No abstract

Intelligent embedded systems (IES) represent a novel and promising generation of embedded systems (ES). IES have the capacity of reasoning about their external environments and adapt their behavior accordingly. Such systems are situated in the intersection of two different branches that are the embedded computing and the intelligent computing. On the other hand, intelligent embedded software (IESo) is becoming a large part of the engineering cost of intelligent embedded systems. IESo can include some artificial intelligence (AI)-based systems such as expert systems, neural networks and other sophisticated artificial intelligence (AI) models to guarantee some important characteristics such as self-learning, self-optimizing and self-repairing. Despite the widespread of such systems, some design challenging issues are arising. Designing a resource-constrained software and at the same time intelligent is not a trivial task especially in a real-time context. To deal with this dilemma, embedded system researchers have profited from the progress in semiconductor technology to develop specific hardware to support well AI models and render the integration of AI with the embedded world a reality.

Embedded systems (ES) ubiquity has increased in the last two decades; it is seldom to find any electronic device that is not controlled by ES. Additionally, there are a large number of ES which need to modify their behavior at run time in response to changing environmental conditions or in the cases where the requirements themselves need to be changed. Up to now, few researchers are interested in the high-level design process of the self-adaptive embedded systems (SAES) specifically in the field of requirement engineering (RE). While there exit some metamodels on RE for SAES, to the best of the authors' knowledge, there is no comprehensive metamodel that can be used as a reference for the development of RE for SAES. For this reason, the objectives of this paper are twofold, first, to review the literature state of the art and practice of requirements engineering for self-adaptive embedded systems and secondly to propose a complete metamodel (MM4SAES) that defines all concepts and relationships that must be taken into account in the development of SAES