CONCEPT
AM2PM
AM2PM envisions a groundbreaking approach to constructing multistorey buildings through an integrated digital and sustainable system. By combining human-robot collaboration, AI-driven design, and Digital Twin technology, our project aims to streamline construction with a focus on efficiency, safety, and environmental impact.
Our unified infrastructure brings together 3D printing, sustainable materials, and computational design to reduce material use by 50%, potentially saving up to 29 million tons of carbon annually. In line with EU climate goals, AM2PM offers a scalable, cost-effective solution that could save over €11 billion each year in construction, leading the way to a more sustainable future for the industry.
AM2PM Digital Twin Architecture Scheme
The AM2PM project designs, tests, and implements an AI-enabled Digital Twin (DT) infrastructure for predictive manufacturing. This infrastructure optimizes building design, material composition, and manufacturing processes across all construction stages. The system transforms construction sites into robotized Cyber-Physical Systems (CPS), enabling dynamic, bi-directional communication between digital and physical systems (robotic, sensing, and processes) and external services.
The DT infrastructure facilitates real-time feedback loops for autonomous Learning-by-Printing (LbP), enabling 3D concrete printing (3DCP) and orchestrating multi-robot construction. It also integrates edge computations, which inform predictive LCA, transfer learning, and automated data analysis. The System-of-Systems approach connects end-users, software applications, and physical systems via a decentralized DT platform that supports data storage, communication, actuation, and orchestration.
Key features of the architecture include:
1. Sustainable Materials: Development of 3DCP cementitious mixes from recycled waste with a 50% reduction in design time.
2. Computational Design: Optimized structural components with 50% less material use and embodied carbon.
3. AI & Feedback: LbP-based AI methods for automated predictions and closed-loop process control.
4. Dynamic DT Models: Real-time construction site representations for robotic planning and management.
5. Human-Robot Interaction: Proactive site management and lean production for large-scale 3DCP deployment.
6. Predictive LCA: Integrated tools for early-stage planning, environmental impact analysis, and resource efficiency.
This holistic approach breaks the linearity of current 3DCP workflows, creating a dynamic, interactive, and sustainable construction ecosystem. The DT infrastructure serves as the backbone for data exchange and communication across software and physical components, bridging digital and physical domains.
Plausibility of the methodology
The research follows an iterative approach, where material synthesis, structural optimization, and predictive AI methods are designed, evaluated, and optimized in relation to each other. This process gradually forms an integrated construction method and computational interface.
The research cycle consists of four phases:
1) Problem Investigation, which identifies areas for improvement
2) Treatment Design, which develops potential solutions
3) Treatment Validation, where the design is tested for effectiveness
4) Implementation Evaluation, which assesses the solution and begins the next research cycle.
The design and analysis are conducted in 6-month sprints, incorporating brainstorming, divergent design, convergence, and evaluation. This research will be implemented using advanced laboratory infrastructures at institutions like Technion, TUM, DTU, TU/E, and WASP SRL Italy, which provide a framework for testing digital fabrication and computational technologies.