OBC-NG: Towards a Reconfigurable On-board Computing Architecture for Spacecraft Daniel Lüdtke German Aerospace Center (DLR) Simulation and Software Technology 38108 Braunschweig, Germany Daniel.Luedtke@DLR.de Karsten Westerdorff German Aerospace Center (DLR) Optical Information Systems 12489 Berlin, Germany Karsten.Westerdorff@DLR.de Kai Stohlmann German Aerospace Center (DLR) Institute of Space Systems 28359 Bremen, Germany Kai.Stohlmann@DLR.de Anko Börner German Aerospace Center (DLR) Optical Information Systems 12489 Berlin, Germany Anke.Boerner@DLR.de Olaf Maibaum, Ting Peng, Benjamin Weps German Aerospace Center (DLR) Simulation and Software Technology 38108 Braunschweig, Germany {Olaf.Maibaum, Ting.Peng, Benjamin.Weps}@DLR.de Görschwin Fey University of Bremen, Institute of Computer Science German Aerospace Center (DLR), Institute of Space Systems 28359 Bremen, Germany Goerschwin.Fey@DLR.de Andreas Gerndt German Aerospace Center (DLR) Simulation and Software Technology 38108 Braunschweig, Germany Andreas.Gerndt@DLR.de Abstract— The computational demands on spacecraft are rapidly increasing. Current on-board computing components and archi- tectures cannot keep up with the growing requirements. Only a small selection of space-qualified processors and FPGAs are available and current architectures stick with the inflexible cold- redundant structure. The objective of the ongoing project OBC- NG (On-board Computer – Next Generation) is to find new concepts for on-board-computer to fulfill future requirements. The concept presented in this paper is based on a distributed re- configurable system, consisting of different nodes for processing, management and interface operations. OBC-NG will exploit the high performance of commercial off-the-shelf (COTS) hardware parts. To compensate the shortcomings of COTS parts the OBC- NG redundancy approach differs from the classic way and error mitigation techniques will work mainly on software level. This paper discusses the hardware and software architecture of the system as well as the redundancy and reconfiguration concept. Our ideas will be proven in an OBC-NG prototype, planned for the next year. TABLE OF CONTENTS 1 I NTRODUCTION .................................. 1 2 RELATED WORK ................................. 2 3 SYSTEM ARCHITECTURE ........................ 3 4 RECONFIGURATION AND REDUNDANCY CON- CEPT ............................................. 6 5 SOFTWARE ARCHITECTURE .................... 7 6 PROTOTYPE AND FUTURE PLANS ............... 9 7 CONCLUSIONS ................................... 10 ACKNOWLEDGMENTS ........................... 11 REFERENCES .................................... 11 BIOGRAPHY ..................................... 12 1. I NTRODUCTION Future space missions face enormous challenges in several technical domains. One is the on-board data processing 978-1-4799-1622-1/14/$31.00 c 2014 IEEE. 1 IEEEAC Paper #2031, Version 3, Updated 12/01/2014. especially in the area of earth observation and robotics. The resolution of sensor systems on new satellites for earth observation increases in all dimensions (spatial, temporal, spectral, etc.). Because of limited communication bandwidth to the ground, demands to the on-board processing (filtering, selection, compression, correction, etc.) are growing as well. In the field of robotic exploration, e.g., deep space probes or rovers, a higher degree of autonomy is required to sup- port complex tasks to widen scientific activities. Current operations are restricted due to the small communication bandwidth and the long commanding delay. More autonomy requires more on-board processing to handle several sensor inputs and to support complex control algorithms. These future applications have in common that on-board sen- sors generate a huge amount of data which must be processed before transmission to the ground or further processing on- board. Future on-board system should provide the necessary computing performance in proximity to the sensor and if possible in real time. On the other hand the constraints of the space domain conflict with the desire for more computing power. Spacecraft in general have strong requirements to durability and reliabil- ity. This leads to very conservative design decisions. Only space-qualified components—usually radiation-hardened— are used, which have predominantly a lower performance than commercial components. Furthermore, today’s redundancy concepts waste available computing power on a spacecraft by considering redundancy only on subsystem level. Most space systems are not devel- oped in a comprehensive approach regarding the on-board computers. Most subsystems (attitude and orbit control, different payloads, data handling, communication) are devel- oped as independent units. Every unit comes with its own on-board computer and every computing unit has usually its own cold or hot redundant counterpart. That means a failing computing unit can normally not be replaced by a spare unit from another subsystem. Hence, a lot of computational resources remain unused. Another reason of non-optimal utilization can be found on 1