The CARM2G project

Predictable real-time behaviour of complex embedded mechatronic systems


The need for higher performance, more variability in usage, improved serviceability and reduction in the cost-of-ownership, drives a continuous need for optimization of complex mechatronic servo systems. Today’s product development methodologies only restrictedly address these needs, such as for accurate predictions of real-time system performance. As a result, all too often, a product development team is confronted with a long and unpredictable test and integration phase. The challenge therefore, is to improve the product development approach to such a level that full insight, overview and predictability of complex mechatronics designs can be guaranteed at earliest possible project stage.

Multi-core solutions

Research results

The approach taken by the CARM2G project is to introduce advanced model-based techniques that can be used as formal specification and design-space exploration means for design of servo-control systems. It is based on a library of parameterized model elements that describe the HW and SW behaviour of a multi-processor, multi-core execution platform. By calibrating and validating those models on the basis of existing product family members, an accurate representation of the future execution platform is obtained to design and dimension the servo application.

The system design methodology is based on an IDE (integrated development Environment) that incorporates the POOSL modelling language and toolset and TRACE visualization tool.

The main results of the project are:

Y-chart

  • Domain-specific languages (DSL’s), following a Y-chart decomposition, of the complete servo domain, including applications, platforms and mappings to specify the servo control domain

  • Automated robust multi-processor, multi-core deployment and scheduling techniques establishing correctness-by-construction with respect to servo control functionality and timing requirements

  • Automated analysis techniques enabling detailed timing and resources usage analysis, including stochastic variability analysis

  • Visualization facilities (Gantt-chart) to present analysis and synthesis results in a human-understandable way 

  • Very high performance, flexible and cost-effective  execution platform architectures (100 kHz+) incorporating  multi-processor, multi-core CPUs, combined with FPGA technologies

  • An Eclipse based IDE integrating the specification, analysis, synthesis and visualization techniques

Value proposition

The techniques can be applied in industries that deal with complex mechatronic or embedded control systems. The key value proposition is found in the overall approach: by investing more effort early in the product design stage, unwanted surprises during system integration and test can be avoided. This will significantly reduce the overall project risk and thereby significantly reduce project costs and time.