Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha by Sergio Mutis

Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha by Sergio Mutis is a highly esteemed winner of the very prestigious 2023 Moputso oa Boronse oa A' Design Award in Moralo wa Tlhahiso, Algorithimi, Paramethrike le AI Competition Category.

BoronseTlhahiso· 2023

Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha

by Sergio Mutis

Moputso oa Boronse oa A' Design Award Winner in Tlhahiso Category

77,941
Verified Award Winner
Verify Award Status for #145846

Visit Site View at A' Design Award Press Release Award Badge

Colors

The Design

About Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha

Diffusive Habitats ke mokgwa wa boqapi ba meaho o kopanyang di-robotiki tse hasaneng, ho ithuta ka moputso (reinforcement learning) le sethaleng sa inthanete e le hore dibaka li ka fetoloa bocha ha nako e ntse e tsamaya. Diyunithi tsa robotiki di sisinya dikarolo tse ka fetolwang (modular parts) ho fetola tlhophiso ya bodulo. Dikarolo tsena tse ka fetolwang di entswe ka polopo e patikaneng e sebetsoang ka mochine wa CNC (CNC-milled laminated plywood), e nang le didikgokahano tse hatisitsweng ka 3D tse kentsweng ka hare. Mefuta ya mathomo ya robotiki e hatisitswe ka 3D ka polasetiki e sebedisitsweng hape. Sethala sena se hlophisa sebaka ho ipapisitswe le ditlhoko tsa baahi, matla a kaho (structural integrity) le tshebetso ya tikoloho. Sepheo ke ho fokotsa ho heletswa ha meaho le ho etsa hore malapa a be a fetohang ha bonolo.

The Recognition

About the Moputso oa Boronse oa A' Design Award

Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha ke mohapi oa Moputso oa Boronse oa A' Design Award, kananelo e hlomphehang haholo e fuoang meralo e ikhethang e tsejoang ka boqapi, bohlale ba ho qapa le bohlale. Bahapi ba Boronse ke lintho tse nang le bokhoni ka ho felletseng le boqapi bo hlollang tse tiisang boiphihlelo bo bongata, mohopolo le bokhoni ba ho iphumana ba baqapi ba tsona. Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha e lebelletsoe ho ba ntho e hlomphuoang haholo, e ratoang le e ratehang, e nang le boprofeshenale bo bonahalang le mehopolo e mecha. Moralo ona o hapileng Moputso oa Boronse oa A' Design Award o kenyelletsa mekhoa e metle ka ho fetesisa bonong, mahlaleng, moralong le theknolojing mme o hlahella ka ho tsepamisa maikutlo lintlheng tse nyenyane. Diffusive Habitats Mokgwa wa Boaho o Kang o Ka Fetoloa Bocha e kutloisisa litlhoko hantle, e ama litakatso, e hapa maikutlo mme e tsosa takatso ea ho tseba.

Behind the Design

Press Kit

Credits, team members and technical details provided by the designer.

Unique Properties: Diffusive Habitats (DH) is a resilient architectural system that builds upon the juncture of distributed robotics, reinforcement learning, and digital platforms to reimagine the way architecture is conceived and inhabited. Challenging traditional praxis, the project embraces constant spatial reconfiguration and distributed ownership. And, via the interplay of applied transdisciplinary research, DH critically explores a spatial system capable of self-assessment, organization, and improvement.
Dimensions & Packaging: Robot Prototypes 3.0.
Dimensions: Width 100mm x Depth 100mm x Height 280mm.
Composition: 3D-Printed recycled plastic (PLA). Weight: 420g. Parts: 5 links of 3D-printed PLA, 8 neodymium magnets 20dx3, 4 Dynamixel AX-12A motors, 1 LiPo 3cell Battery, 1 U2D2 Power Hub, 1 Raspberry Pi Zero.

// Material Prototypes 3.0.
Standard Dimensions: Width 140mm x Depth 420mm x Height 140mm. Standard Weight: 477g.
Composition: CNC-milled laminated plywood with embedded 3D-printed PLA connection interfaces.
Tags: Living Architecture, Distributed Mechatronics, Reinforcement Learning, Multi-Agent Systems, Generative Architecture
Production Technology: The DH system consists of multiple correlated elements that were heuristically prototyped, tested, and redesigned. Its space generation platform was scripted in C-sharp for Unity3D with the ML Agents U3D reinforcement learning framework. Laminated plywood prototypes were CNC-milled and assembled with friction joints. 3D-printed PLA robotic prototypes were developed with a Unity3D digital twin environment, wireless Raspberry Pi communication, Dynamixel AX-12A motors, and Optitrack motion capture.
Design Challenge: Besides technological and transdisciplinary challenges, the shift in authorship paradigm was perhaps the greatest obstacle in the development of the project. Rather than the desires of individual designers, this research required a heuristic process of trial and error, where the design was the result of cascading iterations of parameter optimization. Once the team embraced that fundamental idea, the once slow and erratic project advanced swiftly beyond its initial research objectives.
Research: Diffusive Habitats employs systematic heuristic research, with reinforcement learning and self-play frameworks for intelligence development. The main research objective was to explore an architectural system capable of self-organization, assessment, and improvement, that would embrace technological advancements to reimagine how we conceive and inhabit spaces. To do so, the project adopted a method of small attainable mechatronic, computational, material, and intelligence milestones that were quickly designed, tested, and analyzed to inform an upcoming iteration. Data collection and analysis included: weight, task success rate, and cost for robotic design / multi-requirement satisfaction for computational generative design / weight, open source fabrication index, and sustainability for material research / task success rate, speed, and energetic efficiency for robotic intelligence. Tests were carried out via simulations in Unity and C# for computational algorithms and robotic intelligence, and via physical tests for mechatronic and material development. Initial results were systematically improved by following iterations leading up to a mechatronic design with diverse capabilities (particularly while collaborating), a computational framework that could organize space under numerous requirements far better than any one designer, a material unit that proved capable of forming multiform structures (though not yet capable of being employed in buildings over 2 stories high), and a robotic intelligence that could calculate, optimize, and execute large reconfiguration tasks.

The project still requires development to be thoroughly applied in architectural practice. Yet, within the constraints of a 1-year Master's degree project, it was considered a vast success and will be used to attract interest and funds to further the research. The effects of this investigation range from avoiding demolition and boosting the sustainability of the built environment, to transforming rigid urban dwellings into flexible hubs for the ever-accelerating change of our times. Altogether, this research explores a radically novel way of living that could be a widespread reality in the future, or allow for new conversations to occur that will ultimately improve the way we live.
Inspiration: Unlike other crafts, arts, and industries, architecture has not seen significant developments during the upheaval of the digital revolution. Traditional praxis continues to produce buildings incapable of substantial programmatic flexibility, that become obsolete within the ever-accelerating change of the times, and quickly end up in queue for demolition. Diffusive Habitats challenges this unsustainable practice, instrumentalizing its technological juncture in the lookout for renewed solutions.
Project Duration: The project was developed as part of the Architectural Design M.Arch. at UCL in London, UK. It started in 2021 and culminated in late 2022 at Bartlett B-Pro Show Expo, where it received the Bartlett AD Gold Medal. The case-study is located in Brixton, London, and the design team comes from India, Colombia, the US, and Greece.
Operation Flow: Diffusive Habitats operates via an online platform facilitating plot search, habitat crowdfunding, collaborative design, and spatial reconfiguration. Behind its intuitive interface, the platform employs sophisticated algorithms to organize space based on inhabitant requirements, structural integrity, and environmental optimization. With it, a community is able to design their ideal habitat, see it constructed by robots, and adjust it to account for changing individual and collective needs.
Team Members: Sergio Mutis
Eric Hughes
Faizunsha Ibrahim Ghousiaa
Garyfallia Papoutsi.
Image Credits: Image #1: Creator Eric Hughes, DH Little Girl Rendering, 2022
Image #2: Creator Sergio Mutis, DH Reinforcement Training Arena, 2022
Image #3: Creator Eric Hughes, DH Skylight Rendering, 2022
Image #4: Creator Sergio Mutis, DH Robotic Reconfiguration Studies, 2022
Image #5: Creator Eric Hughes, DH Speculative Urbanism, 2022