Communication Patterns: a Novel Modeling Approach for Software Defined Radio Systems
Résumé
Efficiently programming Software Defined Radio applications still remains a
pending challenge.
While most of the efforts are focused on the processing part of a design,
communications are a great source of performance and portability issues that is often neglected.
Within the frame of a Model Driven Engineering methodology for the design of
dataflow processing applications, this paper proposes a novel approach to model
complex communication interactions. This approach relies on communication
patterns to capture communication protocols and standards at system-level,
independently of computations. A case study for cognitive radio shows how
communication patterns are efficiently used to generate cross-platform models
that can be ported and refined to specific applications and architectures.