ZeroWaste — Towards Computing Cooperative Robotic Sequences for the Disassembly and Reuse of Timber Frame Structures.

Abstract

ZeroWaste is a project about repositioning existing timber building stock within a circular economy framework. Rather than disposing of these buildings at the end of their life, the goal is to view them as stores of valuable resources that can be readily reused. By doing this, material life cycle becomes an integral design consideration alongside planning for the efficient disas- sembly and reuse of these structures. In this paper, the computational workflow is presented for the first phase of the project: planning a cooperative robotic disassembly sequence for the scaf- fold-free removal of members from existing timber structures. A pavilion-scale prototype is first constructed, in the Embodied Computation Lab at Princeton University, to represent an existing timber structure built according to conventional North American stick frame construction prac- tices. A multi-directed graph data structure, representing structural member connectivity and support hierarchy, is then coupled with a breadth-first search algorithm to plan potential scaf- fold-free robotic disassembly sequences given a member removal target. In parallel, computer vision is integrated and implemented through the robotic setup to create an accurate as-built point cloud scan of the whole structure. This as-built information is then used to inform the evaluation of potential robotic sequences from the point of view of robotic reachability and structural performance. This work-in-progress paper first presents a high-level overview of the various components in this workflow, followed by its demonstration in planning the removal of a specific member in the prototype structure. Upcoming project developments will include the planning, and physical demonstration, of more complex disassembly sequences, coupled with reassembly and reuse of the removed members for various regions of the prototype structure.