EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Novel Virtual Geometry Shadowing Approach For Multi-Robot Formation Control In Dynamic Environments

Hoolark Jung, Donghyun Lee, Hyunchul Tae

Journal of Intelligent & Robotic Systems · 16.11.2026

DOI: 10.1007/s10846-026-02488-5 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Intelligent & Robotic Systems, online erschienen am 16.11.2026. Verfasst von Hoolark Jung und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Korea Institute of Industrial Technology.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Existing multi-robot cooperative transport techniques suffer from continuous communication dependency to maintain load-supporting configurations. Leader-Follower schemes require constant communication between robots, while centralized virtual structure approaches demand continuous position feedback. As multi-robot fleets scale up in logistics automation and smart factory environments, continuous inter-robot communication loads the shared wireless channel with control-rate traffic and creates computational bottlenecks, posing critical risks to the stable transport of fragile payloads. This paper proposes the Shadowing technique to resolve this communication dependency. The key innovation is that robots maintain formation and continue transport during communication outages by following a pre-transmitted time-space trajectory of the virtual polygon framework. Each robot independently maintains its assigned position based on pre-stored timestamps. Communication is only required when dynamic obstacles are detected or positional errors exceed thresholds. Through experiments across 16, 200 trials per path planner (3 map types $$\times $$ × 4 obstacle densities $$\times $$ × 3 formation methods $$\times $$ × 3 formation sizes (2, 4, and 6 robots) $$\times $$ × 150 trials), using Anytime D* as the path planning backbone, Shadowing achieved an average reduction in communication message count of $$93.8\%$$ 93.8 % , $$94.3\%$$ 94.3 % , and $$95.5\%$$ 95.5 % relative to Leader-Follower for 2-, 4-, and 6-robot formations respectively, and $$97.8\%$$ 97.8 % , $$97.0\%$$ 97.0 % , and $$97.4\%$$ 97.4 % relative to Centralized-VS; these communication-reduction figures refer to message counts, and a complementary data-volume analysis is reported as well. In this quantitative campaign, each dynamic obstacle appears at a random time uniformly distributed over 1– $$2\,\textrm{s}$$ 2 s and remains static thereafter; continuously moving obstacles are evaluated separately in a physics-based Gazebo study. Shadowing completed missions $$25.0\%$$ 25.0 % , $$21.2\%$$ 21.2 % , and $$21.5\%$$ 21.5 % faster than Leader-Follower across the three formation sizes, while incurring only 4.6– $$10.0\%$$ 10.0 % longer mission time than Centralized-VS in exchange for a 42– $$48\%$$ 48 % reduction in formation error. Under the A* path planning baseline, the corresponding figures were $$93.7\%$$ 93.7 % , $$93.7\%$$ 93.7 % , and $$94.8\%$$ 94.8 % message-count reduction relative to Leader-Follower; $$97.8\%$$ 97.8 % , $$96.8\%$$ 96.8 % , and $$97.0\%$$ 97.0 % relative to Centralized-VS; $$19.4\%$$ 19.4 % – $$23.9\%$$ 23.9 % faster than Leader-Follower; and 45– $$51\%$$ 51 % lower formation error than Centralized-VS. The absolute message gap between Shadowing and the baselines grows roughly linearly with the number of robots, since Shadowing’s broadcast count is governed by event rate (initial trajectory, dynamic-obstacle replanning, and realignment) rather than by control frequency, while the reduction ratio remains above $$93\%$$ 93 % relative to Leader-Follower and above $$96\%$$ 96 % relative to Centralized-VS across all evaluated formation sizes. Zero mission failures and zero formation collapses were recorded for the Shadowing technique across all conditions under both planners. Throughout all conditions, communication consists of a single initial trajectory broadcast augmented by event-triggered broadcasts for dynamic-obstacle replanning and for realignment when the position error exceeds $$\varepsilon _{\max }$$ ε max .

Förderung

  • Korea Institute of Industrial Technology

Zitieren

APA

Jung, H., Lee, D., & Tae, H. (2026). A Novel Virtual Geometry Shadowing Approach For Multi-Robot Formation Control In Dynamic Environments. Journal of Intelligent & Robotic Systems. https://doi.org/10.1007/s10846-026-02488-5

Vancouver

Jung H, Lee D, Tae H. A Novel Virtual Geometry Shadowing Approach For Multi-Robot Formation Control In Dynamic Environments. Journal of Intelligent & Robotic Systems. 2026. doi:10.1007/s10846-026-02488-5

BibTeX
@article{Jung2026,
  author = {Jung, Hoolark and Lee, Donghyun and Tae, Hyunchul},
  title = {A Novel Virtual Geometry Shadowing Approach For Multi-Robot Formation Control In Dynamic Environments},
  journal = {Journal of Intelligent & Robotic Systems},
  year = {2026},
  doi = {10.1007/s10846-026-02488-5}
}
Kostenlos starten →