pith. machine review for the scientific record. sign in

arxiv: 2604.23693 · v2 · submitted 2026-04-26 · 💻 cs.RO

Recognition: unknown

Decentralized Heterogeneous Multi-Robot Collaborative Exploration for Indoor and Outdoor 3D Environments

Yuxiang Li , Kun Chen , Jiancheng Wang , Shihao Fang , Haoyao Chen , Yunhui Liu
Authors on Pith no claims yet

Pith reviewed 2026-05-12 00:45 UTC · model grok-4.3

classification 💻 cs.RO
keywords heterogeneous multi-robot systemscollaborative exploration3D mappingdecentralized coordinationHMDMTSPgenetic algorithmsupervoxel segmentationindoor and outdoor environments
0
0 comments X

The pith

A decentralized framework assigns exploration tasks to heterogeneous robots by modeling their distinct capabilities and solving a constrained multi-depot traveling salesman problem.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out a complete pipeline that lets robots with different speeds, sensors, and ranges explore cluttered 3D spaces without a central controller. It builds a shared perception map, simplifies it with supervoxels, groups viewpoints by what each robot can actually do, and hands the resulting clusters to an improved genetic algorithm that respects capability constraints. A reader cares because real deployments often mix fast ground robots with slower aerial ones; if the assignment step works, the team covers space faster and exchanges far less data than uniform-robot baselines. The claim rests on both simulation runs and field trials in indoor and outdoor clutter.

Core claim

By first building a basic perception map that fuses terrain and observation metrics, then applying improved supervoxel segmentation for lightweight sharing, modeling each robot's traversal and observation limits, clustering task views accordingly, and casting the assignment as a heterogeneous multi-depot multi-traveling salesman problem solved by an improved genetic algorithm, the framework produces collision-free routes that achieve higher exploration efficiency and lower communication volume than prior methods in both simulated and real cluttered 3D environments.

What carries the argument

The heterogeneous multi-depot multi-traveling salesman problem (HMDMTSP) that encodes view-cluster requirements together with each robot's distinct traversal and observation capabilities, solved by an improved genetic algorithm that maintains global consistency before redundant views are pruned and motion conflicts are resolved.

If this is right

  • Each robot receives only the viewpoint clusters its sensors and mobility can actually handle, so coverage improves without idle time or unreachable targets.
  • Supervoxel compression plus elimination of redundant views inside each cluster sharply reduces the data that must be exchanged to keep the shared map consistent.
  • Once assignments are fixed, simple local path-conflict resolution is sufficient to let the robots move simultaneously without central replanning.
  • The same pipeline works for both indoor rooms and outdoor terrain because the perception map and capability models are environment-agnostic.
  • Global consistency of the genetic-algorithm solution prevents the fragmented coverage that arises when robots greedily claim nearby views.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same capability-modeling step could be reused for other heterogeneous tasks such as search-and-rescue or persistent surveillance once the exploration phase ends.
  • If the genetic algorithm's runtime grows too quickly with team size, a distributed approximation or auction-based alternative would be needed to keep the system truly decentralized at scale.
  • Adding online map updates that re-trigger clustering when new obstacles appear would test whether the offline assignment remains stable in dynamic scenes.
  • The approach implicitly assumes static environments; extending the capability models to include energy budgets or time-varying sensor ranges would broaden its applicability.

Load-bearing premise

The traversal and observation models used to score task views must match actual robot performance closely enough that the genetic algorithm's assignments remain useful once the robots start moving in the real world.

What would settle it

A side-by-side run in the same cluttered indoor or outdoor site where the proposed method's total exploration time or total bytes transmitted exceeds the best competing decentralized planner would falsify the efficiency and communication claims.

Figures

Figures reproduced from arXiv: 2604.23693 by Haoyao Chen, Jiancheng Wang, Kun Chen, Shihao Fang, Yunhui Liu, Yuxiang Li.

Figure 1
Figure 1. Figure 1: The proposed decentralized heterogeneous multi-robot collaborative exploration framework for view at source ↗
Figure 2
Figure 2. Figure 2: The voxel-view relationship from a top-down view. view at source ↗
Figure 3
Figure 3. Figure 3: Voxel segmentation results in two scenes. (a) Supervoxel segmentation view at source ↗
Figure 4
Figure 4. Figure 4: Results of two supervoxel-segmentation methods for noisy ground view at source ↗
Figure 5
Figure 5. Figure 5: Visibility check between ground and facade supervoxels. view at source ↗
Figure 6
Figure 6. Figure 6: Stochastic individual generation of multi-chromosome and mutation view at source ↗
Figure 7
Figure 7. Figure 7: Generation process from task-view clusters to task-view sequences. (a) Candidate views are produced by extracting ground-ISE supervoxels near the current cluster. (b) A new view sequence is obtained via the MNSS. Sec. V-A). As shown in view at source ↗
Figure 9
Figure 9. Figure 9: Simulated collapsed building scenes and their ground-truth point view at source ↗
Figure 10
Figure 10. Figure 10: Metric results of the four approaches in the simulated single building view at source ↗
Figure 11
Figure 11. Figure 11: Top view of reconstructed maps and exploration trajectories of the view at source ↗
Figure 14
Figure 14. Figure 14: Computation runtime of each component of the proposed approach view at source ↗
Figure 17
Figure 17. Figure 17: The real-world scene and its ground truth point cloud. (a) Overview view at source ↗
Figure 18
Figure 18. Figure 18: The real-world scene and reconstructed maps in different regions. view at source ↗
Figure 20
Figure 20. Figure 20: Two bird-eye views (a) and (b) of reconstructed map and exploration view at source ↗
read the original abstract

Heterogeneous multi-robot systems feature significant adaptability for complex environments. However, effective collaboration that fully exploits the robots' potential remains a core challenge. This paper proposes a decentralized collaborative framework for heterogeneous multi-robot systems to autonomously explore indoor and outdoor 3D environments. First, a basic perception map that integrates terrain and observation metrics is designed. Improved supervoxel segmentation is developed to simplify the map structure and form a high-level representation that supports lightweight communication. Second, the traversal and observation capabilities of heterogeneous robots are modeled to evaluate the requirements of task views derived from incomplete supervoxels. These task views are grouped by requirements and clustered to streamline assignment. Subsequently, the view-cluster assignment is formulated as a heterogeneous multi-depot multi-traveling salesman problem (HMDMTSP) that incorporates constraints between view-cluster requirements and robot capabilities. An improved genetic algorithm is developed to efficiently solve this problem while ensuring global consistency. Based on the assignments, redundant views within clusters are eliminated to refine exploration routes. Finally, conflicts between robots' motion paths are resolved. Simulations and field experiments in cluttered indoor and outdoor environments demonstrate that our approach effectively coordinates exploration tasks among heterogeneous robots, achieving superior exploration efficiency and communication savings compared to state-of-the-art approaches.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes a decentralized collaborative exploration framework for heterogeneous multi-robot systems operating in cluttered indoor and outdoor 3D environments. It constructs a perception map integrating terrain and observation metrics, applies improved supervoxel segmentation to create a lightweight high-level representation, models the traversal and observation capabilities of heterogeneous robots to evaluate task views derived from incomplete supervoxels, clusters these views by requirements, formulates the view-cluster assignment as a heterogeneous multi-depot multi-traveling salesman problem (HMDMTSP) solved via an improved genetic algorithm that enforces global consistency, eliminates redundant views within clusters, and resolves inter-robot motion conflicts. Simulations and field experiments are reported to demonstrate superior exploration efficiency and communication savings relative to state-of-the-art baselines.

Significance. If the quantitative performance gains hold under rigorous validation, the work would offer a practical, communication-efficient method for coordinating robots with differing sensor and mobility profiles in complex 3D settings. The combination of supervoxel-based abstraction with an HMDMTSP formulation addresses a recognized gap in heterogeneous multi-robot exploration; reproducible code or machine-checked elements are not mentioned, but the approach is falsifiable via direct comparison on standard benchmarks.

major comments (3)
  1. [§5 and abstract] §5 (Experiments) and abstract: the central claim of 'superior exploration efficiency and communication savings' is stated without any reported quantitative metrics (e.g., exploration time, coverage percentage, bytes communicated), baselines, error bars, or statistical tests. This absence prevents evaluation of the magnitude or statistical significance of the claimed improvements.
  2. [§3.2] §3.2 (Robot capability modeling): the traversal and observation models are derived directly from incomplete supervoxels and used to assign tasks, yet no prediction-error analysis (e.g., modeled vs. measured traversal times on real robots) or sensitivity study is provided. If these models deviate in cluttered 3D environments, the downstream efficiency gains over SOTA cannot be trusted.
  3. [§4.3] §4.3 (Improved GA for HMDMTSP): the algorithm is asserted to produce globally consistent assignments 'efficiently' and without local-optima entrapment, but no optimality-gap results on benchmark HMDMTSP instances, runtime scaling curves, or ablation of the proposed improvements are reported. This is load-bearing for the global-consistency guarantee.
minor comments (2)
  1. [§2–4] Notation for HMDMTSP and supervoxel parameters is introduced without a consolidated table; a single reference table would improve readability.
  2. [§5] Figure captions for the experimental environments should explicitly state the number of robots, heterogeneity parameters, and map sizes used in each trial.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thorough and constructive review. The three major comments identify areas where the presentation of quantitative results, model validation, and algorithmic analysis can be strengthened. We address each point below and will incorporate revisions to improve clarity and rigor without altering the core contributions.

read point-by-point responses

  1. Referee: [§5 and abstract] §5 (Experiments) and abstract: the central claim of 'superior exploration efficiency and communication savings' is stated without any reported quantitative metrics (e.g., exploration time, coverage percentage, bytes communicated), baselines, error bars, or statistical tests. This absence prevents evaluation of the magnitude or statistical significance of the claimed improvements.

    Authors: We acknowledge that while §5 presents simulation and field experiment results comparing our method against state-of-the-art baselines (with performance visualized in figures), the text does not sufficiently extract and tabulate explicit numerical metrics, error bars, or statistical significance tests. The abstract similarly summarizes the outcome qualitatively. In the revised manuscript we will add a dedicated results table in §5 listing key quantitative values (exploration time, coverage percentage, communication volume) with means, standard deviations, and direct baseline comparisons, and we will revise the abstract to include the principal numerical gains. revision: yes

  2. Referee: [§3.2] §3.2 (Robot capability modeling): the traversal and observation models are derived directly from incomplete supervoxels and used to assign tasks, yet no prediction-error analysis (e.g., modeled vs. measured traversal times on real robots) or sensitivity study is provided. If these models deviate in cluttered 3D environments, the downstream efficiency gains over SOTA cannot be trusted.

    Authors: The traversal and observation models in §3.2 are constructed from the geometric attributes of incomplete supervoxels together with the known sensor ranges and mobility constraints of each robot class; they are not intended as high-fidelity physics simulators but as lightweight estimators sufficient for view-cluster assignment. The field experiments demonstrate that tasks assigned under these models were executed without collision or failure, providing indirect validation. Nevertheless, we agree that an explicit prediction-error analysis and sensitivity study would increase confidence. We will add a new subsection (or appendix) that compares modeled versus measured traversal times from the real-robot trials and reports sensitivity of assignment quality to variations in the model parameters. revision: yes

  3. Referee: [§4.3] §4.3 (Improved GA for HMDMTSP): the algorithm is asserted to produce globally consistent assignments 'efficiently' and without local-optima entrapment, but no optimality-gap results on benchmark HMDMTSP instances, runtime scaling curves, or ablation of the proposed improvements are reported. This is load-bearing for the global-consistency guarantee.

    Authors: Global consistency in our HMDMTSP formulation is enforced by the explicit constraints linking view-cluster requirements to heterogeneous robot capabilities, by the cluster-level assignment, and by the subsequent conflict-resolution step; the genetic algorithm is engineered to respect these constraints through its encoding and repair operators. The manuscript reports wall-clock runtimes observed during the experiments, but we concede that dedicated optimality-gap measurements on standard HMDMTSP benchmark sets, scaling curves, and ablation studies of the algorithmic improvements are absent. In revision we will include an ablation study isolating the effect of the proposed operators, runtime scaling plots with increasing numbers of robots and clusters, and, where feasible, optimality-gap results on small synthetic HMDMTSP instances. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper's approach chains standard components—perception map integration, improved supervoxel segmentation for lightweight representation, explicit modeling of heterogeneous traversal/observation capabilities, HMDMTSP formulation with constraints, improved genetic algorithm assignment, view pruning, and conflict resolution—without any step reducing a claimed result to a fitted parameter, self-definition, or self-citation that is itself unverified. Simulations and field experiments are presented as empirical outcomes rather than derivations that loop back to the same inputs by construction. No load-bearing uniqueness theorems, ansatzes smuggled via prior self-work, or renamings of known patterns appear in the described chain.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Assessment is limited to the abstract; no explicit free parameters, invented entities, or non-standard axioms are identifiable beyond routine robotics modeling assumptions.

axioms (1)
  • domain assumption Heterogeneous robots possess distinct, accurately modelable traversal and observation capabilities that determine task suitability
    Invoked when evaluating requirements of task views derived from incomplete supervoxels.

pith-pipeline@v0.9.0 · 5531 in / 1278 out tokens · 55655 ms · 2026-05-12T00:45:57.362820+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages

  1. [1]

    Large-scale multirobot coverage path planning on grids with path deconfliction,

    J. Tang, Z. Mao, and H. Ma, “Large-scale multirobot coverage path planning on grids with path deconfliction,”IEEE Trans. Robot., vol. 41, pp. 3348–3367, 2025

    work page 2025

  2. [2]

    Search and rescue under the forest canopy using multiple uavs,

    Y . Tian, K. Liu, K. Ok, L. Tran, D. Allen, N. Roy, and J. P. How, “Search and rescue under the forest canopy using multiple uavs,”Int. J. Robot. Res., vol. 39, no. 10-11, pp. 1201–1221, 2020

    work page 2020

  3. [3]

    An exploration-enhanced search algorithm for robot indoor source searching,

    M. Wang, B. Xin, M. Jing, and Y . Qu, “An exploration-enhanced search algorithm for robot indoor source searching,”IEEE Trans. Robot., vol. 40, pp. 4160–4178, 2024

    work page 2024

  4. [4]

    Multi-robot collaborative dense scene reconstruction,

    S. Dong, K. Xu, Q. Zhou, A. Tagliasacchi, S. Xin, M. Nießner, and B. Chen, “Multi-robot collaborative dense scene reconstruction,”ACM Trans. Graph., vol. 38, no. 4, pp. 1–16, 2019

    work page 2019

  5. [5]

    A multirobot system for 3-d surface reconstruction with centralized and distributed architectures,

    G. Hardouin, J. Moras, F. Morbidi, J. Marzat, and E. M. Mouaddib, “A multirobot system for 3-d surface reconstruction with centralized and distributed architectures,”IEEE Trans. Robot., vol. 39, no. 4, pp. 2623–2638, 2023

    work page 2023

  6. [6]

    Fast multi-uav decentralized exploration of forests,

    L. Bartolomei, L. Teixeira, and M. Chli, “Fast multi-uav decentralized exploration of forests,”IEEE Robot. Autom. Lett., vol. 8, no. 9, pp. 5576–5583, 2023

    work page 2023

  7. [7]

    Racer: Rapid collaborative exploration with a decentralized multi-uav system,

    B. Zhou, H. Xu, and S. Shen, “Racer: Rapid collaborative exploration with a decentralized multi-uav system,”IEEE Trans. Robot., vol. 39, no. 3, pp. 1816–1835, 2023

    work page 2023

  8. [8]

    Mui-tare: Cooperative multi-agent exploration with unknown initial position,

    J. Yan, X. Lin, Z. Ren, S. Zhao, J. Yu, C. Cao, P. Yin, J. Zhang, and S. Scherer, “Mui-tare: Cooperative multi-agent exploration with unknown initial position,”IEEE Robot. Autom. Lett., vol. 8, no. 7, pp. 4299–4306, 2023

    work page 2023

  9. [9]

    Fast and communication-efficient multi-uav exploration via voronoi partition on dynamic topological graph,

    Q. Dong, H. Xi, S. Zhang, Q. Bi, T. Li, Z. Wang, and X. Zhang, “Fast and communication-efficient multi-uav exploration via voronoi partition on dynamic topological graph,” inIEEE Int. Conf. Intell. Robots Syst. IEEE, 2024, pp. 14 063–14 070

    work page 2024

  10. [10]

    Cerberus in the darpa subterranean challenge,

    M. Tranzatto, T. Miki, M. Dharmadhikari, L. Bernreiter, M. Kulkarni, F. Mascarich, O. Andersson, S. Khattak, M. Hutter, R. Siegwartet al., “Cerberus in the darpa subterranean challenge,”Sci. Robot., vol. 7, no. 66, p. eabp9742, 2022

    work page 2022

  11. [11]

    Nebula: Team costar’s robotic autonomy solution that won phase ii of darpa subterranean challenge,

    B. Morrell, R. Thakker, `A. Santamaria Navarro, A. Bouman, X. Lei, J. Edlund, T. Pailevanian, T. S. Vaquero, Y . L. Chang, T. Toumaet al., “Nebula: Team costar’s robotic autonomy solution that won phase ii of darpa subterranean challenge,”Field Robot., vol. 2, pp. 1432–1506, 2022

    work page 2022

  12. [12]

    A clustering-based coverage path planning method for autonomous heterogeneous uavs,

    J. Chen, C. Du, Y . Zhang, P. Han, and W. Wei, “A clustering-based coverage path planning method for autonomous heterogeneous uavs,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 12, pp. 25 546–25 556, 2021

    work page 2021

  13. [13]

    Tightly-coupled perception and navigation of heterogeneous land-air robots in complex scenarios,

    Y . Yue, M. Wen, Y . Putra, M. Wang, and D. Wang, “Tightly-coupled perception and navigation of heterogeneous land-air robots in complex scenarios,” inIEEE Int. Conf. Robot. Autom.IEEE, 2021, pp. 10 052– 10 058

    work page 2021

  14. [14]

    Autonomous exploration and mapping system using heterogeneous uavs and ugvs in gps-denied environments,

    H. Qin, Z. Meng, W. Meng, X. Chen, H. Sun, F. Lin, and M. H. Ang, “Autonomous exploration and mapping system using heterogeneous uavs and ugvs in gps-denied environments,”IEEE Trans. Veh. Technol., vol. 68, no. 2, pp. 1339–1350, 2019

    work page 2019

  15. [15]

    Riemannian optimization for active mapping with robot teams,

    A. Asgharivaskasi, F. Girke, and N. Atanasov, “Riemannian optimization for active mapping with robot teams,”IEEE Trans. Robot., vol. 41, pp. 1077–1097, 2025

    work page 2025

  16. [16]

    Multi-robot multi-room exploration with geometric cue extraction and circular decomposition,

    S. Kim, M. Corah, J. Keller, G. Best, and S. Scherer, “Multi-robot multi-room exploration with geometric cue extraction and circular decomposition,”IEEE Robot. Autom. Lett., vol. 9, no. 2, pp. 1190–1197, 2023

    work page 2023

  17. [17]

    Meeting-merging-mission: A multi-robot coordinate framework for large-scale communication-limited exploration,

    Y . Gao, Y . Wang, X. Zhong, T. Yang, M. Wang, Z. Xu, Y . Wang, Y . Lin, C. Xu, and F. Gao, “Meeting-merging-mission: A multi-robot coordinate framework for large-scale communication-limited exploration,” inIEEE Int. Conf. Intell. Robots Syst.IEEE, 2022, pp. 13 700–13 707

    work page 2022

  18. [18]

    A common optimization framework for multi-robot exploration and coverage in 3d environments,

    A. Renzaglia, J. Dibangoye, V . Le Doze, and O. Simonin, “A common optimization framework for multi-robot exploration and coverage in 3d environments,”J. Intell. Robot. Syst., vol. 100, no. 3, pp. 1453–1468, 2020

    work page 2020

  19. [19]

    Decentralized topological mapping for multi-robot autonomous exploration under low-bandwidth communication,

    J. Bayer and J. Faigl, “Decentralized topological mapping for multi-robot autonomous exploration under low-bandwidth communication,” inProc. Eur. Conf. Mobile Robots. IEEE, 2021, pp. 1–7

    work page 2021

  20. [20]

    Multirobot unknown environment exploration and obstacle avoidance based on a voronoi diagram and reinforcement learning,

    H. Zhao, Y . Guo, Y . Liu, and J. Jin, “Multirobot unknown environment exploration and obstacle avoidance based on a voronoi diagram and reinforcement learning,”Expert Syst. Appl., vol. 264, p. 125900, 2025

    work page 2025

  21. [21]

    Active metric-semantic mapping by multiple aerial robots,

    X. Liu, A. Prabhu, F. Cladera, I. D. Miller, L. Zhou, C. J. Taylor, and V . Kumar, “Active metric-semantic mapping by multiple aerial robots,” inIEEE Int. Conf. Robot. Autom.IEEE, 2023, pp. 3282–3288

    work page 2023

  22. [22]

    Aage: Air- assisted ground robotic autonomous exploration in large-scale unknown environments,

    L. Zheng, M. Wei, R. Mei, K. Xu, J. Huang, and H. Cheng, “Aage: Air- assisted ground robotic autonomous exploration in large-scale unknown environments,”IEEE Trans. Robot., 2025

    work page 2025

  23. [23]

    Energy-efficient ground traversability mapping based on uav-ugv col- laborative system,

    J. Li, Y . Cheng, J. Zhou, J. Chen, Z. Liu, S. Hu, and V . C. Leung, “Energy-efficient ground traversability mapping based on uav-ugv col- laborative system,”IEEE Trans. Green Commun. Netw., vol. 6, no. 1, pp. 69–78, 2021

    work page 2021

  24. [24]

    Fast active aerial exploration for traversable path finding of ground robots in unknown environments,

    S. Zhang, X. Zhang, T. Li, J. Yuan, and Y . Fang, “Fast active aerial exploration for traversable path finding of ground robots in unknown environments,”IEEE Trans. Instrum. Meas., vol. 71, pp. 1–13, 2022

    work page 2022

  25. [25]

    Enhancing cooperative exploration and planning: Uav-legged robot synergy,

    Z. Li, P. Xu, L. Ding, J. Shan, W. Yang, Z. Deng, H. Gao, T. Liu, and H. Yang, “Enhancing cooperative exploration and planning: Uav-legged robot synergy,”IEEE Robot. Autom. Lett., 2024

    work page 2024

  26. [26]

    Real- time multilevel terrain-aware path planning for ground mobile robots in large-scale rough terrains,

    Y . Li, K. Chen, Y . Wang, W. Zhang, J. Wang, H. Chen, and Y . Liu, “Real- time multilevel terrain-aware path planning for ground mobile robots in large-scale rough terrains,”IEEE Trans. Robot., vol. 41, pp. 4159–4179, 2025

    work page 2025

  27. [27]

    Object-aware view planning for autonomous 3-d model reconstruction of buildings using a mobile robot,

    Y . Li, J. Wang, H. Chen, X. Jiang, and Y . Liu, “Object-aware view planning for autonomous 3-d model reconstruction of buildings using a mobile robot,”IEEE Trans. Instrum. Meas., vol. 72, pp. 1–15, 2023

    work page 2023

  28. [28]

    A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas,

    J. Terven, D.-M. C ´ordova-Esparza, and J.-A. Romero-Gonz ´alez, “A comprehensive review of yolo architectures in computer vision: From yolov1 to yolov8 and yolo-nas,”Mach. Learn. Know. Extr., vol. 5, no. 4, pp. 1680–1716, 2023

    work page 2023

  29. [29]

    Rangenet ++: Fast and accurate lidar semantic segmentation,

    A. Milioto, I. Vizzo, J. Behley, and C. Stachniss, “Rangenet ++: Fast and accurate lidar semantic segmentation,” inIEEE Int. Conf. Intell. Robots Syst., 2019, pp. 4213–4220

    work page 2019

  30. [30]

    V oxel cloud connectivity segmentation - supervoxels for point clouds,

    J. Papon, A. Abramov, M. Schoeler, and F. Worgotter, “V oxel cloud connectivity segmentation - supervoxels for point clouds,” inProc. IEEE Conf. Comput. Vis. Pattern Recog., 2013, pp. 2027–2034

    work page 2013

  31. [31]

    Multidimensional binary search trees used for associative searching,

    J. L. Bentley, “Multidimensional binary search trees used for associative searching,”Commun. ACM, vol. 18, no. 9, pp. 509–517, 1975

    work page 1975

  32. [32]

    Expanding wavefront frontier detection: An approach for efficiently detecting frontier cells,

    P. Quin, A. Alempijevic, G. Paul, and D. Liu, “Expanding wavefront frontier detection: An approach for efficiently detecting frontier cells,” inAustralas. Conf. Robot. Autom., 2014

    work page 2014

  33. [33]

    A*-based pathfinding in modern computer games,

    X. Cui and H. Shi, “A*-based pathfinding in modern computer games,” Int. J. Comput. Sci. Netw. Secur., vol. 11, no. 1, pp. 125–130, 2011

    work page 2011

  34. [34]

    A multi-stage cuda kernel for floyd-warshall,

    J. W. Smith and B. Lund, “A multi-stage cuda kernel for floyd-warshall,” ArXiv, 2010

    work page 2010

  35. [35]

    Gmp: A genetic mission planner for heterogeneous multirobot system applications,

    B. Miloradovi ´c, B. C ¸¨ur¨ukl¨u, M. Ekstr ¨om, and A. V . Papadopoulos, “Gmp: A genetic mission planner for heterogeneous multirobot system applications,”IEEE Trans. Cybern., vol. 52, no. 10, pp. 10 627–10 638, 2021

    work page 2021

  36. [36]

    Optimization of multiple traveling salesmen problem by a novel representation based genetic algorithm,

    A. Kir ´aly and J. Abonyi, “Optimization of multiple traveling salesmen problem by a novel representation based genetic algorithm,” inIntell. Comput. Optim. in Eng.: Tech. and Appl.Berlin: Springer, 2011, pp. 241–269

    work page 2011

  37. [37]

    Soar: Simultaneous exploration and photographing with heterogeneous uavs for fast autonomous reconstruction,

    M. Zhang, C. Feng, Z. Li, G. Zheng, Y . Luo, Z. Wang, J. Zhou, S. Shen, and B. Zhou, “Soar: Simultaneous exploration and photographing with heterogeneous uavs for fast autonomous reconstruction,” inIEEE Int. Conf. Intell. Robots Syst., 2024, pp. 10 975–10 982

    work page 2024

  38. [38]

    Fast-lio2: Fast direct lidar- inertial odometry,

    W. Xu, Y . Cai, D. He, J. Lin, and F. Zhang, “Fast-lio2: Fast direct lidar- inertial odometry,”IEEE Trans. Robot., vol. 38, no. 4, pp. 2053–2073, 2022

    work page 2053