Programming Model for Cyber-Physical Systems

Martin Richter, Matthias Werner

The importance of Cyber-Physical Systems (CPS) is growing in our daily lives as phenomena like the Internet of Things, Smart Grid, Smart Logistics, Smart Farming, or Industry 4.0 emerge.
Such systems consist of heterogeneous and mobile sensors and actuators which observe and influence their environment.

Current programming models facilitate abstractions that are close to the application domain of digital systems.
Therefore, the semantics of CPS applications are often not clear with respect to their influence on the physical world (the application domain of CPS).

Our approach introduces a physical programming model to make these semantics more explicit.
According to this model, the application developers program the physical environment instead of the digital system by describing a desired future state of the physical world.
From such a description, the system can determine what actuator and sensor actions are required to move towards the target state.

Our programming model is based on a physical machine model that provides a framework for describing the physical semantics of the application.
Additionally, it enables the estimation and prediction of the outcomes of future physical actions.
In this talk, we will present such a machine model in conjunction with the programming model as described above.
We have found that the complexity for deducting required actuator actions rises exponentially for non-trivial scenarios.
Therefore, new abstractions need to be found which potentially pave the way for a more efficient solution.