Capability-oriented construction task allocation method for human-robot collaboration

CHEN Songchun, GUO Hongling

Journal of Tsinghua University(Science and Technology) ›› 2026, Vol. 66 ›› Issue (11) : 2383-2394.

PDF(4086 KB)
PDF(4086 KB)
Journal of Tsinghua University(Science and Technology) ›› 2026, Vol. 66 ›› Issue (11) : 2383-2394. DOI: 10.16511/j.cnki.qhdxxb.2026.26.068
CONSTRUCTION MANAGEMENT

Capability-oriented construction task allocation method for human-robot collaboration

  • {{article.zuoZhe_EN}}
Author information +
History +

Abstract

[Objective] Human-robot collaboration (HRC) is increasingly regarded as an important approach to improving productivity, reducing labor intensity, and enhancing safety on construction sites. However, effective HRC requires that tasks be allocated to workers and robots according to their heterogeneous capabilities and the logical requirements of onsite execution. Existing approaches usually retain workflow-or operation-level divisions, assigning selected steps to robots largely based on managerial experience. Such coarse allocation granularity cannot fully reveal capability requirements within individual operations or provide a unified quantitative basis for comparing workers and robots. The result is often capability mismatch, unnecessary occupation of highly capable resources, frequent agent switching, and discontinuous execution. To address these limitations, this research develops a capability-oriented task allocation method that decomposes construction tasks to the action level, enabling human-robot assignment. [Methods] First, a hierarchical task decomposition framework was established, extending construction task structures to the action level. Drawing on therblig-based work analysis adapted to HRC, construction activities were represented as action units (the minimum allocable elements) covering perception, decision-making, and execution. Second, a unified capability representation framework was developed for human workers and robots: domain-specific capabilities described competence in construction functions, whereas general capabilities characterized perception, decision-making, and execution through indicators such as accuracy, range, latency, adaptability, load, speed, and endurance. Task requirements were then represented within the same capability space. Third, continuity relationships between successive actions were classified as hard continuity, soft continuity, and anti-continuity. Accordingly, a model was formulated to minimize capability redundancy and agent-switching costs, subject to capacity, capability, and logical constraints. After infeasible agents and transitions were eliminated, the problem was transformed into a minimum-cost path problem and solved using the Viterbi dynamic programming algorithm. Finally, the method was validated through a case study of column formwork installation involving two workers, two robots, and ten actions for side-form installation. [Results] The proposed method generated an allocation sequence that assigned early positioning, planning, moving, and grasping actions to the lower-capability robot; posture adjustment and holding to the more capable robot; and installation to the worker. Compared with five experience-based schemes developed by construction managers, the proposed solution achieved the lowest total cost (2.48), whereas the manual solutions yielded total costs ranging from 2.71 to 3.20. The results demonstrated that action-level decomposition distinguished capability requirements within the same construction operation, while the capability representation framework excluded infeasible assignments and reduced unnecessary use of highly capable resources. Continuity-aware optimization further limited fragmented allocation and maintained stable execution chains. Sensitivity analyses indicated that the allocation remained generally stable under variations in robot capability, switching cost, and objective weights, although local adjustments occurred when capability thresholds or optimization preferences changed. [Conclusions] Construction human-robot task allocation can be improved by shifting from coarse workflow-level assignment to action-level, capability-oriented optimization. The proposed framework integrates task decomposition, heterogeneous capability representation, continuity constraints, and sequence optimization within a unified methodology. It provides a transparent and computationally efficient basis for matching construction actions with suitable human or robotic agents, thereby reducing resource mismatches and coordination losses. This research establishes a methodological foundation for task planning and robot deployment in construction HRC and provides theoretical support for advancing automation and intelligence in construction.

Key words

construction robot / human-robot collaboration / task allocation

Cite this article

Download Citations
CHEN Songchun, GUO Hongling. Capability-oriented construction task allocation method for human-robot collaboration[J]. Journal of Tsinghua University(Science and Technology). 2026, 66(11): 2383-2394 https://doi.org/10.16511/j.cnki.qhdxxb.2026.26.068

References

[1] GUAN D, YANG N, LAI J, et al. Kinematic modeling and constraint analysis for robotic excavator operations in piling construction[J]. Automation in Construction, 2021, 126: 103666.
[2] LIU M, GUO J J, DENG L, et al. Enhanced vision-based 6-DoF pose estimation for robotic rebar tying[J]. Automation in Construction, 2025, 171: 105999.
[3] WANG Y, XIE L X, CHEN J, et al. Mortar spraying and plastering integrated robot for wall construction[J]. Automation in Construction, 2024, 165: 105533.
[4] ITURRALDE K, FEUCHT M, ILLNER D, et al. Cable-driven parallel robot for curtain wall module installation[J]. Automation in Construction, 2022, 138: 104235.
[5] WANG Y, XIE L X, WANG H, et al. Intelligent spraying robot for building walls with mobility and perception[J]. Automation in Construction, 2022, 139: 104270.
[6] JIANG D Q, KONG L J, WANG H, et al. Precise control mode for concrete vibration time based on attention-enhanced machine vision[J]. Automation in Construction, 2024, 158: 105232.
[7] WU M H, LIN J R, ZHANG X H. How human-robot collaboration impacts construction productivity: An agent-based multi-fidelity modeling approach[J]. Advanced Engineering Informatics, 2022, 52: 101589.
[8] COSTA MATEUS J E, AGHEZZAF E H, CLAEYS D, et al. Method for transition from manual assembly to human-robot collaborative assembly[J]. IFAC-PapersOnLine, 2018, 51(11): 405-410.
[9] ZHANG F Q, ZHANG Y R, XU S L. Collaboration effectiveness- based complex operations allocation strategy towards to human- robot interaction[J]. Autonomous Intelligent Systems, 2022, 2(1): 20.
[10] LIAU Y Y, RYU K. Task allocation in human-robot collaboration (HRC) based on task characteristics and agent capability for mold assembly[J]. Procedia Manufacturing, 2020, 51: 179-186.
[11] WANG Y Z, FANG Y H, BAI Y. Task planning for human-robot collaboration in structure assembly[J]. Automation in Construction, 2025, 179: 106464.
[12] PETZOLDT C, HARMS M, FREITAG M. Review of task allocation for human-robot collaboration in assembly[J]. International Journal of Computer Integrated Manufacturing, 2023, 36(11): 1675-1715.
[13] MÜLLER R, VETTE M, MAILAHN O. Process-oriented task assignment for assembly processes with human-robot interaction[J]. Procedia CIRP, 2016, 44: 210-215.
[14] PELLEGRINELLI S, ORLANDINI A, PEDROCCHI N, et al. Motion planning and scheduling for human and industrial-robot collaboration[J]. CIRP Annals, 2017, 66(1): 1-4.
[15] SCHMIDBAUER C, SCHLUND S, IONESCU T B, et al. Adaptive task sharing in human-robot interaction in assembly[C]// 2020 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM). Singapore: IEEE, 2020: 546-550.
[16] ANTONELLI D, ASTANIN S, BRUNO G. Applicability of human-robot collaboration to small batch production[M]// AFSARMANESH H, CAMARINHA-MATOS L M, SOARES A L. Collaboration in a Hyperconnected World. Cham: Springer, 2016: 24-32.
[17] BERNOLD L E. Automation and robotics in construction: A challenge and a chance for an industry in transition[J]. International Journal of Project Management, 1987, 5(3): 155-160.
[18] GILBRETH F B, KENT R T. Motion study: A method for increasing the efficiency of the workman[M]. New York: D. Van Nostrand Company, 1911.
[19] 郭红领, 叶啸天, 任琦鹏, 等. 基于BIM和规则推理的施工进度计划自动编排[J]. 清华大学学报(自然科学版), 2022, 62(2): 189-198. GUO H L, YE X T, REN Q P, et al. Automatic generation of construction schedules based on BIM and rule reasoning[J]. Journal of Tsinghua University (Science and Technology), 2022, 62(2): 189-198. (in Chinese)
[20] BUILDINGSMART. IFC 4.3.2[EB/OL]. (2026-04-12) [2026-04- 22]. https://standards.buildingsmart.org/IFC/DEV/IFC4_3/HTML/lexical/IfcTaskType.html.
[21] MA X Y, MAO C, LIU G W. Can robots replace human beings?—Assessment on the developmental potential of construction robot[J]. Journal of Building Engineering, 2022, 56: 104727.
[22] FORNEY G D JR. The Viterbi algorithm[J]. Proceedings of the IEEE, 1973, 61(3): 268-278.
[23] YIN X J, HUANG J W, HE W, et al. Group task allocation approach for heterogeneous software crowdsourcing tasks[J]. Peer-to-Peer Networking and Applications, 2021, 14(3): 1736-1747.
PDF(4086 KB)

Accesses

Citation

Detail

Sections
Recommended

/

〈 〉