Papers – Robotic Building (RB)

Computer Vision and Human–Robot Collaboration Supported Design-to-Robotic-Assembly

Abstract
While half of all construction tasks can be fully automated the other half relies to a certain degree on human support. This paper presents a Computer Vision (CV) and Human–Robot Interaction/Collaboration (HRI/C) supported Design-to-Robotic-Assembly (D2RA) approach that links computational design with robotic assembly. This multidisciplinary approach has been tested on a case study focusing on urban furniture and involving experts from respective disciplines and students.

SpringerLink

MULTI-SPECIES AND MULTI-SCALAR URBAN TOPOLOGIES (PREPRINT)

Abstract
Over half of the world’s population lives in urban areas, with an increase to 68% by 2050 (UN report, 2018) and their ecosystems are in need of repair and restoration. Hence, understanding urban ecosystems across different scales and knowledge domains is vital to global sustainability. Presented research focuses on multispecies interactions, in the context of shifting environmental conditions due to climate change. It introduces adaptive strategies that integrate landscapes, built environments, and multi-species infrastructures. In this work, strategies and knowledge domains of sustainability, ecology and species diversity are combined with architecture, computational design, advanced manufacturing, and material systems performance. By investigating multi-species across scales ranging from micro to macro and by designing new topologies, urban ecosystems are established that are serving equally all species. In this context, computational design, simulation, analysis, and advanced manufacturing allow evaluation, understanding, and forecasting trends and consequences of connected impacts across multiple knowledge domains and scales as well as developing and operating sustainable ecosystems.

Full-text PDF

Aspects of Ambient UX Design Within Design-to-Robotic-Production and -Operation Processes

Abstract
Ambient User Experience design in architecture implies consideration of various intersecting and sometimes overlapping design fields such as interaction and architectural design with the aim to achieve a continuous and cohesive user experience across devices, time, and space. In this paper, Ambient User Experience design is explored in relation to Design-to-Robotic-Production and -Operation processes developed at TU Delft, which link computational design with robotic production and operation. Several case studies involving the integration of sensor-actuators into the built environment are discussed with respect to mapping activities through constraints and enablers and designing with the time as a variable.

SpringerLink

Developing Responsive Environments based on Design-to-Robotic-Production and -Operation Principles

Abstract
The development of physical and computational mechanisms aimed at augmenting architectural environments has been one of the foci of research implemented at the Faculty of Architecture and the Built Environment, Delft University of Technology (TUD) for more than a decade. This paper presents the integration of distributed responsive climate control into the built environment based on Design-to-Robotic-Production and -Operation (D2RP&O) principles. These connect computational design with robotic production and operation of buildings. In the presented case study structural elements meet load bearing as well as functional requirements. Their spatial arrangement creates variable densities for accommodating sensor-actuators that are operating heating and cooling. This mechatronic operation relies on activity recognition for achieving responsive climate control in the built-environment.

ResearchGate

A Wearable Interface in a Design-to-Robotic-Production and -Operation Development

Abstract
This paper presents the development of an Internet of Things (IoT) wearable device as a personal interfacing node in a high-resolution intelligence framework, which is informed by Design-to-Robotic-Production and -Operation (D2RP&O) principles developed at Delft University of Technology (TUD). The device enables the user to act as an active node in the built-environment’s underlying Wireless Sensor and Actuator Network (WSAN), thereby permitting a more immediate and intuitive relationship between the user and his/her environment, where this latter is integrated with physical / computational adaptive systems and services. Two main resulting advantages are identified and illustrated. On the one hand, the device’s sensors provide personal (i.e., body temperature / humidity, physical activity) as well as immediate environmental (i.e., personal-space air-quality) data to the built-environment’s embedded / ambulant systems. Moreover, rotaries on the device enable the user to override automatically established illumination and ventilation settings in order to accommodate user-preferences. On the other hand, the built-environment’s systems provide notifications and feedback with respect to their status to the device, thereby raising user-awareness of the state of his/her surroundings and corresponding interior environmental conditions. In this manner, the user becomes a context-aware node in a Cyber-Physical System (CPS). The present work promotes a considered relationship between the architecture of the built-environment and the Information and Communication Technologies (ICTs) embedded and/or deployed therein in order to develop highly effective alternatives to existing Ambient Intelligence (AmI) solutions.

ResearchGate

Deep learning object-recognition in a design-to-robotic-production and -operation implementation

Abstract
This paper presents a new instance in a series of discrete proof-of-concept implementations of comprehensively intelligent built-environments based on Design-to-Robotic-Production and -Operation (D2RP&O) principles developed at Delft University of Technology (TUD). With respect to D2RP, the featured implementation presents a customized design-to-production framework informed by optimization strategies based on point clouds. With respect to D2RO, said implementation builds on a previously developed highly heterogeneous, partially meshed, self-healing, and Machine Learning (ML) enabled Wireless Sensor and Actuator Network (WSAN). In this instance, a computer vision mechanism based on open-source Deep Learning (DL) / Convolutional Neural Networks (CNNs) for object-recognition is added to the inherited ecosystem. This mechanism is integrated into the system's Fall-Detection and -Intervention System in order to enable decentralized detection of three types of events and to instantiate corresponding interventions. The first type pertains to human-centered activities / accidents, where cellular- and internet-based intervention notifications are generated in response. The second pertains to object-centered events that require the physical intervention of an automated robotic agent. Finally, the third pertains to object-centered events that elicit visual / aural notification cues for human feedback. These features, in conjunction with their enabling architectures, are intended as essential components in the on-going development of highly sophisticated alternatives to existing Ambient Intelligence (AmI) solutions.

doi:10.1109/ETCM.2017.8247495

The use of variable stiffness joints in adaptive structures

Abstract
Adaptive structures are defined here as structures capable of counteracting actively the effect of external loads via controlled shape changes and redirection of the internal load path. These structures are integrated with sensors (e.g. strain, vision), control intelligence and actuators. This paper investigates the use of variable stiffness joints in adaptive structures to achieve large shape changes. Large shape changes are employed as a structural adaptation strategy to counteract the effect of the external load. The structure is designed to 'morph' into optimal shapes as the load changes. This way the stress can be homogenized avoiding peak demands that occur rarely. Numerical results show that when large shape changes are considered, material mass (and thus embodied energy) reduction is achieved with respect to both adaptive structures limited to small shape changes and optimised passive structures. Embodied energy savings become substantive when shape changes are allowed to go beyond conventional deflection limits. However, large shape changes require significant flexibility of the joints because their fixity can affect load-path and shape control. To address this problem, a variable stiffness joint is proposed. During shape/load-path control, the joint reduces its stiffness so that required deformation patterns can be achieved with low actuation energy. After shape control the joint recovers rigidity. Experimental studies are presented to show the potential for application of joints with variable stiffness in adaptive structures.

http://repository.tue.nl/870046

A high-resolution intelligence implementation based on Design-to-Robotic-Production and -Operation strategies

Abstract
This paper presents an initial proof-of-concept implementation of a comprehensively intelligent built-environment based on mutually informing Design-to-Robotic-Production and -Operation (D2RP&O) strategies and methods developed at Delft University of Technology (TUD). In this implementation, D2RP is expressed via deliberately differentiated and function-specialized components, while D2RO expressions subsume an extended Ambient Intelligence (AmI) enabled by a Cyber-Physical System (CPS). This CPS, in turn, is built on a heterogeneous, scalable, self-healing, and partially meshed Wireless Sensor and Actuator Network (WSAN) whose nodes may be clustered dynamically ad hoc to respond to varying computational needs. Two principal and innovative functionalities are demonstrated in this implementation: (1) costeffective yet robust Human Activity Recognition (HAR) via Support Vector Machine (SVM) and k-Nearest Neighbor (k-NN) classification models, and (2) appropriate corresponding reactions that promote the occupant's spatial experience and wellbeing via continuous regulation of illumination with respect to colors and intensities to correspond to engaged activities. The present implementation attempts to provide a fundamentally different approach to intelligent built-environments, and to promote a highly sophisticated alternative to existing intelligent solutions whose disconnection between architectural considerations and computational services limits their operational scope and impact.

doi:10.22260/ISARC2017/0014

Kite-powered design-to-robotic-production for affordable building on demand

Abstract
Building technologies employed today in 2nd and 3rd world countries are imported, expensive, outdated and unsustainable. Highly developed countries, on the other hand, rapidly advance in developing affordable, numerically controlled and robotically supported material- and energy-efficient methods for building on demand. The research team proposes to close this gap by applying advanced design-to-robotic-production (D2RP) technologies developed at Technical University Delft (TUD) to construction problems in 2nd and 3rd world countries. The provided tool base uses refurbished robotic technology, which is retrofitted with state-ofthe-art open source control software, and by employing local approaches and available materials the dependency on imported materials and processes is drastically reduced. The D2RP unit is coupled with the electricity generating Kite Power (KP) system developed at TUD to create a mobile sustainable autarkic unit that can be deployed everywhere.

doi:10.22260/ISARC2017/0078

Adaptive building-skin components as context-aware nodes in an extended cyber-physical network

Abstract
This paper presents an adaptive building-skin system that attempts to establish the foundations for an intuitive and responsive interface between interior and exterior spaces with respect to environmental, thermal, acoustic, and user-comfort considerations. It does this by enabling each of its components to act as individual, context-aware, sensor-actuator nodes capable of differentiated - yet correlated - actions, reactions, and interactions. The proposal situates the system within an intelligent environment whose ecosystem's operational scope subsumes yet extends beyond interior environments to include exterior domains via wearable devices. Accordingly, as the sensed data of any device is accessible across all devices in a topology of meshed nodes, the computationally processed behavior of any node is potentially informed by and informing of the status of individual and/or sets of other nodes. In this manner, the building-skin is not construed as a mere envelope, but rather as a system comprised of agents that, in conjunction with all other embedded, ambulant, or wearable agents, actively promote the well-being, comfort, and spatial experience of users.

doi:10.1109/WF-IoT.2016.7845436

Towards embedding high-resolution intelligence into the built environment

Abstract
Prevailing architectural design paradigms, identified as those informed by historically conservative positions and methods, are incompatible with the intelligent built-environment discourse. Two core considerations inform this assessment. The first asserts that such paradigms produce spaces and programmatic distributions in terms of discrete, precisely delimited, and artificially ordered static partitions. The second asserts that said paradigms preclude (at best) or exclude (at worst) discussions of technological intelligence from the early stages of the design process, thereby negating the possibility of imbuing the built-environment with inherent intelligence. The rigidity expressed in the first consideration, and the disregard for technological intelligence expressed in the second, produce very low-resolution and-adaptability architectures. As a result , occupants are compelled to conform to their built-environment rather than the expected vice versa, as it is fundamentally incapable of actively, reactively, and interactively promoting their well-being. In this paper, two key positions (i.e., high resolution space and high resolution intelligence) motivated by the above considerations are promoted as part of a fundamentally different design paradigm, one expressly geared towards personalization, interaction, and intelligence in a parametrically fluid and self-adapting built-environment capable of intuitive physical, spatial, and computational feedback-loops.

ResearchGate

An Extended Ambient Intelligence Implementation for Enhanced Human-Space Interaction

Abstract
This paper proposes an extended Ambient Intelli-gence (AmI) solution that expresses intelligence with respect to both Information and Communications Technologies (ICTs) and spatial reconfiguration in the built-environment. With respect to the former, a solution based on a decentralized yet unified Wireless Sensor Network (WSN) is proposed. This is deployed across exterior, interior, and wearable domains, equipped with heterogeneous platforms across em-bedded and ambulant nodes, and open to a variety of proprietary and non-proprietary communication protocols. With respect to the latter, a corresponding functionally and physically reconfigurable built-environment pertinent to the Adaptive Architecture discourse is revisited. The ICTs component aims to demonstrate the advantages of a cohesive and in-teroperable heterogeneity distributed along local and web-based proprietary and non-proprietary services over a prevalent locally based homogeneity with re-spect to both development platforms and communica-tion protocols in a WSN. The architectural compo-nent aims to demonstrate that a highly adaptive and transformable built-environment is better suited to complement and to sustain assistive as well as inter-ventive services enabled by said WSN. As a unified solution, the proposal showcases that the merging of technological and architectural considerations in the design of an intelligent environment enables more intuitive solutions that actively adapt to, interact with, intervene on the user to promote comfort and well-being via computational as well as physical feedback-loops.

doi:10.22260/ISARC2016/0094

Materially Informed Design to Robotic Production: A Robotic 3D Printing System for Informed Material Deposition

Abstract
This paper presents and discusses the development of a materially informed Design-to-Robotic-Production (D2RP) process for additive manufacturing aiming to achieve performative porosity in architecture at various scales. An extended series of experiments on materiality employing robotic fabrication techniques were implemented in order to finally produce a prototype on one-to-one scale. In this context, design materiality has been approached from both digital and physical perspectives. At digital materiality level, a customized computational design framework has been implemented for form finding of compression only structures combined with a material distribution optimization method. Moreover, the chained connection between the parametric design model and the robotic production setup has enabled a systematic study of specific aspects of physicality that cannot be fully simulated in the digital medium. This established a feedback loop for not only understanding material behaviours and properties but also robotically deposit material in order to create an informed material architecture.

doi:10.1007/978-3-319-26378-6_27

Informed Design to Robotic Production Systems; Developing Robotic 3D Printing System for Informed Material Deposition

Abstract
This paper discusses the development of an informed Design-to-Robotic-Production (D2RP) system for additive manufacturing to achieve performative porosity in architecture at various scales. An extended series of experiments on materiality, fabrication and robotics were designed and carried out resulting in the production of a one-to-one scale prototype. In this context, design materiality has been approached from both digital and physical perspectives. At digital materiality level, a customized computational design framework is implemented for form finding of compression only structures combined with a material distribution optimization method. Moreover, the chained connection between parametric design model and robotic production setup has led to a systematic study of certain aspects of physicality that cannot be fully simulated in the digital medium, which then establish a feedback loop for underrating material behaviors and properties. As a result, the D2RP system proposes an alternative method of robotic material deposition to create an informed material architecture

http://resolver.tudelft.nl/uuid:02bf0811-41f0-4ac9-a270-af796fc77c7c

PAPERS / CHAPTERS