Sergio Mutis 的作品 Diffusive Habitats 可重构建筑系统

设计师 Sergio Mutis 的作品 Diffusive Habitats 可重构建筑系统荣获了极具声望的 2023 A' Design Award 铜奖，竞赛类别为生成式、算法、参数化与人工智能辅助设计。

铜生成式设计· 2023

Diffusive Habitats 可重构建筑系统

设计师： Sergio Mutis

生成式设计类别A' Design Award 铜奖获奖者

78,281

已验证的获奖者

验证 #145846 的获奖状态

访问网站在A' Design Award上查看新闻稿获奖徽章

色彩

设计作品

关于 Diffusive Habitats 可重构建筑系统

《扩散式栖居》(Diffusive Habitats)是一套融合分布式机器人技术、强化学习与在线平台的建筑系统,使空间能够随时间推移不断重新配置。机器人单元通过搬运模块化构件来改变栖居空间的布局。模块化构件采用数控铣削层压胶合板制成,并嵌入3D打印的连接接口。机器人原型则以再生塑料3D打印而成。该平台依据居住者需求、结构完整性与环境性能来组织空间,旨在减少拆除行为,让居所更具灵活性。

荣誉认可

关于A' Design Award 铜奖

Diffusive Habitats 可重构建筑系统是 A' Design Award 铜奖得主，这是授予以创造力、独创性和敏锐见长的杰出设计的极具声望的认可。铜奖得主是技艺娴熟、创意惊人的作品，印证了其设计师丰富的经验、想象力与机智。预计Diffusive Habitats 可重构建筑系统将成为一件备受推崇、令人钦佩且被珍爱的物品，兼具显著的专业性与创新性。这件赢得 A' Design Award 铜奖的设计融入艺术、科学、设计和技术领域的最佳实践，并因注重细节而出类拔萃。Diffusive Habitats 可重构建筑系统出色地满足需求，触动渴望，迷住情感，并吸引好奇心。

设计幕后

媒体资料包

由设计师提供的致谢、团队成员及技术详情。

独特属性: 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.
尺寸与包装: 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.
标签: Living Architecture, Distributed Mechatronics, Reinforcement Learning, Multi-Agent Systems, Generative Architecture
生产技术: 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.
设计挑战: 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.
研究: 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.
灵感: 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.
项目周期: 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.
操作流程: 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.
团队成员: Sergio Mutis
Eric Hughes
Faizunsha Ibrahim Ghousiaa
Garyfallia Papoutsi.
图片版权: 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