A Review on Software/Systems Architecture Description for Autonomous Systems

Background: The design of Autonomous Systems must consider multiple elements of the system, such as agents, physical objects and their software counterparts, control mechanisms, sensors, actuators, and other components. All these distributed elements in the environment make the necessity of creating multiple views for design, including system coordination, development, structural, and behavior views. Therefore, Software/Systems Architectures have been recognized as an important element in process development to manage the systems' complexity. Objective: The objective of this article is to describe a review of architecture characteristics, approaches, styles, and standards that are commonly considered for the development of autonomous systems. Method: First, we describe important elements of software architecture, as well as the standards used in this field. Then, we describe the types of approaches for architecture design. In addition, we provide a classification of software/systems architectures for autonomous systems. Results: As a result, we present a review on the Software/Systems Architecture description for Autonomous Systems. We also find that formal architecture standards are rarely considered in practice, and a large variety of nonfunctional requirements is mentioned. Conclusion: As autonomous systems deal with many components interacting with the real world, under certain quality constraints, considering trade-offs and decisions, software/system architectures are highly relevant for managing all this complexity. A list of main challenges for autonomous systems is described and then discussed according to a review of the literature. This review can be useful for professionals and researchers in identifying software/systems architecture as an important technical element for developing autonomous systems.