From initial design to life-cycle planning, the creation and operation of the built environment involves a number of complex decisions. These decisions must take into account disparate aesthetic and socioeconomic factors, inevitably intertwined with the impacts of aging and hazards, through data-informed dynamic approaches. This increasing level of complexity has exceeded the limits of existing computational methods used to optimise decisions for structures and infrastructure.
Recent developments in AI offer new opportunities for supporting this complex optimisation process in architectural design and engineering. Big data from the built environment are ubiquitous, in the form of images, videos, and sensory measurements of structural health. Moreover, the computing power for data processing is growing. Harnessing these opportunities, AI provides us with unprecedented capabilities to analyse, optimise and automate the different phases of decision-making in the built environment.
At AiDAPT, data-driven intelligence and model-based engineering come together to support long-term, adaptive, and evidence-based abstraction and synthesis of structural and architectural choices, towards a more sustainable and resilient built environment. We study and develop state-of-the-art machine-learning and deep learning methods from automatic recognition of visual architectural elements to the dynamic feature extraction, inference, optimisation and autonomous decision-making under uncertainty.