Implementation Challenges in Real-Time Middleware for Distributed Autonomous Systems

摘要

Exploration missions will achieve a sustainable humanrobotic presence in space and on the surface of the Moon and of Mars. Sustained operations require higher performance, which in turn implies a higher degree of asset autonomy. For example, construction, maintenance, or insitu resource utilization will be accomplished by efficiently executing complex tasks in the absence of long-haul teleoperation from Earth. Moreover, assets will communicate with each other and form a distributed system. For example, multiple units will collaborate on, say, a construction task. The resulting system is both autonomous (i.e., it can operate in the absence of direct Earth tele-operation) and distributed (i.e., multiple assets communicate and collaborate with each other). Robustness and economy further demand the re-usability of tested middleware whenever possible. Finally, the distributed autonomous systems will include assets that operate in real-time: examples are surface robots, rovers, CEV on-board actuation units, or space construction robots. In this paper, we describe the transversal challenges in implementing real-time middleware components for distributed autonomous systems. The emphasis is on reusable components that address issues that are shared among distributed autonomous real-time systems. Specific applications are not the focus of this paper.