Our team and mission

This position is based at the European Space Research and Technology Centre (ESTEC) – Noordwijk, Netherlands

Under the direct authority of the Radio Frequency Systems and Payloads Division, the Radio Frequency Payloads and Technology Division is responsible for RF payloads and technologies for space and ground applications and associated laboratory facilities.

More specifically, its responsibilities encompass:

• Payloads with RF interface for telecommunication and navigation exploiting different technologies (e.g. analogue, digital, optical) including design and performance analysis tools and testing;
• Earth observation and scientific RF active and passive instruments, including design and performance analysis, engineering and testing up to sub-millimetre waves;
• Wave-propagation and interaction relevant to space communications, navigation and remote sensing, including interference and regulatory aspects;
• Antenna systems, architecture, technologies and techniques for all space applications, including space vehicle TT&C and user segment terminals, as well as antenna engineering, and RF testing of antenna and material;
• RF technologies and RF equipment, also including vacuum electronics and high-power RF phenomena (multipactor, corona and passive intermodulation);
• Time and frequency references, modelling, design tools, measurements, performance characterisation and calibration techniques.

For further information visit our web site: http://www.esa.int

Field(s) of activity for the internship

Topic: Deep learning for on-board SAR data processing

Synthetic Aperture Radar (SAR) satellites have been growing in popularity in recent years and the amount of data to be acquired is set to increase substantially. This data increase is heavily linked to data downlinking costs, as well as computational and environmental costs due to the enormous amount of ground data processing necessary. Traditional data processing pipelines involve both ground and on-board operations. Usually, raw SAR data are acquired, compressed, and transmitted to the ground. Then, decompression, image focusing, and image analysis follows. Not only this is costly, but it can also take a significant amount of time, which is not compatible with situations when early detection and fast responses are crucial (e.g., disaster monitoring, detection of illegal shipping activities).

Recently, research has been undertaken on the possibility of performing more tasks on-board a satellite, using deep learning. This includes on-board data compression, on-board image focusing and on-board data analysis (by-passing the image focusing step). Depending on the requirements of each module, different deep learning models can be designed, resulting in trade-offs of accuracy and computational complexity.

Building on prior work, the aim of this project is to research and develop novel deep learning models that are suitable for on-board SAR data processing. The suitability will be determined based on predefined accuracy metrics of interest and computational performance. It is envisioned that a suite of algorithmic modules would be available on-board and chosen autonomously depending on the complexity of the observed environment.

Behavioural competencies

Result Orientation
Operational Efficiency
Fostering Cooperation
Relationship Management
Continuous Improvement
Forward Thinking

Education

You must have student status and be enrolled at university for the entire duration of the internship. You should preferably be in your final or second to last year of a university course at master’s level in a technical or scientific discipline.

Additional requirements

The working languages of the Agency are English and French. A good knowledge of one of these is required. Knowledge of another Member State language would be an asset.

Additional Requirements:

• Good knowledge of linear algebra, probability, and statistics.
• Basic knowledge of signal processing.
• Good knowledge of Python programming and of the Linux environment. Prior experience using a deep learning framework is an asset.