The AADRL is a post-professional MArch (Architecture & Urbanism) graduate design programme at the Architectural Association School of Architecture, in London. Led by Theodore Spyropoulos, it has four research labs that respond to an umbrella agenda that the team set for a period of 3-4 years. Most recently they have been working on a theme of ‘Behavioral Complexity‘, which explores design that is proto-typical, scenario driven and examines behaviors through design enquiry. A feature of this research agenda between the four labs is examining robotics within architecture. They have two studios that are exploring 3D printing at the scale of buildings, one augmenting robotic arms by developing custom end effectors and the other exploring drones and swarm printing and finally the fourth studio is looking at parametric approaches towards kinetic architecture.
The following is the work from AADRL Spyropoulos Design Lab, which has been exploring behavior based design systems that are self-aware, mobile, and self-structure and assemble. The research explores high population mobility agents that evolve design that moves beyond finite conceptions of space towards one that co-evolves and adapts. The research looks at developing strategies of intelligent design units that through organisation develop body plans that afford new behaviors to emerge. An example of this is OWO project below which as a singular unit is conceptualised as a unit that through embedded pneumatics allows the unit to expand / contract, curl, and roll. With respect to mobility as an individual unit it can roll in its mobility mode, but for example a unit connects to two other units then the body plan allows for a tripod configuration, which allows the creature to walk.
The primary modes of that we develop our digital and analogue research are 1.) mobility 2.) self-structuring . the thesis take the position that every unit in its simple form (singular) should have the capacity to be mobile and search for other units. Our Self-structuring mode is higher level organsiations that are collective in principal and allow units to construct spatial structures. Driving this is evolutionary algorithmic strategies that take into account the physical constraints of the designed unit. The projects do not have a blueprint that they follow and they do not have secondary scaffolding systems to assist. The interactions are designed through playpen behaviors that are learned through iterative interactions. We use evaluation criteria like load paths to allow for reconfigurations after a stable form is achieved and in general it is formative strategy.
The physical prototypes are use various actuation strategies, some use small scale hydraulic pistons, other soft robot strategies through custom silcone casting, servoes and electromagnets. The software platforms include Processing / openFrameworks, Maya, Miarmy, along with Arduino micro controlling, camera tracking etc. They generally use voxel based computational strategies for most of the digital generation and communication.
The research references for this work looks to research of Hod Lipson, Karl Sims, and John Henry Holland’s research with adaptive systems. All the projects develop a computational back end that examines evolutionary algorithmic design principles that the lab explores.