From Architecture Models to a Target Platform: A Systematic Approach to Model-Driven Development of Dynamically Reconfigurable Avionics Systems

Abstract

Integrated Modular Avionics (IMA) is an advanced avionics architecture with a unified design, standardized components, and a unique computing paradigm. By allocating multiple functions to single processing units, IMA brought breakthroughs regarding efficiency, modularity, and maintainability to the aviation industry. The innovations in avionics, such as IMA, are ever-growing, with one of the more recent ones being dynamically reconfigurable avionics. Dynamically reconfigurable avionics refers to the state, i.e., configuration, change of the avionics system while it is executing. This allows the system to adapt and rearrange its software partitions and resource utilization as needed during runtime. By enabling the system to dynamically allocate and reallocate available resources relating to processing power, memory, and I/O devices, it allows for adapting to changing operational requirements and handling failures. This ensures the functional system availability for a longer time. The study introduced in this paper presents a model-driven methodology that addresses the complexity of developing such advanced avionics systems. The paper consists of three major aspects with regards to developing dynamically reconfigurable avionics: the architecture specification, model-to-text transformation for code and configuration generation, and embedding on a target platform. The paper demonstrates a systematic approach for capturing all essential architecture parameters of an avionics configuration followed in a model-driven approach. This approach can significantly enhance the development of dynamically reconfigurable avionics systems.