Data Model for Adaptive Robotic Construction in Architecture

Abstract

Robotic fabrication plays an essential role for the construction industry in the pursuit of higher productivity and greater flexibility. Modular robotic systems enable the off-site prefabrication of components for high-performance building systems from wood or fiber composites, while digital planning methods and computational design enable the efficient planning of highly specialized and unique lightweight structures. However, existing manufacturing systems do not provide an interface between the two domains that allows to utilize their combined potential. Manufacturing instructions for the individual components are currently generated directly from the design model and don’t consider the precise setup and configuration of the modular system. The parametric description in computational design thus ends in the planning phase and is converted into a non-parametric description in native robot code, losing a lot of information in the process. For a continuous description of the building components from planning to fabrication, a data model for the parametric description of the fabrication tasks is therefore needed. We present and such a data model for timber and fiber prefabrication applications. The model allows the generic description of the manufacturing process within a task graph, which gives the manufacturing system a way of adapting the tasks to its specific configuration and setup. In addition to the description of jobs, tasks and actions, the model also includes a description of the considered manufacturing environment, as well as a monitoring component for the acquisition of sensor values and system configurations for individual tasks. This not only allows the identification of manufacturing issues like collisions ahead of time, but also enables the implementation of a feed-back loop for iterative optimization of the paths for any given environment. In this work, we give the design considerations for this data model.