Modeling, simulation and emulation of Intelligent Domotic Environments

Abstract

Intelligent Domotic Environments are a promising approach, based on semantic models and commercially off the shelf domotic technologies, to realize new buildings that provide users with increased care, adaptability and safety. Intelligent buildings exhibit complex behaviors and are composed of various subsystems, and of many layers and components: such complexity requires innovative design methodologies and tools for ensuring correctness and effectiveness. Therefore suitable simulation and emulation approaches must be adopted, and tools must be provided, to allow architects and designers to experiment with their ideas and to incrementally verify the effects of designed policies. Incremental design and implementation of new strategies or new functionalities requires a validation scenario where the environment is partly emulated and partly composed of real devices. This paper describes a framework, which exploits UML2.0 state diagrams for automatic generation of device simulators from ontology-based descriptions of domotic environments. The device models are managed by the DogSim simulator, that may either simulate a complete building automation system in software, or may be integrated inside the Domotic OSGi Gateway (Dog) as a new emulation driver (EmuDog), allowing partial simulation of virtual devices alongside with actual interface to real devices. The proposed approach has been tested both on a synthetic 6-room flat equipped with 120 devices and on a real home comprising 3 flats located in northern Italy. Synthetic and real experiments show that the approach is feasible and can easily address both simulation and emulation, thus supporting incremental design and validation of home automation policies in realistic home scenarios.

DoMAIns: Domain-based modeling for Ambient Intelligence

“…On the other hand, the DogSim [23] simulator exploits the DogOnt ontology and a library of state-machine templates to offer IDE simulation capabilities, with a particular focus on design of automation plants. Finally, the EmuDog [32] framework extends DogSim to support mixed interaction of real and simulated entities, i.e., it supports the emulation use case reported in Table 2.…”

“…Many combinations of simulated, real and formally represented domain entities are allowed. Whenever a given application scenario only involves ontology-based representations and/or other formal models carrying a well defined and verifiable semantics, simulation-based validation and formal verification can be performed [31,32]. For example, state diagrams can be exploited to describe formally verifiable end-to-end automation scenarios, from user models to AmI.…”

“…The Bank Security Booth use case described in the previous section has been implemented both as a fully simulated model (row 2 in Table 2, Section 4) by exploiting the DogSim framework [23] and as a partially emulated model with real touch sensors and door actuators (10th row of Table 2), by exploiting the EmuDog libraries [32].…”

SmartTags: IoT Product Passport for Circular Economy Based on Printed Sensors and Unique Item-Level Identifiers

“…A number of semantic appliance models have been defined in the smart home domain. Prominent among them is the DogOnt ontology [24], which describes homes in a device and network-agnostic way. It specifies both appliances that can be controlled as well as the structural elements of the home, such as a room or garden.…”

Emulating home automation installations through component-based web technology

“…Regarding simulation capabilities (i.e., only virtual components are deployed in the architecture and no real devices are present), the DogSim approach [22] describes the devices' behaviors through state machines and this proposal provides an SCXML compliant simulation engine for injecting new events and reacting to events and state changes of these simulated machines. Based on this previous work, the authors of EmuDog [23] integrates the DogSim features and the Dog domotic gateway [24] for mixed environments in which events can be simulated or caused by real device interactions. EmuDog approach discovers devices needing emulation from the house model and builds the corresponding state machine.…”