6235 – Data Driven Platform for Smarter Testing
Neil Loftus – Airbus UK Tony Goff – Dassault Systemes UK Cedric Laurent – Dynaworks (Airbus)
In a desire to make transportation increasingly autonomous, the need for a robust and precise positioning system is essential. Global Navigation Satellite Systems (GNSS) is typically the technology of choice to provide positioning, velocity and timing (PVT) information. Standard GNSS receivers are based on scalar tracking loops (STL) in order to track the signal of each satellite in view independently. Although, in harsh environments, such as urban canyons, the GNSS signals are strongly degraded by multipaths, signal attenuation and non-line-of-sight (NLOS) phenomena which highly impacts the stability of the tracking loops and thus affects the robustness and the precision of the PVT solution. Other GNSS receiver architectures are proposed within the literature which overcome some of the weakness of a STL architecture. They are based on vector tracking loops (VTL) where each tracking loops is intimately related with the others through an Extended Kalman Filter (EKF). Moreover, enhancing the VTL with inertial measurement unit (IMU) information through an inertial navigation system (INS), in an ultra-tightly coupled manner, allows to account for local receiver dynamics. Thereby it is possible to narrow down the bandwidth of phase lock loop (PLL) thus improving the resilience to noise and the stability of the tracking loops. Works on vector delay and frequency lock loops (VDFLL) architectures show good results in specific environments but the real-time hardware implementation remains complex and data synchronization and calculation time management between the EKF and the tracking loops is not trivial to handle.
In this contribution, an alternative approach is proposed for a GNSS/INS ultra-tight coupling which is based on the VTL architecture but in which the PLL bandwidth is adapted according to inertial navigation system (INS) information. This approach uses the calculated pseudoranges and Doppler frequencies from the receiver hardware which are tightly coupled with the output of an INS in order to provide an estimated position, velocity and time (PVT) solution using an extended Kalman filter (EKF). The solution is then used to calculate Doppler frequency rates which are fed back to the loop filter of every channel at the same time, in order to adapt the phase lock loop (PLL) bandwidth. The proposed architecture has the advantage to be easily implementable on a System-on-Chip (SoC) component such as an FPGA (Field-Programmable Gate Arrays), with minor modifications on an existing GNSS receiver platform. Moreover, compared to classical vector-based solutions, it has two main implementation advantages: 1) data synchronization and timing issues are easily handled; 2) the navigation message is decoded in the loop, without the need to run scalar loops in parallel or having to store pre-downloaded ephemeris data, limiting therefore the area occupied on the FPGA and the use of additional resources for storage.
The proposed GNSS receiver architecture uses GPS L1/C and Galileo E1 signals and is composed of one acquisition module and 16 tracking channels (8 GPS and 8 Galileo) which are implemented within a FPGA (Zynq-Ultrascale). The EKF module is implemented on the ARM of the Zynq. The approach is tested with synthetic and real-world data, where different signal degradations occur. The ultra-tight hybrid STL/VTL architecture’s real-time implementation is analyzed and compared to a reference software defined receiver designed with Matlab. A detailed description of the the GNSS receiver hardware is given.
4953 – TESTING OF THE NEW ADVANCED AVIONICS PLATFORM SOFTWARE FOR THE GRIPEN E FIGHTER
Maxime DAUDIER – SAAB
It’s hard to imagine that the smartphone in your pocket has anything in common with Gripen E, one of the most modern fighters in aviation history. But they have more similarities between them than meets the eye. Both rely on technology that can be upgraded and updated without the need for costly replacement, to ensure continual performance at optimal levels. And both have been developed with built-in flexibility, allowing the original product to evolve and to be customised to meet the changing needs of the user.
The same way that you download apps for your smartphone that fit your individual preferences, with Gripen, software adaptations can be made to address new and evolving types of threats.
The ability to customise the functionality of the fighter to address future needs is due to the adaptability of the advanced Avionics Platform Software (APS) architecture that is embedded in Gripen E software. This new avionics architecture is based on a Distributed Integrated Modular Avionics (DIMA) system, where flight critical functions are separated from tactical features. This in turn leads to a fast loop of iterative tactical upgrades with no effect on the basic aircraft nor its system- or flight safety.
This is a real technology leap, tactical functions can be upgraded without having to retest safety critical functions.
Since most of the complexity of the core avionics software resides in the avionics platform, the challenge for the Flight Test department was to find a reliable regression test suite of the APS meant to mainly ensure that the basic properties for scheduling and data communication are met and to test the robustness of the avionics platform in many aspects. The testing has to include also stress tests of the separation between the flight critical functions and the tactical system.
This paper describes the test suite built at aircraft level by the Flight Test department to achieve this and how the APS was live monitored during flight tests and then evaluated post-flight.
The results include successful flight testing of the new avionics platform software. The APS was monitored during many flights and reached the expected safety of flight. The test suite was then performed against many new software editions upgrading the Flight Critical functions, these editions were then monitored during flights and they all flew safely.
A lot of editions with only software impacts on tactical features were then allowed to fly without performing any test on APS and without performing any long Flight Safety Test Campaign in integration rigs and simulators. This gave shorter time between updated software deliveries and start of flight test, so the expected technology leap was reached. »
7835 – AITA. Automatic validation of parameters
Francisca Coll & Pedro Rubio – Airbus Defence and Space Spain
Most of the Test analyses are repetitive and therefore can be automated. The instrumentation of prototypes is becoming heavier and all parameters require validation. Most Test analyses tasks are manual, time consuming and prone to human errors.
AITA is an Airbus R&T project with the aim of developing a framework based on AI to automate most of these tasks and replace the obsolete tools currently in place.
This paper will describe the Machine Learning technique used for the validation of instrumentation parameters, which is part of the AITA project.
The need to validate automatically the instrumented parameters is clear, since thousands of Flight Test Instrumented parameters are not initially validated and they are not ready when required. Additionally, the automatic validation reduces workload and human errors. The technique used for the parameter validation is based on Decision Tree.
Decision trees are a type of machine learning model that is often used in classification problems. They work by dividing the dataset into smaller segments based on specific features, and then using that information to make predictions.
This paper explains the steps followed for the validation, the features selected to create the tree, the data set used to train the tree and finally, the decision tree prediction on flight instrumented parameters belonging to ADS aircraft fleet.
7691 – Anomaly detection: a first AI-based application for OASIS
Fanny Morel, Clémentine Barreyre, Anaïs Charcosset, Oihana Coustie, Jules-Edouard Denis, Linda Hammoud, Bruno Rouzier and Sébastien Zajac – Airbus Defence and Space SAS France
Airbus Defence and Space (ADS) has developed a big data platform named OASIS, which stands for Open Analytic ServIces for Space. Since June 2020, this big data platform is in charge of real-time telemetry ingestion, analytics and visualization for a wide and global ADS Satellite Fleet Supervision. It is used all along the satellite lifetime, from design and test to in-orbit support. The supervised fleet is now composed of 55 earth observation and telecommunication satellites and the platform aggregates 300Gb per day.
Every day, 600 active Airbus users take advantage of the capacities of the platform. In particular, in-orbit satellite operators are using OASIS twenty-four hours a day, seven days a week, for fleet supervision purposes. The OASIS capabilities to store, index, analyze data, alert in case of anomaly during in-flight phases are vital. In this context, Artificial Intelligence (AI) solutions are of high interest, as they are key enablers for decision-making systems, helping the operators to automatize routine tasks and focus on potential critical events.
In particular, the present paper will focus on the AI-based anomaly detection algorithms that are designed, developed and tested on OASIS. Unsupervised multivariate and monovariate algorithms are studied. Their development follows a three steps strategy: data analytics techniques investigation, algorithm implementation on a subset of telemetry and involvement of satellite experts. The OASIS platform offers two strong advantages for the development of such data science applications: easy access to analytics tools and easy access to a huge amount of telemetry data. The designed anomaly detection algorithms are now deployed in production for specific earth observation satellite use-cases, efficiently ensuring a daily monitoring service. This is a first step towards industrialization of AI-based algorithms on a larger and less specific satellite fleet. »
8772 – Airbus Defence & Space Solution for TVAC tests data supervision : DynaThermaNeo\rThibaut Le Goffic & Rémi Lamandé & Yann Bouetard & Cédric Laurent – Airbus Defence and Space France
Space simulation in Thermal Vacuum Chambers is one of the major tests performed for Spacecraft qualification.
During these tests, huge amount (several thousands) of parameters need to be acquired for near real time real supervision or post-test analysis.
These parameters can come from :
– The test facility such as the command control data
– The instrumentation on test such as the thermocouples used for the qualification
– The Spacecraft telemetries which are running also during the test
– Other sources of data such as simulations, test rigs, gas analyser, …
To aggregate these data and be able to deliver them to the different stakeholders of the test (test operators, satellite architects and analysts, remote customers), Airbus Defence & Space developed DynaThermaNeo solution.
DynaThermaNeo ensures the test preparation, piloting interface and near real-time acquisition. DynaThermaNeo solution is based on DynaWorks software platform developed by Airbus Defence & Space and used in-house and by external customers in Aerospace and Defence industries.
The main objectives (and stakes) of the solution are :
– Improve process efficiency –> test cost reduction
– Develop flexible interfaces and configurations (open solution) –> adaptability and durability
– Develop an user friendly user interface –> training time reduction
– Be compliant with network constraints and test means
– Reliability on critical systems »
