Dsplogic Inc — Department of Defense SBIR Phase I: AF181-037

Description

Highly efficient gas turbine engines require precise combustion, flow, and stability controls.To meet stringent future performance demands, engine control systems are transitioning to fully distributed networked architectures with smart sensors and actuators.While a networked control architecture has substantial benefits (e.g. reduced weight, improved robustness), it also presents new design challenges.The salient feature of networked architectures is a shared communications data bus with limited bandwidth and added latency.The performance of control algorithms is affected by the accuracy and rate at which data can be acquired from sensors and/or commands sent to actuators over the shared bus.Likewise, excessive latency between the sensor input and actuator output can cause instability.Limited processing resources must also be considered.We propose a unified framework and methodology for modeling, simulation, iterative design and optimization of communication-aware distributed control-systems.The framework is built upon a novel model of computation and communication that supports deterministic behavior. It will allow us to assess design tradeoffs of alternative communication architectures and scheduling strategies to arrive at the optimized networked control system design, with potential for automated implementation and code generation for reduced defects and development time.