Dry Stacking Strategies for Automated Construction with Irregular Objects
[ Watch Movie ]
We propose a ”next best step” planning method that autonomously chooses assembly actions to allow construction with objects that contain a significant amount of variation.
Raw Material Dataset
We generated synthetic datasets to test our assembly planning method.
A
2D simulation framework
to test the proposed planning algorithm was developed.
Planning Algorithm
The use of physics and geometric analysis combined with a heuristic approach has led us to develop an algorithm that functions well for a distribution of irregular objects with relatively high irregularities.
A strategy that includes heuristics from masonry manuals.
A structure built in the simulation framework using the proposed algorithm.
Experiments
The results demonstrate a feasible planning approach for building dry stacked structures that functions with reasonably high values of irregularity in the object shape distribution.
It also highlight the close relationship between the planner and the object types they operate on.