Abstract
Additive manufacturing methods1,2,3,4 using static and mobile robots are being developed for both on-site construction5,6,7,8 and off-site prefabrication9,10. Here we introduce a method of additive manufacturing, referred to as aerial additive manufacturing (Aerial-AM), that utilizes a team of aerial robots inspired by natural builders11 such as wasps who use collective building methods12,13. We present a scalable multi-robot three-dimensional (3D) printing and path-planning framework that enables robot tasks and population size to be adapted to variations in print geometry throughout a building mission. The multi-robot manufacturing framework allows for autonomous three-dimensional printing under human supervision, real-time assessment of printed geometry and robot behavioural adaptation. To validate autonomous Aerial-AM based on the framework, we develop BuilDrones for depositing materials during flight and ScanDrones for measuring the print quality, and integrate a generic real-time model-predictive-control scheme with the Aerial-AM robots. In addition, we integrate a dynamically self-aligning delta manipulator with the BuilDrone to further improve the manufacturing accuracy to five millimetres for printing geometry with precise trajectory requirements, and develop four cementitious–polymeric composite mixtures suitable for continuous material deposition. We demonstrate proof-of-concept prints including a cylinder 2.05 metres high consisting of 72 layers of a rapid-curing insulation foam material and a cylinder 0.18 metres high consisting of 28 layers of structural pseudoplastic cementitious material, a light-trail virtual print of a dome-like geometry, and multi-robot simulations. Aerial-AM allows manufacturing in-flight and offers future possibilities for building in unbounded, at-height or hard-to-access locations.
This is a preview of subscription content, access via your institution
Relevant articles
Open Access articles citing this article.
-
Aerodynamic effect for collision-free reactive navigation of a small quadcopter
npj Robotics Open Access 26 October 2023
-
Space-filling and print path generation methods for large-area 3D concrete printing pavements
Architectural Intelligence Open Access 29 June 2023
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 51 print issues and online access
$199.00 per year
only $3.90 per issue
Rent or buy this article
Prices vary by article type
from$1.95
to$39.95
Prices may be subject to local taxes which are calculated during checkout
Data availability
The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information. Each data point corresponding to figures that describe the results from experimental and simulation studies are provided as separate Source Data for Figs. 3a,b,d and 4a–c and Extended Data Figs. 5a,b, 6b–d and 7a–d. Other source data related to the study are available from the corresponding author upon reasonable request.
Code availability
The custom code for all algorithms developed in this work are available from the corresponding author on reasonable request.
References
Lim, S. et al. Developments in construction-scale additive manufacturing processes. Autom. Constr. 21, 262–268 (2012).
Gosselin, C. et al. Large-scale 3D printing of ultra-high performance concrete-a new processing route for architects and builders. Mater. Design 100, 102–109 (2016).
Keating, S., Leland, J., Cai, L. & Oxman, N. Toward site-specific and self-sufficient robotic fabrication on architectural scales. Sci. Robot. 2, 15 (2017).
Buswell, R., Silva, W., Jones, S. & Dirrenberger, J. 3D printing using concrete extrusion: a roadmap for research. Cem. Concr. Res. 112, 37–49 (2018).
The 21st Century Revolution in Building Technology Has A Name (Monolite Ltd, 2009); https://www.cadblog.pl/podcasty/luty_2012/d_shape_presentation.pdf
BigDelta WASP 12MT. WASP https://www.3dwasp.com/en/giant-3d-printer-bigdelta-wasp-12mt (2016).
Sustarevas, J. et al. MAP-a mobile agile printer robot for on-site construction. In Proc. 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2441–2448 (IEEE, 2018).
Zhang, X. et al. Large-scale 3D printing by a team of mobile robots. Autom. Constr. 95, 98–106 (2018).
Off-site Manufacture for Construction: Building for Change (House of Lords Science and Technology Committee, 2018); https://publications.parliament.uk/pa/ld201719/ldselect/ldsctech/169/16902.htm
Khoshnevis, B. Automated construction by contour crafting related robotics and information technologies. Autom. Constr. 13, 5–19 (2004).
Hansell, M. Built by Animals: the Natural History of Animal Architecture (OUP, 2007).
Theraulaz, G., Bonabeau, E. & Deneubourg, J. The mechanisms and rules of coordinated building in social insects. In Information Processing In Social Insects (eds Detrain, C. et al.) 309–330 (Birkhäuser, 1999).
Petersen, K., Napp, N., Stuart-Smith, R., Rus, D. & Kovac, M. A review of collective robotic construction. Sci. Robot. 4, eaau8479 (2019).
Delgado, J. et al. Robotics and automated systems in construction: understanding industry-specific challenges for adoption. J. Build. Eng. 26, 100868 (2019).
Tay, Y. et al. 3D printing trends in building and construction industry: a review. Virtual Phys. Prototyp. 12, 261–276 (2017).
Lindsey, Q., Mellinger, D. & Kumar, V. Construction of cubic structures with quadrotor teams. In Robotics: Science And Systems VII 177–184 (MIT Press, 2012).
Mirjan, A., Augugliaro, F., D’Andrea, R., Gramazio, F. & Kohler, M. in Robotic Fabrication In Architecture, Art And Design 2016 (eds Reinhardt, D. et al.) 34–47 (Springer, 2016).
Stuart-Smith, R. Behavioural production: autonomous swarm-constructed architecture. Archit. Des. 86, 54–59 (2016).
Furrer, F. et al. Autonomous robotic stone stacking with misc next best object target pose planning. In 2017 IEEE International Conference On Robotics And Automation (ICRA) 2350–2356 (IEEE, 2017).
Kayser, M. et al. Design of a multi-agent, fiber composite digital fabrication system. Sci. Robot. 3, eaau5630 (2018).
Hutter, M. et al. in Field And Service Robotics (eds Hutter, M. & Siegwart, R.) 575–589 (Springer, 2018)
Chermprayong, P., Zhang, K., Xiao, F. & Kovac, M. An integrated delta manipulator for aerial repair: a new aerial robotic system. IEEE Robot. Autom. Mag. 26, 54–66 (2019).
Bellingham, J. & Rajan, K. Robotics in remote and hostile environments. Science 318, 1098–1102 (2007).
Helm, V., Ercan, S., Gramazio, F. & Kohler, M. Mobile robotic fabrication on construction sites: DimRob. In IEEE/RSJ International Conference on Intelligent Robots And Systems (IROS) 4335–4341 (IEEE, 2012).
Werfel, J., Petersen, K. & Nagpal, R. Designing collective behavior in a termite-inspired robot construction team. Science 343, 754–758 (2014).
Willmann, J. et al. Aerial robotic construction towards a new field of architectural research. Int. J. Archit. Comput. 10, 439–459 (2012).
Augugliaro, F. et al. The flight assembled architecture installation: cooperative construction with flying machines. IEEE Control Syst. Mag. 34, 46–64 (2014).
Minibuilders. IAAC https://iaac.net/project/minibuilders/ (2016).
Speich, S., Jones, H. & Benedict, E. Review of the natural nesting of the barn swallow in North America. Am. Midl. Nat. 115, 248–254 (1986).
Miriyev, A. & Kovač, M. Skills for physical artificial intelligence. Nat. Mach. Intell. 2, 658–660 (2020).
Zareiyan, B. & Khoshnevis, B. Interlayer adhesion and strength of structures in contour crafting-effects of aggregate size, extrusion rate, and layer thickness. Autom. Constr. 81, 112–121 (2017).
Zareiyan, B. & Khoshnevis, B. Effects of mixture ingredients on interlayer adhesion of concrete in contour crafting. Rapid Prototyp. J. 24, 584–592 (2018).
Le, T. et al. Mix design and fresh properties for high-performance printing concrete. Mater. Struct. 45, 1221–1232 (2012).
Wangler, T. et al. Digital concrete: opportunities and challenges. RILEM Tech. Lett. 1, 67–75 (2016).
Lloret Fritschi, E. Smart Dynamic Casting-a Digital Fabrication Method for Non-standard Concrete Structures (ETH Zurich, 2016).
Hack, N. et al. Mesh mould: an on site, robotically fabricated, functional formwork. In Proceedings of 11th High Performance Concrete (11th HPC) and Second Concrete Innovation Conference (2nd CIC) (ed Hammer, T. A.) 330–340 (2017).
Bos, F., Wolfs, R., Ahmed, Z. & Salet, T. Additive manufacturing of concrete in construction: potentials and challenges of 3D concrete printing. Virtual Phys. Prototyp. 11, 209–225 (2016).
Bos, F., Ahmed, Z., Jutinov, E. & Salet, T. Experimental exploration of metal cable as reinforcement in 3D printed concrete. Materials 10, 1314 (2017).
Wolfs, R., Bos, F. & Salet, T. Early age mechanical behaviour of 3D printed concrete: numerical modelling and experimental testing. Cem. Concr. Res. 106, 103–116 (2018).
Ghaffar, S., Corker, J. & Fan, M. Additive manufacturing technology and its implementation in construction as an eco-innovative solution. Autom. Constr. 93, 1–11 (2018).
Barnett, E. & Gosselin, C. Large-scale 3D printing with a cable-suspended robot. Addit. Manuf. 7, 27–44 (2015).
Subrin, K. et al. Improvement of the mobile robot location dedicated for habitable house construction by 3D printing. IFAC-PapersOnLine 51, 716–721 (2018).
Marchon, D., Kawashima, S., Bessaies-Bey, H., Mantellato, S. & Ng, S. Hydration and rheology control of concrete for digital fabrication: potential admixtures and cement chemistry. Cem. Concr. Res. 112, 96–110 (2018).
Vespa, E., Funk, N., Kelly, P. & Leutenegger, S. Adaptive-resolution octree-based volumetric SLAM. In 2019 International Conference On 3D Vision (3DV) 654–662 (IEEE, 2019).
Jiao, D. et al. Effect of constituents on rheological properties of fresh concrete—a review. Cem. Concr. Compos. 83, 146–159 (2017).
Lootens, D., Hébraud, P., Lécolier, E. & Van Damme, H. Gelation, shear-thinning and shear-thickening in cement slurries. Oil Gas Sci. Technol. 59, 31–40 (2004).
Studart, A. Additive manufacturing of biologically-inspired materials. Chem. Soc. Rev. 45, 359–376 (2016).
Chen, M. et al. Rheological and mechanical properties of admixtures modified 3D printing sulphoaluminate cementitious materials. Constr. Build. Mater. 189, 601–611 (2018).
Group, C. National Structural Concrete Specification for Building Construction (British Cement Association, 2000).
Ander, M., Olsson, J., Shepherd, P., Stuart-Smith, R. & Williams, C. A building of unlimited height. Proc. IASS Annu. Symp. 2019, 1465–1472 (2019).
Acknowledgements
We acknowledge the support of the Engineering and Physical Sciences Research Council (EPSRC) awards under grant agreements EP/N018494/1, EP/K005030/1 and EP/S031464/1, the EPSRC Centre for Decarbonisation of the Built Environment (dCarb) under grant agreement EP/L016869/1, the Royal Society Wolfson Fellowship under grant number RSWFR1180003, EU H2020 AeroTwin project under grant number 810321 (M.K.), the Royal Thai Government Scholarship (P.C.), the University of Bath Research Scholarship (B.D.) and the Department of Aeronautics of Imperial College London. We thank T. Al-Hinai and R. Siddall for their contributions in the early conceptualization stage of the project, Z. Jiang, C. Liu and Y. F. Kaya for their assistance in experimental tests and multi-media files preparation, and A. Cully for pre-reviewing the early versions of the paper.
Author information
Authors and Affiliations
Contributions
K.Z., S.S., L.M., V.M.P., R.J.B., C.W., P.S., S.L., R.S.-S. and M.K. conceived the study. K.Z., P.C., B.B.K., L.O. and M.K. designed and engineered the Aerial-AM robots and material extrusion system. D.T., W.L., C.C., P.C., K.Z., M.K. and S.L. designed and analysed the controller for the Aerial-AM robots. B.D., S.A.N. and R.J.B. engineered the material mixtures and performed material tests. S.K., V.M.P., S.H., K.Z., P.C., F.X., D.T., S.L., M.K. and R.S.-S. designed the multi-agent framework and performed the light-trace virtual AM demonstration and simulations. C.W., P.S. and R.S.-S. performed design of proof-of-concept geometries. K.Z., P.C., F.X., D.T., B.D., S.K., B.B.K., A.B., D.D.D., A.C., L.M., V.M.P., S.L. and M.K. carried out system integration and Aerial-AM printing experiments with the robots. K.Z., B.D., V.M.P., S.L., R.S.-S. and M.K. wrote the manuscript. M.K. and R.S.-S. conducted pilot research and initiated the research. All authors contributed to and approved the final draft of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature thanks Theo Salet and Fu Zhang for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Aerial-AM robots.
a, The BuilDrone for foam printing. The foam material canisters which store the dual components of the expansion foam are mounted underneath the quadrotor platform. The nozzle for spraying the foam material is then fixed to the bottom of the canister holder. b, The BuilDrone for cementitious material print. The cementitious material extruder is placed in the holder underneath the wheelbase of the BuilDrone while the upper platform of delta parallel manipulator is attached to the holder. The nozzle is mounted on the end-effector of the delta manipulator and connected to the extruder though tubing. (d: distance from nozzle to the substrate; h: single layer height; w: layer width.) c, The ScanDrone with an integrated RGBD camera for 3D mapping of the printed structure.
Extended Data Fig. 2 Deviation compensating using delta manipulator.
a, The setting of BuilDrone with upper platform of delta parallel manipulator mounted underneath the quadrotor platform. b, Kinematic diagram of the lightweight delta parallel manipulator which has three limbs with identical kinematic structure. The end-effector with geometric centre Oe implements pure translational motion with respect to the upper platform with geometric centre Oc. c, Schematic diagram of the deviation compensation principle: the nozzle tip F keeps at desired position though the BuilDrone platform may drift to the pose at \({O}^{{\prime} }\) away from the reference pose at Ob. This method results in higher positional accuracy of the nozzle tip for depositing the material at target position T.
Extended Data Fig. 3 Geometry designs and printed layer/s of sample printing path for thin-walled cylinders.
a, The printing path with four concentric circles (Separation: 8 × 10−3 m, inner diameter (ID): 272 × 10−3 m, outer diameter (OD): 320 × 10−3 m). b, The printing path with rounded Peano curve (ID: 260 × 10−3 m, OD: 320 × 10−3 m, Period of pattern: 50 × 10−3 m, Amplitude of pattern: 30 × 10−3 m, Closest approach between successive shapes: 8 × 10−3 m. c, The hybrid printing path including concentric circles and compact rounded Peano curve in alternative layers (ID: 255 × 10−3 m, OD: 335 × 10−3 m, Straight lines separation: 20 × 10−3 m, Sinusoidal period: 18 × 10−3 m, Sinusoidal amplitude: 52 × 10−3 m). d, The first layer of a printed sample using pure concentric circles. e, The first layer and the half-unit offset second layer of a printed sample using rounded Peano curve printing path. f, The first two layers printed using the hybrid printing path. g–i, The top view of the printed samples with 5 layers using three different path designs respectively. j–l, Front view of the five-layer structures.
Extended Data Fig. 4 Robot Operating System (ROS) based control architecture for Aerial-AM robot platforms.
a, High-level control architecture. b, Model Predictive Control diagram for trajectory tracking for both BuilDrone and ScanDrone. c, Control architecture of BuilDrone deviation compensation using the integrated delta manipulator.
Extended Data Fig. 5 Position errors of the BuilDrone platform during the foam printing in flight.
a, BuilDrone position error measured using the centre of mass. b, Absolute position error of the centre of mass of BuilDrone. For the box plots, the middle quartile denotes the median and the lower and upper quartiles indicate the 25th and 75th percentiles, respectively; the whiskers denote 1.5 times the interquartile range from the upper or lower quartiles.
Extended Data Fig. 6 The four cementitious–polymeric composite mixes trialled with the BuilDrone.
No.1 (green), No.2 (orange), No.3 (red) and No.4 (blue), with mix 1 possessing the best buildability (the ability of the material to retain shape and resist deformation following extrusion due to subsequently deposited layers) and mix 4 the best workability (the ability of a material to be pushed through and extruded from a deposition device). a: Potential constituents plotted to show contribution to the properties of mixes. Workability was considered to be the primary parameter, with the selected constituents for mix formulation highlighted. b: The full constituent specifications of mixes No.1-No.4 in kg/m3 to three significant figures. Key: CEM1=Portland Cement, PFA=Pulverised Fuel Ash, Xan=Xanthan gum, hemc=Hydroxyethyl methyl cellulose, Foam=EAB Associates foaming agent mixed with water and brought to a stiff-peak consistency, Plast.=Adoflow ‘S’ plasticiser. Fresh mix densities: No.1: 1793 kg/m3, No.2: 1741 kg/m3, No.3: 1757 kg/m3No.4: 1760 kg/m3. c: Viscosity flow profiles for mixes 1-4 and viscosity values relating to the four mixes while at rest, in the cartridge vessel and in the tubing indicated. d: Selected material parameters giving an overview of properties of cementitious mix 1-4. Key: phase angle δ (∘), complex modulus G*, 28-day compressive strength f28c, 28-day flexural strength f28f (all MPa) and the force required to process the material through the deposition device and tubing (N), the value shown on the figure being the true value divided by a factor of 10. For purposes of clarity and presentation, error bars for the individual material properties are included in the respective cementitious materials test sections and the table (Supplementary Table 5) providing an summary of tests in Supplementary Experiment S1, which also contains information on sample size and additional material parameters including yield stress, which ranged from 0.7 (Mix 4) to 1.1 kPa (Mixes 1-3).
Extended Data Fig. 7 Position errors of the BuilDrone platform and the printing nozzle tip during the cementitious material printing in flight.
a, BuilDrone position error. b, Position error of the tip of depositing nozzle mounted on delta manipulator’s end-effector. During the print, the tubing was filled with material and becomes stiffer. This led to negative errors in x- and y-direction. c, BuilDrone absolute position error. d, Absolute position error of the tip of the depositing nozzle. For the box plots, the middle quartile denotes the median and the lower and upper quartiles indicate the 25th and 75th percentiles, respectively; the whiskers denote 1.5 times the interquartile range from the upper or lower quartiles.
Supplementary information
Supplementary Information
This file contains Supplementary Methods, which do not readily fit within the constraints of the Methods/Extended Data formats, and Supplementary experiments related to the research article titled. The file has 40 display items, including 26 assembled figures and 14 tables. The figures illustrate pseudo-code of algorithms, experimental set-ups, and results of statistic analyses and simulations related to the results reported in the main paper.
Supplementary Tables
This file includes all the tables, including 2 for Extended Data and 14 for Supplementary Information, prepared using the table menu in Microsoft Word. These tables structurally document key parameters and values in experiments and terminologies used.
Supplementary Video 1
Aerial additive manufacturing (Aerial-AM): Concept differences in constraints, capabilities and applications between the proposed Aerial-AM method and established approaches in the construction industry. Prefabricated construction methods rely heavily on large ground-based systems such as static gantry and crane systems for placing large elements at the building site. Using the multidisciplinary physical artificial intelligence development method, we developed the Aerial-AM system to provide a new capability suitable for on-site manufacturing in hard-to-reach, remote or non-standard geometrical site conditions.
Supplementary Video 2
Multi-agent Aerial-AM framework. To enable Aerial-AM for building-scale manufacturing using multiple networked aerial robots, it requires solutions to multi-robot coordination beyond currently available methodologies. We developed a multi-agent Aerial-AM framework that benefits from the principle of incremental manufacturing—employing large numbers of trips between material supplies and the manufacturing site, carrying small amounts of materials at a time. The framework provides capabilities for live autonomous task allocation, spatial collision awareness, collective organization and system robustness through redundancy.
Supplementary Video 3
Aerial-AM: a tall cylinder print with 3D scan in the loop. We presented a 3D printed cylinder of 2.05 m in height using foam material to demonstrate the capability of the Aerial-AM system for manufacturing geometries at a relatively large scale. It took a BuilDrone 24 seconds to print 1 layer and 29 minutes in total to complete the tall cylinder. During the manufacturing process, we deployed a ScanDrone equipped with an onboard RGBD sensor to measure the mean height of the top layer. The trajectory to print the next layer is then simply offset accordingly to nominally fly at the correct relative reference above the built structure.
Supplementary Video 4
Aerial-AM virtual printing: Peano curve trajectory light-painting flight. We developed a comprehensive control system by integrating a custom-developed position controller and finite-horizon model-predictive-control (MPC) scheme, which can adapt to all aerial robots for the Aerial-AM system. By further incorporating the control system and the kinematics-based controller of the self-aligning delta manipulator, the BuilDrone can achieve an improved manufacturing accuracy to 5 mm in the experimental space with accurate state estimation. We demonstrated the high-precision capability of the BuilDrone with an integrated Delta manipulator by virtual printing a cylindrical geometry with Peano curve printing path.
Supplementary Video 5
Aerial-AM: printing path tests for cementitious material deposition. In this work, three scalable paths were designed for constructing cylindrical geometries: (1) multiple adjacent concentric circles effectively forming a solid wall; (2) a rounded Peano curve, with alternating layers staggered around the circle with a half-unit offset; and (3) a hybrid design with three non-adjacent concentric circles alternating with a rounded Peano curve. We printed samples using the BuilDrone with integrated delta manipulators to help qualitative assessment of each path’s likely tolerance to geometrical inaccuracies on the scale of those produced by the BuilDrone.
Supplementary Video 6
Aerial-AM: cylinder print using Peano curve printing path. Experimental tests of samples indicated that among the three different designs the rounded Peano curve design has the best balance of material use and structural performance. Using two BuilDrones with integrated delta manipulators, we additively manufactured a 28-layer thin-walled cylinder with rounded Peano curve path. In the printing process of the 28-layer structure, the two BuilDrones print in sequence to guarantee continuous construction. it took 4 minutes and 45 seconds to print 1 layer and in total 2 hours and 13 minutes were used for material deposition in printing the 28-layered cylinder.
Supplementary Video 7
Aerial-AM: cylinder print using Peano curve printing path. Experimental tests of samples indicated that among the three different designs the rounded Peano curve design has the best balance of material use and structural performance. Using two BuilDrones with integrated delta manipulators, we additively manufactured a 28-layer thin-walled cylinder with rounded Peano curve path. In the printing process of the 28-layer structure, the two BuilDrones print in sequence to guarantee continuous construction. it took 4 minutes and 45 seconds to print 1 layer and in total 2 hours and 13 minutes were used for material deposition in printing the 28-layered cylinder.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Zhang, K., Chermprayong, P., Xiao, F. et al. Aerial additive manufacturing with multiple autonomous robots. Nature 609, 709–717 (2022). https://doi.org/10.1038/s41586-022-04988-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41586-022-04988-4
This article is cited by
-
3D-printed PEDOT:PSS for soft robotics
Nature Reviews Materials (2023)
-
Propelling the widespread adoption of large-scale 3D printing
Nature Reviews Materials (2023)
-
Aerodynamic effect for collision-free reactive navigation of a small quadcopter
npj Robotics (2023)
-
Analysis of 3D printing techniques for building construction: a review
Construction Robotics (2023)
-
Space-filling and print path generation methods for large-area 3D concrete printing pavements
Architectural Intelligence (2023)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
