Demonstrating Controlled Change for Autonomous Space Vehicles

Abstract

Recent research discusses concepts of infield changes to overcome the drawbacks of conventional lab-based system design processes. In this paper, we evaluate the concept of controlled change by applying it to a demonstration of a potential future space exploration scenario with mobile robots. The robots are capable of executing several image computations for exploration, object detection and pose estimation, which can be allocated to both FPGA-and processor resources of a System-on-Chip. The demonstrator addresses three scenarios which cover application-, environment-, and platform change. The system adapts itself to any of the named changes. This capability can increase the autonomy of future space missions. Exemplary, the demonstrator executes adaption of applications during operation to fulfill the mission goals, adaption of reliability under changing environment conditions, and adaption to sensor failure.

Automating integration under emergent constraints for embedded systems

“…For space robots, we are more concerned about heterogeneous resource-constrained devices in a mission-critical setting where solutions to unexpected situations (e.g. hardware failures, environmental conditions, power shortage) shall be found autonomously [8].…”

“…Our application scenario originates from the CCC project in which a space robot was developed for exploring an unknown environment in search for certain objects (cf. [8]). The robot is equipped with a stereo camera and executes various imageprocessing algorithms.…”

MIMOS: A Deterministic Model for the Design and Update of Real-Time Systems

“…A model-based approach to the design and dynamic updates for cyber-physical systems is proposed in [27]. The work of [11] demonstrates that autonomous systems in operation can be updated through contract negotiation and run-time enforcement of contracts.…”