Mixed Substrate Computation. Sensor Based Artificial Cognition for Architectural Design and Modification

Doktorsavhandling, 2016

A mutual relationship exists between activities and their physical environment. Change of circumstances in the built environment cause change of activities, and vice versa, change of activities cause changes of physical environment. How does information originating from activities influence the environment? And how does the environment provide relevant information for activities to take place? Digital and material computation has — within architectural design — been used extensively to strengthen the capacity to build novel and more geometrically enhanced structures. However on large and small physical as well as long and short temporal scale there is only poor understanding the activities and phenomena taking place in the built environment. In particular, a lack in understanding of the relationships between changes in physical conditions and changes of activities. The rationale behind implementations and modifications of the built environment is constituted by many actants simultaneously at play, mainly based on human heuristics, sensemaking and semantics. Digital computation can be combined with material, morphological and other types of computation to create models of the past, current and future scenarios. Here this concept is coined !Mixed !Substrate !Computation and relies heavily on successful !embodiment and !embodied !computation. A technology is needed for tracing, extracting and computing both embodied memory and data from activities residing in the environment on different spatial and temporal scales. This thesis presents a set of methods that combine sensors and algorithms to a novel technique of perceiving activities and phenomena over time.. Consequently, a kind of artificial cognition is demonstrated able to detect recurrent phenomena and in turn perform predictions in seemingly chaotic situations. This discovery can bring about a paradigm shift in design, taking us from our current situation where architects, designers and planners predict and design for future needs using their present day point-of-view, to a situation where design tools are able to learn from complex situations and can predict future needs autonomously. This capacity for perception and prediction contributes to the current discourse on mixed material and digital practices within architectural design — filling an increasingly widening gap between material and formal computation: the concept of embodiment. The idea of integrating body and soul—or the physical and the abstract—is a concept which is key for understanding the relationships between phenomena and matter.