Convertible furniture design

Abstract

Convertible furniture is used widely because of its multi-purpose and space-saving properties, but substantial professional skills are required to design practical convertible furniture. In this study, we introduce a system that allows unskilled users to design convertible furniture objects. By using a set of desired furniture units as inputs, our system generates a group of suggested convertible furniture objects, which can collapse into a compact form. The geometric problem involves connecting the furniture units via junctions so they occupy the smallest space in their compact form while still fulfilling the desired function in the expanded form. This optimization problem is non-trivial because of the very large discrete search space of the connected structures. Furthermore, there is a huge continuous space of geometric parameters where the connected units make contact due to the complex dependencies of their movements. We propose an efficient algorithm based on a tree selection approach and pre-computed junction-dependent configuration spaces to accelerate the optimization process in a high-dimensional, mixed discrete–continuous search space. We evaluated our system based on a wide range of inputs and a series of convertible furniture objects were generated. The experimental results demonstrate that our system can provide design suggestions as well as inspiring creativity among users.