The Allegio project

Composable embedded systems for healthcare

Constant requests for additional features, performance improvements and a wider range of configuration options is driving up the complexity of embedded software in many types of high-end systems. The development methods commonly in use do not scale and many companies are facing a reduction in their rate of innovation. Especially, faults found late in the development process are costly and labour-intensive to rectify. This leads to a test and integration phase which is long and unpredictable, thus reducing the rate of innovation. The challenge is to find faults much earlier in the development process using modelling techniques.

Allegio

Research results

The approach taken by the Allegio project is to introduce and prove advanced model-based software design techniques that formally evaluate the correctness of designs and automate labour-intensive activities. This includes both design models, that are part of the design flow which ultimately result in code, and analysis models that evaluate certain aspects such as performance, evolvability, and functional correctness. To combine and relate these two types of models, domain specific languages (DSLs) and model transformations are used.

The main results of the project include:

  • Modelling, simulation, and visualization of requirements
  • Validation of software architectures
  • An Eclipse-based integrated development environment (IDE) using DSL technology
  • A manual for the constructions of DSLs and model transformations (including code generation)
  • Techniques for the semi-automatic transformation of legacy code

This system design methodology, using the POOSL modelling language and toolset, is proven in use within actual Philips iXR projects.

Value proposition

The techniques proven by the Allegio project are general in nature and can be applied to many industries involved in the development of complex embedded systems. Investing more effort upfront reduces the time needed for test and integration. Simulation and visualization of requirements and architectures can reveal faults and ambiguities early in the development process. The first experiences with the semi-automatic migration of legacy code show a large reduction of effort and code size. This frees up valuable staff to focus on the development of additional innovative product features. The overall impact is a significant increase in the rate of innovation.