Periodic Report Summary - TARES (Theory Applied to Real Embedded Systems)
Project objectives
Our project, entitled 'Theory Applied to Real Embedded Systems (TARES)', aimed at bridging a gap between theory and practise for the design of Embedded Systems. We have investigated how to apply recent advanced theories and tools developed for timed systems to the design of existing and commercialised software products.
Project results
During the course of the project we have used high-level modelling formalisms like Hybrid Automata and Timed Automata to model manufacturing plants as well as complex hardware. The models we have built enabled us to obtain interesting preliminary results concerning the automatic synthesis of controllers and the profiling of binary programs that run on complex hardware.
We investigated the use of controller synthesis methods to build correct device drivers and came up with a list of key issues that would need theoretical foundations before applications can be expected. We also addressed security issues and contributed to results on how to ensure some security properties (opacity, non-interference). Finally, we proposed a new methodology to compute worst-case execution times for binary programs running on complex hardware (pipelines, caches). The methodology was tested on ARM processors and provided encouraging results.
Our project, entitled 'Theory Applied to Real Embedded Systems (TARES)', aimed at bridging a gap between theory and practise for the design of Embedded Systems. We have investigated how to apply recent advanced theories and tools developed for timed systems to the design of existing and commercialised software products.
Project results
During the course of the project we have used high-level modelling formalisms like Hybrid Automata and Timed Automata to model manufacturing plants as well as complex hardware. The models we have built enabled us to obtain interesting preliminary results concerning the automatic synthesis of controllers and the profiling of binary programs that run on complex hardware.
We investigated the use of controller synthesis methods to build correct device drivers and came up with a list of key issues that would need theoretical foundations before applications can be expected. We also addressed security issues and contributed to results on how to ensure some security properties (opacity, non-interference). Finally, we proposed a new methodology to compute worst-case execution times for binary programs running on complex hardware (pipelines, caches). The methodology was tested on ARM processors and provided encouraging results.