Self-organizing, -adaptive, and hybrid robotic control in architectural technology

Robotic control for construction automation and building adaptation represents a quickly developing field of research, spanning applications in several sectors. Embedded mechatronics or smart building systems may sense and respond to complex systems such as climate, city dynamics, or occupancy dynamics, while on-site construction robots may operate in strongly unstructured environments. In such complex or highly stochastic conditions, certain tasks stand to benefit from self-organized or self-adaptive robotic control and communication, which can offer attractive features like robustness in uncertain conditions. The process of designing for and constructing with self-organized control has thus far been underexplored in architectural research. As these technologies inherently involve nondeterministic behaviors, or some degree of random events, their use will require new approaches from those currently dominant in architecture, engineering, and construction sectors, which largely depend on predetermined outcomes. This thesis develops approaches for design and implementation of architectural technology that uses self-organizing or self-adaptive robotic control, inclusive of hybrid systems. It investigates methods for steering such engineered systems in simulation and reality setups, using approaches from evolutionary robotics, machine learning, interactive evolution, and complex systems science. Development of the thesis proceeds within the context of a bio-hybrid robotics project, which explores collaborative decision-making between plants, robots, and human occupants. Within the broader field of architecture, the thesis contributes to key topics of bio-hybrids, self-organization, complex systems analysis, high-level objectives for low-level control, and the incorporation of human judgment within real-time adaptation. The individual studies are interdependently related as mechanisms within a posited “self-building” framework, which encompasses self-organizing, -adaptive, and hybrid robotic control within both construction automation and embeddedmechatronics in buildings and infrastructure.