BIM for Smart Engineering Centre
Jointly established by Cardiff University, Dalian University of Technology, China Communications Construction Company
Jointly established by Cardiff University, Dalian University of Technology, China Communications Construction Company

Ontology supported Digital Building for Energy Monitoring

OntoFM: EPSRC & ActivePlan Ltd

PhD student: Michael Dibley, 2007-11 (now working as a research associate in Cardiff BIM group)

Main Supervisor: Prof. Haijiang LI, co-supervisors: Prof. John Miles

During 2007-2011, supported by an EPSRC-CASE studentship, a holistic smart sensor network with underlying ontology reasoning based decision making was developed to enable multi-agent based intelligent systems integration to conduct the autonomous building energy monitoring. It solves several key problems for complex building energy monitoring domain, e.g. it enables autonomous decision making based on smart resource negotiation and knowledge acquisition through ontology. The 4-year pioneering research started the ontology research theme within my research group, and it directly led to the success of several prestigious research projects in the area of building energy monitoring and management, e.g. EU FP7 – KNOWHOIEM, ref:285229, total value: £2.6M; SPORTE2, ref: 260124, total value: £3.9M; and RESILIENT, ref:314671, total value: £4.2M.

The simplified architecture is shown below. The information layer hosts a network of sensor nodes which store data collected from sensors, with added context, metadata, and ontological descriptions. These sensor nodes provide: an endpoint for hardware (bottom layer) to register its presence in the model, act as proxy for the low level devices, as well as providing long term data storage in relational databases. Interfaces to other resources such as databases storing building related documentation such as designs and schedules from any stage of the building lifecycle will be added later to the information layer. Returning to the multi agent layer, which is currently being developed, the deployed agents in general are characterised by the strong notion of agency (as described in the literature). The popular BDI (belief, desire, intention) model of agency is employed as a powerful abstraction tool, suited to applications where the complex and detailed processes cannot be fully modelled.

Agents within the software specialise different aspects of the model domain as a whole, adopting appropriate logical theories and axiomisation, as well as some specialist ontologies scoped to the various sub domains. This approach is critical in accurate knowledge representation and to preserve integrity in inference, while collectively providing comprehensive and diverse coverage. The goals of agents are typically to maximise some defined utility, and resulting behaviours are individual and collaborative, aligned and opposed. Agents collaborate to build confidence in query responses, and to complement incomplete and uncertain knowledge. Agent learning is realised via trusted message content from other agents, and inference and analogy of that content, as well as advice initiated by users. To maintain system integrity, agents’ behaviours conform to commonly accepted standards, establishing a network of trust between agents.

Relevant Publications

  1. Michael Dibley; Haijiang Li; Yacine Rezgui; John Miles, Towards Smart Building Monitoring Driven by Domain Knowledge Reasoning, Advances in Engineering Software, impact factor:1.4, 2014 (under revision)
  2. Dibley M, Li H, Rezgui Y, Miles JC, An integrated framework utilizing software agent reasoning and ontology models for sensor based building monitoring, Journal of Civil Engineering and Management, 2015, Impact Factor: 2.016, DOI:10.3846/13923730.2014.890645,
  3. Dibley M, Li H, Rezgui Y, Miles JC, Software agent reasoning supporting non-intrusive building space usage monitoring, Computers in Industry, (2013), DOI: 10.1016/j.compind.2013.03.008, Impact Factor: 2.062.
  4. Dibley M, Li H, Rezgui Y, Miles JC, Cost effective and scalable sensor network for intelligent building monitoring, International Journal ofComputing, Information and Control, 2012, 2010 Impact Factor: 2.932,
  5. Dibley M, Li H, Rezgui Y, Miles JC, An Ontology Framework for Intelligent Sensor-based Building Monitoring, Automation in Construction, Impact Factor: 1.8, (2012), DOI: 10.1016/j.autcon.2012.05,
  6. Dibley M, Li H, Miles JC, Rezgui Y, Towards intelligent agent based software for buildings related decision support, Advanced Engineering Informatics, Impact Factor: 1.7, (2011), v25: p311-329, doi:10.1016/j.aei.2010.11.002 (Top 25 downloads in 2011),
  7. Dibley M, Rezgui Y, Li H, Towards an integrated system for building energy management, ICT for Sustainable Places, 2013 annual international conference and exhibition, Supported by the European Commission, 2013, Nice, France
  8. Dibley M, Li H, Rezgui Y, Miles J, Wang B, Multi-Agent System for Real Time Building Monitoring, EG-ICE 2012, German(2012)
  9. Dibley M, Li H, Miles J, The application of intelligent agency in a software model for buildings,ICCCBE2010 and EG-ICE10, Nottingham,, (2010)
  10. Dibley M, Li H, Miles J C, Rezgui Y, Towards a Synchronised Semantic Model to Support Aspects of Building Management, 7th IEEE International Conference on Industrial Informatics - 24-26 June 2009, Cardiff, DOI: 10.1109/INDIN.2009.5195822
  11. Dibley MJ, Li H, Miles JC, Rezgui Y, A semantic model provision for the digital building, Book Series of the Department of Civil Engineering, Technische Universitat Berlin, Book No., (2009) 83-90 Huhnt W, Shaker Verlag, Aachen ISSN 1868-8357