Virtual prototyping of distributed control systems
A lighting system (e.g. for an office building) is a prime example of a complex system: it is a large-scale distributed system, containing thousands of sensor and actuator components, connected via a (wired or wireless) communication network. The system can have very many configurations, but is comprised of only a limited number of types of components. Typical for modern lighting systems is the complicating factor that they have to cooperate with other systems (HVAC, network, power, security), sometimes leading to conflicting requirements.
The goal of the Prisma project was to reduce the effort for the full product lifecycle of distributed control systems: specification, development, validation, installation, commissioning, upgrade, etc. Installation and commissioning are very costly phases in a lighting system's lifecycle. Installation involves the installation of a lighting system's physical components in its environment (e.g. an office building) and commissioning involves the configuration of the system's software to guarantee the required system behavior. Commissioning requires a lot of (manual) testing in a late phase of development. Detecting and solving errors in a late development phase is typically very expensive, especially because testing is done on-site. A specific goal of the Prisma project is to replace on-site testing of a physical system by off-site analysis of virtual prototypes.
IoT systems are be specified in terms of a number of aspect DSLs, which separate the different concerns in the specification of lighting systems. For instance, an IoT system's physical devices (i.e. its sensors and actuators) are described independent of their behavior. The DSLs allow the specification of reusable behavioral templates that can be deployed onto many physical devices of a system. Next to the system specification DSLs, the Prisma approach includes DSLs to specify a system's environment. These DSLs describe the interaction of a systems with its users. For instance, for lighting systems, these DSLs describe how system sensors are being triggered by the corresponding building's inhabitants.
Figure 1: DSL Flow
Together, a specification of a system and its environment allows a system's behavior and performance to be analyzed by transforming the specification into a simulation model (POOSL, Java) or a model checking model (Uppaal). Two kinds of simulation are supported: closed simulations that fully capture the system's environment and interactive simulations that allow a user to trigger the system and observe its response. The visualization used to trigger a lighting system is shown in Figure 2; it is coupled to a lighting system simulation via Java-based co-simulation framework.
Figure 2: Witte Dame - simulation
Philips Lighting (now Signify)
Prisma project period
January 2013 - February 2018
+31 88 866 54 20 (secretariat)
“My drive is to scale up knowledge valorisation, balancing research objectives against the industrial expectations, for architecting and designing complex systems into the high-tech embedded industry for real industry valorisation.”