Advanced intelligent agents for optimised dynamic process monitoring and defect inspection in construction projects

Defects and errors in new or recently completed work continually pervade the construction industry. Whilst inspection and monitoring processes are established vehicles for their 'control', the procedures involved are often process driven, time consuming, and resource intensive. Paradoxically therefore, they can negatively impinge upon the broader aspects of project time, cost, and quality outcomes. Acknowledging this means appreciating concatenation effects such as the potential for litigation, impact on other processes and influence on stakeholders' perceptions textendash that in turn, can impede progress and stifle opportunities for process optimisation and innovation. That is, opportunities relating to for example, logistics, carbon reduction, health and safety, efficiency, asset underutilisation, and efficient labour distribution. This study evaluates these kinds of challenge from a time, cost, and quality perspective, with a focus on identifying opportunities for process innovation and optimisation. It reviews textendash within the construction domain textendash state of the art technologies that support optimal use of artificial intelligence, cybernetics, and complex adaptive systems. From this, a conceptual framework for development of a real-time intelligent observational platform (RtIOP) supported by advanced intelligent agents, is presented and discussed. RtIOP actively, autonomously, and seamlessly manages intelligent agents (cameras, RFID scanners, remote sensors, etc.) in order to identify, report, and document 'high risk' defects. Findings underpin a new ontological model that supports ongoing development of a dynamic, self-organised sensor (agent) network, for capturing and reporting real-time construction site data. The RtIOP model is a 'stepping stone' towards advancement of independent intelligent agents, embracing sensory and computational support, able to perform complicated (previously manual) tasks that provide optimal, dynamic, and autonomous management functions.