Building automation (BA) systems ensure comfort and safety of the occupants while optimizing energy consumption. Malfunctions of electro-mechanical equipment or the associated control function result not only in incorrect operation, but also wastage of energy. Automated fault detection and diagnostics (AFDD) of processes like heating, ventilation, and air-conditioning (HVAC) is a challenging task due to the diversity of the applications and inter-dependencies between the equipment. Semantic description of the building and its equipment can help AFDD agents to better understand the system and continuously examine its operation. However, AFDD using only the structural description of the building topology and the contained equipment without relating it to the control functions requires manual verification of the root cause. In this work, we added a model of the control functionality to the semantic information of the structural and topological description and represented it using W3C Web of Things Thing Descriptions. Based on these descriptions, we have demonstrated in a case study of a building installation that fault detection rules which have access to the semantic description of underlying control functions can be formulated more precisely and thereby avoid false positives.

Building automation systems control many aspects of today's buildings - lightning control, air condition, shading, access control, and surveillance, just to name a few. This diversity and the diversity of related technologies and protocols entails that the extension or integration of building automation systems requires a major effort. To address these challenges, the ITEA3 Building as a Service (BaaS) project developed a reference architecture that is based on a service oriented approach enhanced with semantic descriptions that aims at model based code generation and simple integration of legacy devices. In this paper, we present an overview of the BaaS reference architecture with a specific focus on the information model and on the envisioned way to take advantage of the semantic descriptions for information filtering, search and discovery throughout the lifecycle of a building automation system.

As of today, building automation systems are present in almost any commercial building. They perform climate control, lightning control, access control, surveillance, and quite a few other tasks. As a result of their evolutionary development, building automation systems are divided into separate silos of disciplines that are not well integrated with each other. As of today, a variety of communication protocols, data models and engineering approaches are used by different vendors. Existing standardized building automation protocols as BACnet or KNX allow integration of some disciplines on the communication level but fail to provide means for common description of devices, services and data on the semantic level. This means that building automation applications that span multiple disciplines require a high effort for development, engineering and maintenance. If devices from multiple vendors are integrated in one installation, a set of different engineering tools and vendor-specific knowledge is required. In the ITEA “Building as a Service” (BaaS) project we try to overcome these deficiencies and define a common way to develop, engineer, commission, operate and maintain building automation systems following a service oriented approach. The whole process will be supported by semantic models to reduce costs and time-to-market, which is a quite new approach. In this paper we will present the current state of the work with special regard to domain modeling and model driven processes that are currently being specified for the BaaS platform.

In this paper large scale multihop sensor networks are established as non-beacon enabled ZigBee mesh networks. The lifetime of the network is increased by putting nodes to sleep and to wakeup state autonomously. To enable a reliable system with sensor nodes sleeping in an asynchronous manner, we propose a cross-layer sleep scheduling solution coupled with ZigBeepsilas proposed AODV routing. It consists of two parts: a) neighbor aware communication (NAC) and b) adaptive resynchronization (AR). NAC avoids sending packets to sleeping nodes while AR allows the sensor nodes to adapt their sleeping schedule to their neighborspsila duty cycles. ns-2-simulations show that the performance of such a cross-layer optimized system in terms of end-to-end delay and packet delivery ratio is comparable to the benchmark case of synchronized sleep schedules.