integrated PANGU within their Harwell Robotics and Autonomy Facility (HRAF), which supports the integration, validation, and verification of autonomy components for Entry, Descent and Landing Systems (EDLS) and surface rover mission scenarios. According to a Space Segment & Robotics Engineer from GMV (UK): “ In the UK, PANGU is being used in the frame of an ambitious ESA activity called Harwell Robotics and Autonomy Facility (HRAF). In HRAF, PANGU is an integral component in the architecture of a large, multi-site distributed simulation environment which includes multiple GMV and ESA sites.” and “PANGU is a core component in HRAF which provides synthetic images of the given scenario to the simulator. … in the context of Verification & Validation of the complex autonomy and GNC systems required in ambitious Space Missions, such as Mars Sample Return and HERA” [E9].

International Impact

PANGU is freely available to any institution working on ESA-funded projects, thus benefitting companies and institutions involved in the development and evaluation of ESA missions. It is also having an increasing impact outside Europe though PANGU licenses sold by STAR-Dundee Ltd for non-ESA missions.

STAR-Dundee is a University of Dundee spin-out company that specialises in spacecraft on-board data-handling and processing technologies. [text removed for publication].

MathWorks (a USA corporation specialising in mathematical software) recently developed a MATLAB workshop to showcase their powerful MATLAB Vision and Deep-learning Toolboxes by running MATLAB algorithms on a PANGU image sequence of descent to the challenging lunar South Pole region. PANGU was attractive for MathWorks because PANGU “ provided realistic simulated images of simulated descent on challenging terrain that would otherwise be difficult for us to obtain” and “ we see growing interest from our customers to use MATLAB and Simulink together with PANGU to develop and validate space navigation algorithms. ” This is corroborated by a letter of support [E12].