A5065689972- Anomaly Detection Techniques and Applications
- Space Satellite Systems and Control
- Spacecraft Design and Technology
- Fault Detection and Control Systems
- Wireless Signal Modulation Classification
- CCD and CMOS Imaging Sensors
- Error Correcting Code Techniques
- Robotics and Automated Systems
- Synthetic Aperture Radar (SAR) Applications and Techniques
- VLSI and Analog Circuit Testing
- Low-power high-performance VLSI design
- Advanced SAR Imaging Techniques
- Modular Robots and Swarm Intelligence
- Optimization and Search Problems
- Technology Assessment and Management
- Artificial Immune Systems Applications
- Radar Systems and Signal Processing
- Adversarial Robustness in Machine Learning
- Parallel Computing and Optimization Techniques
- Radiation Effects in Electronics
- Distributed and Parallel Computing Systems
- Network Security and Intrusion Detection
- Machine Learning and Data Classification
European Space Agency
2023-2024
European Space Research and Technology Centre
2023-2024
Airbus (Germany)
2022
Satellite-borne Synthetic Aperture Radar (SAR) technology has revolutionized remote sensing applications by providing high-resolution and all-weather imaging capabilities. With the increasing availability of SAR data, need for efficient data processing become crucial. On-board emerged as a promising solution to address challenges associated with limited downlink capacity final products latency. Performing compression directly on airborne platform reduces raw transmitted ground stations,...
Higher autonomy in satellite operation is seen as the key game changer for space systems market next decade, with a considerable amount of agencies and startups focusing on bringing machine learning to space.
Compute in space, e.g., miniaturized satellites, requires dealing with special physical and boundary constraints, including the limited energy budget. These constraints impose strict operational conditions on on-board data processing system its capability sophisticated workloads suchlike Machine Learning (ML). In meantime, breakthroughs ML based Deep Neural Networks (DNNs) last decade promise innovative solutions to expand functional capabilities of drive space industry forward. Therefore,...
This paper presents the MLAB project, a research and development activity funded by ESA General Support Technology Programme under lead of Airbus Defence Space GmbH, with goal developing machine learning application benchmark for space applications. First, need dedicated to applications in spacecraft is explained, examples are described including their design challenges. Then presented, rules metrics, guidelines scenarios references. These include description reference workloads that have...
Artificial intelligence has found its way into space, and similar to the situation on ground demands powerful hardware unfold full potential. With heterogeneous compute platform that is offered by space-grade variant of Versal, AMD Xilinx presents a system particularly targeted at accelerating AI inference in space. This paper investigates design flow achievable performance this novel device. We present benchmark results terms concrete figures measurements, i.e., throughput, latency, power...
Within the scope of an ESA funded activity, Airbus Defence and Space GmbH completed a research development study in order to provide novel dataset develop flight-ready system for on-board anomaly detection. This work includes extraction satellite telemetry data, identification anomalies, machine learning models finally deployment algorithms via hardware acceleration. We present benchmarking results three accelerated ML from within final system.
This paper presents reference implementations for a multitude of space applications from the Machine Learning Application Benchmark. Reference include respective model, its on-board hardware implementation, test scripts and final benchmarking results. In publishing these implementations, we make significant contribution to benchmark provide more insight into viability machine learning applications.
Artificial intelligence has found its way into space and necessitates a powerful flexible hardware platform to keep up with the fast-paced AI domain. With space-grade variant of Versal, AMD-Xilinx offers one first space-ready accelerators that combine multiple compute paradigms, i.e., scalar processing (CPU), adaptive engines (FPGA), vector (AI-Engine array) an System-on-Chip. This paper provides thorough analysis capabilities respect throughput power efficiency for Multi-Layer Perceptrons...