Pith. sign in

REVIEW 3 major objections 5 minor 68 references

Real-Time Sampling-Based Safe Motion Planning for Robotic Manipulators in Dynamic Environments

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that a sampling-based manipulator planner can meet hard real-time deadlines on a single CPU core and, using dynamic expanded bubbles, guarantee safe motion whenever obstacle speeds are bounded.

desk verdict Solid new safety structure and extensive experiments, but the hard real-time claim is contradicted by the paper's own timing data and needs reframing. read the letter →

arxiv 2501.00507 v2 pith:JYUFWWS7 submitted 2024-12-31 cs.RO

Nermin Covic , Bakir Lacevic , Dinko Osmankovic , Tarik Uzunovic This is my paper · ORCID
classification cs.RO
keywords real-timemotionplanningdynamicenvironmentssampling-basedexpandedbubblegeneralizedburroboticmanipulatorshardschedulinghuman-robotcollaboration
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish that a sampling-based motion planner for multi-DoF robotic manipulators can run under hard real-time constraints on one CPU core and still carry a provable safety guarantee in environments where obstacles move unpredictably, as long as their speed is bounded. It upgrades the DRGBT planner with a two-task earliest-deadline-first schedule and with a new geometric structure, the dynamic expanded bubble, which shrinks the known-safe region in configuration space by the distance an obstacle could cover during the planning interval. If the central claims hold, a six- or ten-DoF arm equipped with two depth cameras can replan at 20 Hz and still guarantee that any collision happens only after the robot has stopped, with joint velocity, acceleration, and jerk limits respected. The paper supports these claims with a 15,000-run randomized simulation study, comparisons against established replanning algorithms, and real experiments including human presence in the workspace.

What carries the argument

The central object is the dynamic expanded bubble (DEB), a configuration-space region defined around the current configuration by the inequality $\sum_{j=1}^i r_{i,j}\,|y_j - q_j| + v_{\mathrm{obs}} t \le d_i$ for every link $i$ and for time $t$ within the planning horizon. Here $d_i$ is the measured minimal distance from link $i$ to the obstacles, $r_{i,j}$ is a radius that bounds how far points on the robot can move when joint $j$ rotates, and $v_{\mathrm{obs}}$ is the assumed maximum obstacle speed. The bubble is exactly the set of configurations the robot can reach during time $t$ without any link being able to collide, even if an obstacle rushes straight toward it at full speed; higher $v_{\mathrm{obs}}$ or smaller $d_i$ shrinks the bubble. The algorithm chains these bubbles along the candidate spline by updating the separating planes between robot links and obstacles, producing a dynamic generalized bur and a sufficient condition for the whole composite trajectory — current spline plus emergency stop — to be collision-free. This condition, together with a quintic spline parameterization satisfying joint velocity, acceleration, and jerk limits, is what Theorem 1 turns into the guarantee of safe motion.

What would settle it

Run the same timing protocol but with, say, 60 obstacles moving at 1.6 m/s and force replanning every iteration at $T = 50$ ms; if Task 1 ever exceeds its 50 ms budget, the hard real-time claim as stated is false for that operating point.

Watch

Extended reading notes

Core claim

The paper's central claim is that DRGBT, a sampling-based planner built on generalized burs of free configuration space, can be decomposed into two tasks — periodic computation of the next configuration and sporadic replanning — and scheduled by earliest-deadline-first on a single CPU core without GPUs or heavy parallelization. The second central claim is Theorem 1: if the current configuration is safe and every obstacle moves at speed at most $v_{\mathrm{obs}}$, then the planner always produces either a safe spline to a new configuration that respects the kinematic constraints $\mathcal{K}$, or an immediate emergency stop along a safe spline; a solution satisfying $\mathcal{K}$ always exists inside the chain of connected dynamic expanded bubbles. In the safe variant, collisions can only occur after the robot has stopped, a property the paper calls type II collision. The authors report real-time operation at 20 Hz with frequencies up to 100 Hz in the simulation analysis, and they present randomized trials and real-robot scenarios supporting the claims.

Load-bearing premise

The hard real-time guarantee assumes that the timing measurements from the paper's 15,000 randomized trials are a valid worst-case bound on Task 1 for every future run, even though the paper notes that more than 50 obstacles will likely make the task slower, and the chosen operating period of 50 ms is smaller than the measured maximum of about 70 ms.

Editorial extensions

If this is right

  • At the chosen operating point, DRGBT(1) is claimed to meet all deadlines at an iteration time $T^* = 50$ ms, i.e., 20 Hz replanning, on a single CPU core; DRGBT-safe is reported to run at 50–100 Hz for small $T$.
  • With safety enabled, any collision that occurs is of type II, meaning contact only after the robot has already come to a stop, provided obstacle speeds stay below $v_{\mathrm{obs}}$ and no obstacle appears outside the occupied halfspace region.
  • The robot automatically slows down as obstacles approach: smaller measured distances $d_i$ or higher assumed obstacle speed produce shorter safe splines and lower peak velocities.
  • Replanning can be interrupted or skipped and the robot can still reach the goal using only the local horizon, though performance degrades; replanning remains beneficial for global progress toward the goal.
  • In the paper's randomized comparison, DRGBT reports higher success rates than the MARS and RRTX baselines across obstacle counts and DoF values, with shorter algorithm times and comparable or shorter path lengths.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension is to feed detected obstacle velocity directions into the bubble computation; the paper's equations show the safety margin could be relaxed by the cosine of the angle between the approach direction and the obstacle velocity, at the cost of monitoring the velocity vector's validity.
  • The two-task earliest-deadline-first scheduling idea transfers to other sampling-based planners: any planner whose per-iteration routines have an empirically bounded worst-case execution time could use the same logic, provided the measured bound is accepted as the true worst case.
  • Running the planner at 100 Hz while the camera supplies perception at 20 Hz means several iterations reuse stale obstacle data; combining the dynamic expanded bubble margin with constant-velocity prediction could make high-frequency replanning useful without weakening the safety claim.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents DRGBT, a sampling-based motion planner for robotic manipulators in dynamic environments, along with a scheduling framework claimed to enable hard real-time execution on a single CPU core, a new safety structure called dynamic expanded bubbles (DEBs), and both extensive simulations and real-robot experiments. The algorithm decomposes planning into two tasks (T1: local motion computation; T2: global replanning), uses EDF scheduling, and introduces a chain of DEBs to guarantee safe motion under bounded obstacle velocity. The paper reports a 15,000-run randomized simulation study, comparisons against RRTX and MARS, and physical experiments with a UFactory xArm6 manipulator and depth cameras.

Significance. If the hard real-time and safety claims were fully substantiated, the paper would be a useful contribution to real-time motion planning for manipulators in human-robot coexistence scenarios. The paper's strengths include a publicly available implementation, a large randomized evaluation over obstacle counts up to 50, a comparative study with state-of-the-art planners, and a real-robot validation with a sensor-based pipeline. The central advertised contribution, however, is the hard real-time property, and that claim is directly contradicted by the paper's own measurements. The safety theorem is also stated without a proof. These issues affect the paper's main thesis and cannot be resolved by minor editing alone.

major comments (3)
  1. [Sec. III, Eq. (4); Sec. V-B; Sec. V-D; Sec. VIII] The hard real-time schedulability claim is internally inconsistent with the reported timing data. Equation (4) requires e1 <= T for the EDF sufficient condition to be meaningful, since e2 = T - e1 becomes negative otherwise. However, Sec. V-B reports a maximum Task 1 execution time of e1 ≈ 70 ms for DRGBT and ≈ 73 ms for DRGBT-safe over 15,000 runs, and Sec. V-D selects T* = 50 ms (20 Hz), with Sec. VIII additionally claiming operation up to 100 Hz (T = 10 ms). At these operating points e1 > T, so T1 cannot be guaranteed to meet its deadline in all runs; the scheduler must interrupt T1, and no worst-case execution time for the partially executed task is established. The CDF evidence is probabilistic (e.g., generateGBur completes within 30 ms in 99.9859% of cases), not a hard bound. Furthermore, the text concedes that e1 'will likely become larger' with more than 50 obstacles, so the empirical maximum is not a valid WCET. This undermines the paper's primary real-time contribution.
  2. [Sec. IV, Theorem 1] Theorem 1 is stated as a guarantee of safe motion but is not proved; the surrounding text only describes the algorithmic procedure (Algs. 2 and 3) and states that the spline is safe if it lies inside a chain of DEBs. The theorem also depends on a discrete time step Δt in the collision checking, while the guarantee is phrased as continuous-time safety. Moreover, the paper later allows type II collisions (zero-speed contacts), as stated in Sec. IV and Sec. V-D, so the claim 'guaranteed safe motion' is weaker than it appears: the robot may still collide once stopped. A precise statement with a proof or a precise characterization of the conditions under which type II collisions can occur is needed. Without this, the safety contribution is not established at the level claimed by Theorem 1.
  3. [Sec. V-D and Sec. VIII] The claim that the algorithm is capable of real-time operation at frequencies up to 100 Hz is not supported by the presented schedulability analysis. The simulation results in Sec. V-D show that DRGBT-safe performs best for T in the range 10–20 ms (i.e., 50–100 Hz), but these values are smaller than the measured maximum e1 ≈ 73 ms, so the EDF condition of Eq. (4) is violated. The 'sweet spot' for DRGBT(1) is reported at T* = 50 ms, which itself already violates the measured worst-case e1. Consequently, the paper does not demonstrate that any chosen operating point satisfies hard real-time constraints; at best it shows soft real-time behavior under the tested scenarios.
minor comments (5)
  1. [Sec. V-A] The text says 'wmin = wmin = 0.5' where one occurrence should presumably be a different symbol or the sentence should be rephrased.
  2. [Sec. IV, Definition IV.3] The symbol '≺' used for element-wise comparison is nonstandard; the manuscript would benefit from a formal definition or a sentence explaining the notation.
  3. [Sec. II-E] The pseudocode in Alg. 1 has the input variable 'saf eon', which appears to be a spacing artifact; this should be corrected to 'safe_o n' or 'safeon' consistently.
  4. [Sec. III-A and Alg. 1] The term 'hard real-time' is applied to T2 even though T2 is forcibly terminated at the next release of T1; this is an unconventional use of the term and should be clarified.
  5. [Sec. V-D] The figures (Figs. 13 and 14) are informative but the captions do not clearly state the meaning of the 'criteria' axes or the distinction between solid and dotted lines; more detailed captions would help.

Circularity Check

1 steps flagged · score 4.0 of 10

The schedulability proof is tautological: Eq. (4) is reduced to 1 <= 1 by defining e2 = T - e1, so the hard real-time guarantee is not actually derived from worst-case timing bounds; the safety theorem itself is independent.

  1. self definitional [Section III, Eqs. (3)-(4) (Scheduling Framework)]
    "Post-T1, the server checks line 20 from Alg. 1, and, if needed, T2 must be executed by lines 21–27 during the time e2 = T − e1, (3) ... Substituting (3) into (4) yields 1 ≤ 1, proving schedulability for this setup."

    The schedulability condition is not evaluated from independent worst-case execution times. Instead, Task 2's execution budget is defined as the leftover T - e1, so Eq. (4) becomes e1/T + (T - e1)/T = 1 identically. The 'proof' is therefore a restatement of the definition of e2, not a derivation of a bound on e1 or e2. The only nontrivial requirement, e1 <= T, is never established analytically; in fact, the paper reports unconstrained maxima e1 ≈ 70-73 ms while later selecting T* = 50 ms, so the measured data do not satisfy that requirement. The formal hard real-time guarantee reduces to a definitional identity rather than to a schedulability analysis.

full rationale

The paper's main safety contribution, Theorem 1 and the DEB/DGBur construction, is not circular: collision freedom is defined through DEB membership in Eq. (5) and verified by Algorithms 2 and 3 using fresh distance data at each spline, yielding a conservative sufficient-condition argument rather than a fit to test data. The reliance on the authors' prior structures [50], [51], [57] is normal use of previously published building blocks; no uniqueness theorem or fitted parameter is imported to force the new safety result. The only load-bearing step that reduces by construction is the schedulability proof in Section III: once e2 is defined as T - e1, Eq. (4) is an algebraic identity, so the formal 'hard real-time' guarantee is not inferred from a worst-case execution-time bound. The paper's own measured e1 values and the chosen T* = 50 ms make this more than a stylistic issue, but the safety theorem and the experimental validation remain independent of that tautology. I therefore score 4 rather than 6 because the paper's strongest safety claim does not reduce to the circular step, while the schedulability argument partially does.

Assumptions & free parameters 8 free parameters · 6 assumptions · 2 invented entities

The central claims rest on a small set of hand-picked parameters (vobs, T, Delta_t, horizon parameters) and domain assumptions about convex obstacles, bounded obstacle velocity, and conservative distance estimates. The most fragile input is the task period T, which is selected from simulation data even though the measured worst-case T1 time exceeds it at the chosen operating point.

free parameters (8)
  • max obstacle velocity vobs = 1.6 m/s (simulations), 0.5 m/s (human-collaboration experiment)
    Assumed upper bound on obstacle speed; the DEB safety guarantee shrinks as vobs grows and fails if actual speed exceeds it. Selected from ISO/TS 15066 recommendations rather than fitted to data.
  • task period T = 50 ms for DRGBT(1) (20 Hz)
    Chosen as a 'sweet spot' from the simulation study trading success rate against deadline misses; at T=50 ms the measured max T1 execution time (about 70 ms) exceeds the deadline, so this is a hand-picked operating point.
  • time step Delta_t = 100 us
    Discretization step for dynamic bur collision checks (Alg. 2); safety guarantee is only as tight as this discretization.
  • initial horizon size Nh0 = 10
    Algorithm parameter controlling number of horizon nodes; affects exploration and runtime.
  • critical distance dcrit = 0.05 m
    Threshold for marking horizon nodes as critical; affects replanning and horizon growth (Eq. 1).
  • weight threshold wmin = 0.5
    Threshold for deciding whether to replan based on node weights.
  • max attempts for bad/critical nodes = 10
    Limits horizon re-sampling attempts.
  • maximal chain size K in Alg. 3 = 5 (used in Fig. 8 example)
    Limits the number of concatenated dynamic burs in the generalized bur; safety coverage depends on this parameter.
assumptions (6)
  • domain assumption Obstacles are represented as a finite set of convex world obstacles WO_j
    Stated in Sec. II-A; non-convex obstacles require decomposition or BVH primitives.
  • domain assumption Obstacle velocity magnitude is bounded by vobs and no new obstacles appear in W_safe during an iteration
    Sec. IV-A states this directly; the DEB safety proof depends on it.
  • domain assumption Distance-to-obstacle profile is piecewise constant per link and computed via distance queries
    Sec. II-E uses per-link distances d_i; the DEB definition (Eq. 5) depends on these values being conservative.
  • standard math EDF schedulability condition sum(e_i/min(D_i,T_i)) <= 1 is sufficient for all deadlines on a single core
    Invoked in Sec. III, Eq. (2), citing [58].
  • domain assumption The measured maximum execution time e1 of Task 1 from randomized trials upper-bounds future executions
    Sec. III and V-B use e1 from 15,000 runs to set deadlines; the paper admits larger obstacle counts may exceed it.
  • domain assumption Robot links are approximated by bounding capsules and collision checks are discrete with step Delta_t
    Sec. V-A.4 and Sec. IV-D; the safety guarantee is approximate, relying on capsule margins to cover discretization.
invented entities (2)
  • Dynamic Expanded Bubble (DEB)
    purpose: Time-varying collision-free region around a configuration that accounts for obstacle motion up to speed vobs; used to certify safe splines and emergency stops.
    Defined in Def. IV.3 (Eq. 5); no external evidence, but it is a mathematical tool internally validated by simulation and experiments.
  • Dynamic generalized bur (DGBur)
    purpose: Chain of dynamic expanded bubbles used to certify longer safe trajectory segments (Alg. 3).
    Defined in Def. IV.5; same status as DEB.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Real-Time Sampling-Based Safe Motion Planning for Robotic Manipulators in Dynamic Environments." pith.science (2026). https://pith.science/paper/JYUFWWS7

@misc{pith2026250100507,
  author       = {Pith},
  title        = {Pith review of: Real-Time Sampling-Based Safe Motion Planning for Robotic Manipulators in Dynamic Environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JYUFWWS7}},
  note         = {Machine review of arXiv:2501.00507}
}
read the original abstract

In this paper, we present the main features of Dynamic Rapidly-exploring Generalized Bur Tree (DRGBT) algorithm, a sampling-based planner for dynamic environments. We provide a detailed time analysis and appropriate scheduling to facilitate a real-time operation. To this end, an extensive analysis is conducted to identify the time-critical routines and their dependence on the number of obstacles. Furthermore, information about the distance to obstacles is used to compute a structure called dynamic expanded bubble of free configuration space, which is then utilized to establish sufficient conditions for a guaranteed safe motion of the robot while satisfying all kinematic constraints. An extensive randomized simulation trial is conducted to compare the proposed algorithm to a competing state-of-the-art method. Finally, an experimental study on a real robot is carried out covering a variety of scenarios including those with human presence. The results show the effectiveness and feasibility of real-time execution of the proposed motion planning algorithm within a typical sensor-based arrangement, using cheap hardware and sequential architecture, without the necessity for GPUs or heavy parallelization.

Figures

Figures reproduced from arXiv: 2501.00507 by the authors.

Figure 1
Figure 1. Generalized bur with 30 “spines” (radial directions) and 5 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Graphical interpretation of some DRGBT components: obstacles (black); predefined/traversed path (gray/green); local (reached) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Generated horizon spines (solid red lines) in the current [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: EDF scheduling of T1 and T2 with two priority levels. popular. They are applied in many real-time operating systems due to their simplicity and predictability. Considering a wide variety of available scheduling algorithms and the correspond￾ing theoretical analysis of …
Figure 5
Figure 5. Figure 5: Workspace of a planar 2-DoF robot in four different config [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Process of computing dynamic bur (solid black lines). [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: (a) Process of computing dynamic generalized bur (solid yellow lines); (b,c,d) The nearest points, [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: The set of collision-free trajectories in [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Using bisection method to find a safe spline. [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 11
Figure 11. Figure 11: Corresponding CDFs for the routines: replan, computeDistances, generateGBur, generateHorizon, updateHorizon and updateCurrState. Execution times are indicated in each subfigure. Legend: DRGBT (red) and DRGBT-safe (green) [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 13
Figure 13. Figure 13 [PITH_FULL_IMAGE:figures/full_fig_p013_13.png]
Figure 15
Figure 15. Figure 15: Criteria w.r.t. Nobs for DRGBT(1) and DRGBT(2) with and without replanning procedure (solid and dotted line, respectively). processor time to T1 (e.g., compare dashed black lines for u1 = 1 and u1 = 0.5). This further supports the choice of setting the higher priority…
Figure 17
Figure 17. Figure 17: Three additional scenarios used in the comparison study. TABLE II: Results for Scenarios 1 and 2 from the comparison study. Scenario 1 Scenario 2 DRGBT MARS RRTX DRGBT MARS RRTX η [%] 94.5 25.8 9 51.6 14 17.9 Talg [s] 6.5 ± 4.5 7.2 ± 1.6 7.4 ± 3.4 16.5 ± 7.7 8.9 ± 1.9…
Figure 18
Figure 18. Figure 18: Architecture of the used real system. Gray text/lines/blocks indicate specific features used within our experimental study. Black [PITH_FULL_IMAGE:figures/full_fig_p017_18.png]
Figure 19
Figure 19. Figure 19: The view from both the left (a) and right camera (b), as well as their overlapped/combined view (c). The obstacle – a car with the [PITH_FULL_IMAGE:figures/full_fig_p017_19.png]
Figure 20
Figure 20. Figure 20: Measured velocities for each joint from Scenario 1. [PITH_FULL_IMAGE:figures/full_fig_p017_20.png]
Figure 21
Figure 21. Figure 21: Snapshots from Scenario 1 – “Moving-car with balloon”. [PITH_FULL_IMAGE:figures/full_fig_p018_21.png]
Figure 23
Figure 23. Figure 23: Snapshots from Scenario 3 – “Human as an obstacle”. [PITH_FULL_IMAGE:figures/full_fig_p018_23.png]
Figure 24
Figure 24. Figure 24: Snapshots from Scenario 4 – “Safe collaboration/coexistence with human”. [PITH_FULL_IMAGE:figures/full_fig_p018_24.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

68 extracted references · 63 canonical work pages

  1. [1]

    Real-time obstacle avoidance for manipulators and mobile robots,

    O. Khatib, “Real-time obstacle avoidance for manipulators and mobile robots,” The international journal of robotics research, vol. 5, no. 1, pp. 90–98, 1986

    1986

  2. [2]

    Toward real-time path planning in changing environments,

    P. Leven and S. Hutchinson, “Toward real-time path planning in changing environments,” in Algorithmic and Computational Robotics: New Directions: The Fourth International Workshop on the Algorithmic Foundations of Robotics, 2001, pp. 363–376

    work page 2001

  3. [3]

    A framework for real-time path planning in changing environ- ments,

    ——, “A framework for real-time path planning in changing environ- ments,” The International Journal of Robotics Research, vol. 21, no. 12, pp. 999–1030, 2002

    work page 2002

  4. [4]

    Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,

    L. E. Kavraki, P. Svestka, J.-C. Latombe, and M. H. Overmars, “Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,” IEEE transactions on Robotics and Automation , vol. 12, no. 4, pp. 566–580, 1996

    1996

  5. [5]

    A PRM-based motion planner for dynamically changing environments,

    L. Jaillet and T. Sim ´eon, “A PRM-based motion planner for dynamically changing environments,” in 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)(IEEE Cat. No. 04CH37566) , vol. 2. IEEE, 2004, pp. 1606–1611

    work page 2004

  6. [6]

    Rapidly-exploring random trees: A new tool for path planning,

    S. M. LaValle, “Rapidly-exploring random trees: A new tool for path planning,” 1998

    1998

  7. [7]

    An efficient neural network method for real-time motion planning with safety consideration,

    S. X. Yang and M. Meng, “An efficient neural network method for real-time motion planning with safety consideration,” Robotics and Autonomous Systems, vol. 32, no. 2-3, pp. 115–128, 2000

    work page 2000

  8. [8]

    Decomposition-based motion planning: A framework for real-time motion planning in high-dimensional configura- tion spaces,

    O. Brock and L. E. Kavraki, “Decomposition-based motion planning: A framework for real-time motion planning in high-dimensional configura- tion spaces,” in Proceedings 2001 ICRA. IEEE International Conference on Robotics and Automation (Cat. No. 01CH37164), vol. 2. IEEE, 2001, pp. 1469–1474

    work page 2001

Show all 68 references
  1. [9]

    Motion planning using dynamic roadmaps,

    M. Kallman and M. Mataric, “Motion planning using dynamic roadmaps,” in IEEE International Conference on Robotics and Automa- tion, 2004. Proceedings. ICRA ’04. 2004 , vol. 5, April 2004, pp. 4399– 4404 V ol.5

    2004

  2. [10]

    RRT-Connect: An efficient approach to single-query path planning,

    J. J. Kuffner and S. M. LaValle, “RRT-Connect: An efficient approach to single-query path planning,” in Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automa- tion. (Cat. No. 00CH37065) , vol. 2. IEEE, 2000, pp. 995–1001

    2000

  3. [11]

    A dynamic subgoal path planner for unpredictable environments,

    H. Liu, W. Wan, and H. Zha, “A dynamic subgoal path planner for unpredictable environments,” in 2010 IEEE International Conference on Robotics and Automation . IEEE, 2010, pp. 994–1001

    2010

  4. [12]

    HDRM: A resolution complete dynamic roadmap for real-time motion planning in complex scenes,

    Y . Yang, W. Merkt, V . Ivan, Z. Li, and S. Vijayakumar, “HDRM: A resolution complete dynamic roadmap for real-time motion planning in complex scenes,” IEEE Robotics and Automation Letters , vol. 3, no. 1, pp. 551–558, 2017

    2017

  5. [13]

    Distance- aware dynamically weighted roadmaps for motion planning in unknown environments,

    A. Knobloch, N. Vahrenkamp, M. W ¨achter, and T. Asfour, “Distance- aware dynamically weighted roadmaps for motion planning in unknown environments,” IEEE Robotics and Automation Letters, vol. 3, no. 3, pp. 2016–2023, 2018

    2016

  6. [14]

    Real-time randomized path planning for robot navigation,

    J. Bruce and M. Veloso, “Real-time randomized path planning for robot navigation,” in IEEE/RSJ international conference on intelligent robots and systems, vol. 3. IEEE, 2002, pp. 2383–2388

    2002

  7. [15]

    Replanning with RRTs,

    D. Ferguson, N. Kalra, and A. Stentz, “Replanning with RRTs,” in Proceedings 2006 IEEE International Conference on Robotics and Automation, 2006. ICRA 2006. IEEE, 2006, pp. 1243–1248

    2006

  8. [16]

    The focussed D* algorithm for real-time replanning,

    A. Stentz et al., “The focussed D* algorithm for real-time replanning,” in IJCAI, vol. 95, 1995, pp. 1652–1659

    1995

  9. [17]

    Multipartite RRTs for rapid replanning in dynamic environments,

    M. Zucker, J. Kuffner, and M. Branicky, “Multipartite RRTs for rapid replanning in dynamic environments,” in Proceedings 2007 IEEE Inter- national Conference on Robotics and Automation . IEEE, 2007, pp. 1603–1609

    2007

  10. [18]

    An incremental learning approach to motion planning with roadmap management,

    Tsai-Yen Li and Yang-Chuan Shie, “An incremental learning approach to motion planning with roadmap management,” in Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292), vol. 4, May 2002, pp. 3411–3416 vol.4

    2002

  11. [19]

    Chance constrained RRT for prob- abilistic robustness to environmental uncertainty,

    B. Luders, M. Kothari, and J. How, “Chance constrained RRT for prob- abilistic robustness to environmental uncertainty,” in AIAA guidance, navigation, and control conference , 2010, p. 8160

    2010

  12. [20]

    Anytime, dynamic planning in high- dimensional search spaces,

    D. Ferguson and A. Stentz, “Anytime, dynamic planning in high- dimensional search spaces,” in Proceedings 2007 IEEE International Conference on Robotics and Automation . IEEE, 2007, pp. 1310–1315

    2007

  13. [21]

    Anytime motion planning using the RRT*,

    S. Karaman, M. R. Walter, A. Perez, E. Frazzoli, and S. Teller, “Anytime motion planning using the RRT*,” in 2011 IEEE International Conference on Robotics and Automation . IEEE, 2011, pp. 1478–1483

    2011

  14. [22]

    Sampling-based algorithms for optimal motion planning,

    S. Karaman and E. Frazzoli, “Sampling-based algorithms for optimal motion planning,” The international journal of robotics research, vol. 30, no. 7, pp. 846–894, 2011

    2011

  15. [23]

    RT-RRT* – A real-time path planning algorithm based on RRT,

    K. Naderi, J. Rajam ¨aki, and P. H ¨am¨al¨ainen, “RT-RRT* – A real-time path planning algorithm based on RRT,” in Proceedings of the 8th ACM SIGGRAPH Conference on Motion in Games , 2015, pp. 113–118

    2015

  16. [24]

    High-frequency replanning un- der uncertainty using parallel sampling-based motion planning,

    W. Sun, S. Patil, and R. Alterovitz, “High-frequency replanning un- der uncertainty using parallel sampling-based motion planning,” IEEE Transactions on Robotics , vol. 31, no. 1, pp. 104–116, 2015

    2015

  17. [25]

    Fast marching tree: A fast marching sampling-based method for optimal motion planning in many dimensions,

    L. Janson, E. Schmerling, A. Clark, and M. Pavone, “Fast marching tree: A fast marching sampling-based method for optimal motion planning in many dimensions,” The International journal of robotics research , vol. 34, no. 7, pp. 883–921, 2015

    2015

  18. [26]

    Online RRT* and Online FMT*: Rapid replanning with dynamic cost,

    B. Chandler and M. A. Goodrich, “Online RRT* and Online FMT*: Rapid replanning with dynamic cost,” in 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2017, pp. 6313–6318

    2017

  19. [27]

    Real-time adaptive motion planning (RAMP) of mobile manipulators in dynamic environments with unforeseen changes,

    J. Vannoy and J. Xiao, “Real-time adaptive motion planning (RAMP) of mobile manipulators in dynamic environments with unforeseen changes,” IEEE Transactions on Robotics , vol. 24, no. 5, pp. 1199– 1212, 2008

    2008

  20. [28]

    Real-time adaptive motion planning for a continuum manipulator,

    J. Xiao and R. Vatcha, “Real-time adaptive motion planning for a continuum manipulator,” in 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems . IEEE, 2010, pp. 5919–5926

    2010

  21. [29]

    Real-time adaptive non-holonomic motion planning in unforeseen dynamic environments,

    S. McLeod and J. Xiao, “Real-time adaptive non-holonomic motion planning in unforeseen dynamic environments,” in 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2016, pp. 4692–4699

    2016

  22. [30]

    RRT x: Asymptotically optimal single-query sampling-based motion planning with quick replanning,

    M. Otte and E. Frazzoli, “RRT x: Asymptotically optimal single-query sampling-based motion planning with quick replanning,” The Interna- tional Journal of Robotics Research , vol. 35, no. 7, pp. 797–822, 2016

    2016

  23. [31]

    Real-time motion plan- ning in the presence of moving obstacles,

    T. Mercy, W. Van Loock, and G. Pipeleers, “Real-time motion plan- ning in the presence of moving obstacles,” in 2016 European Control Conference (ECC). IEEE, 2016, pp. 1586–1591

    2016

  24. [32]

    Horizon-based lazy optimal RRT for fast, efficient replanning in dynamic environment,

    Y . Chen, Z. He, and S. Li, “Horizon-based lazy optimal RRT for fast, efficient replanning in dynamic environment,” Autonomous Robots, vol. 43, no. 8, pp. 2271–2292, 2019

    2019

  25. [33]

    A heuristic rapidly- exploring random trees method for manipulator motion planning,

    C. Yuan, W. Zhang, G. Liu, X. Pan, and X. Liu, “A heuristic rapidly- exploring random trees method for manipulator motion planning,” IEEE Access, vol. 8, pp. 900–910, 2019

    2019

  26. [34]

    An efficient RRT cache method in dynamic environments for path planning,

    C. Yuan, G. Liu, W. Zhang, and X. Pan, “An efficient RRT cache method in dynamic environments for path planning,” Robotics and Autonomous Systems, vol. 131, p. 103595, 2020

    2020

  27. [35]

    Predicted composite signed- distance fields for real-time motion planning in dynamic environments,

    M. N. Finean, W. Merkt, and I. Havoutis, “Predicted composite signed- distance fields for real-time motion planning in dynamic environments,” in Proceedings of the International Conference on Automated Planning and Scheduling, vol. 31, 2021, pp. 616–624

    2021

  28. [36]

    Rmmi: Enhanced obstacle avoidance for reactive mobile manipulation using an implicit neural map,

    N. Marticorena, T. Fischer, J. Haviland, and N. Suenderhauf, “Rmmi: Enhanced obstacle avoidance for reactive mobile manipulation using an implicit neural map,” 2024. [Online]. Available: https: //arxiv.org/abs/2408.16206

    2024 arXiv

  29. [37]

    PiP-X: Funnel-based online feedback motion planning/replanning in dynamic environments,

    M. K. M. Jaffar and M. Otte, “PiP-X: Funnel-based online feedback motion planning/replanning in dynamic environments,” in International Workshop on the Algorithmic Foundations of Robotics. Springer, 2022, pp. 132–148

    2022

  30. [38]

    Collision-free motion planning method based on online trajectory generation in high dimen- sional dynamic workspace,

    H. Liu, D. Qu, F. Xu, Z. Du, K. Jia, and M. Liu, “Collision-free motion planning method based on online trajectory generation in high dimen- sional dynamic workspace,” in 2022 IEEE International Conference on Real-time Computing and Robotics (RCAR) . IEEE, 2022, pp. 93–98

    2022

  31. [39]

    ST-RRT*: Asymptotically-optimal bidirectional motion planning through space- time,

    F. Grothe, V . N. Hartmann, A. Orthey, and M. Toussaint, “ST-RRT*: Asymptotically-optimal bidirectional motion planning through space- time,” in 2022 International Conference on Robotics and Automation (ICRA). IEEE, 2022, pp. 3314–3320

    2022

  32. [40]

    Real- time RRT* with signal temporal logic preferences,

    A. Linard, I. Torre, E. Bartoli, A. Sleat, I. Leite, and J. Tumova, “Real- time RRT* with signal temporal logic preferences,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2023, pp. 8621–8627

    2023

  33. [41]

    Real- time constraint-based planning and control of robotic manipulators for safe human–robot collaboration,

    K. Merckaert, B. Convens, M. M. Nicotra, and B. Vanderborght, “Real- time constraint-based planning and control of robotic manipulators for safe human–robot collaboration,” Robotics and Computer-Integrated Manufacturing, vol. 87, p. 102711, 2024

    2024

  34. [42]

    Online trajectory genera- tion with local replanning for 7-DoF serial manipulator in unforeseen dynamic environments,

    J. Shao, H. Zhang, S. Zhu, and W. Song, “Online trajectory genera- tion with local replanning for 7-DoF serial manipulator in unforeseen dynamic environments,” IEEE Robotics and Automation Letters , 2024

    2024

  35. [43]

    Robot motion planning on a chip,

    S. Murray, W. Floyd-Jones, Y . Qi, D. J. Sorin, and G. D. Konidaris, “Robot motion planning on a chip,” in Robotics: Science and Systems , vol. 6, 2016

    2016

  36. [44]

    CuRobo: Parallelized collision-free robot motion generation,

    B. Sundaralingam, S. K. S. Hari, A. Fishman, C. Garrett, K. Van Wyk, V . Blukis, A. Millane, H. Oleynikova, A. Handa, F. Ramos et al. , “CuRobo: Parallelized collision-free robot motion generation,” in 2023 IEEE International Conference on Robotics and Automation (ICRA) . IEEE...

    2023

  37. [45]

    Motions in microsec- onds via vectorized sampling-based planning,

    W. Thomason, Z. Kingston, and L. E. Kavraki, “Motions in microsec- onds via vectorized sampling-based planning,” in 2024 IEEE Interna- tional Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 8749–8756

    2024

  38. [46]

    Anytime informed multi-path replanning strategy for complex environments,

    C. Tonola, M. Faroni, M. Beschi, and N. Pedrocchi, “Anytime informed multi-path replanning strategy for complex environments,” IEEE Access, vol. 11, pp. 4105–4116, 2023

    2023

  39. [47]

    A survey of robotic motion planning in dynamic environments,

    M. Mohanan and A. Salgoankar, “A survey of robotic motion planning in dynamic environments,” Robotics and Autonomous Systems, vol. 100, pp. 171–185, 2018

    2018

  40. [48]

    A review of motion planning algorithms for intelligent robots,

    C. Zhou, B. Huang, and P. Fr ¨anti, “A review of motion planning algorithms for intelligent robots,” Journal of Intelligent Manufacturing , vol. 33, no. 2, pp. 387–424, 2022

    2022

  41. [49]

    A review of recent trend in motion planning of industrial robots,

    M. G. Tamizi, M. Yaghoubi, and H. Najjaran, “A review of recent trend in motion planning of industrial robots,” International Journal of Intelligent Robotics and Applications , vol. 7, no. 2, pp. 253–274, 2023

    2023

  42. [50]

    Path planning for robotic manipulators in dynamic environments using distance information,

    N. Covic, B. Lacevic, and D. Osmankovic, “Path planning for robotic manipulators in dynamic environments using distance information,” in 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2021, pp. 4708–4713

    2021

  43. [51]

    Improved C-space exploration and path planning for robotic manipulators using distance information,

    B. Lacevic and D. Osmankovic, “Improved C-space exploration and path planning for robotic manipulators using distance information,” in 2020 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2020

    2020

  44. [52]

    Approximate convex decomposition of polyhedra and its applications,

    J.-M. Lien and N. M. Amato, “Approximate convex decomposition of polyhedra and its applications,” Computer Aided Geometric Design , vol. 25, no. 7, pp. 503–522, 2008

    2008

  45. [53]

    FCL: A general purpose library for collision and proximity queries,

    J. Pan, S. Chitta, and D. Manocha, “FCL: A general purpose library for collision and proximity queries,” in 2012 IEEE International Conference on Robotics and Automation . IEEE, 2012, pp. 3859–3866

    2012

  46. [54]

    Quinlan, Real-time modification of collision-free paths

    S. Quinlan, Real-time modification of collision-free paths . Stanford University Stanford, 1994, no. 1537

    1994

  47. [55]

    Burs of free C- space: A novel structure for path planning,

    B. Lacevic, D. Osmankovic, and A. Ademovic, “Burs of free C- space: A novel structure for path planning,” in 2016 IEEE International Conference on Robotics and Automation (ICRA) , 2016, pp. 70–76

    2016

  48. [56]

    Asymptotically optimal path planning for robotic manipulators: Multi-directional, multi-tree approach,

    N. Covic, D. Osmankovic, and B. Lacevic, “Asymptotically optimal path planning for robotic manipulators: Multi-directional, multi-tree approach,” Journal of Intelligent & Robotic Systems , vol. 109, no. 1, pp. 1–19, 2023

    2023

  49. [57]

    Path planning for robotic manipulators using expanded bubbles of free C-space,

    A. Ademovic and B. Lacevic, “Path planning for robotic manipulators using expanded bubbles of free C-space,” in 2016 IEEE International Conference on Robotics and Automation (ICRA) , 2016, pp. 77–82

    2016

  50. [58]

    J. W. S. Liu, Real-Time Systems. Prentice Hall, 2000

    2000

  51. [59]

    P. A. Laplante et al. , Real-time systems design and analysis . Wiley New York, 2004

    2004

  52. [60]

    Mall, Real-time systems: Theory and practice

    R. Mall, Real-time systems: Theory and practice . Pearson Education India, 2009

    2009

  53. [61]

    McNaughton, Parallel algorithms for real-time motion planning

    M. McNaughton, Parallel algorithms for real-time motion planning . Carnegie Mellon University, 2011

    2011

  54. [62]

    nanoflann: a C++ header-only fork of FLANN, a library for nearest neighbor (NN) with KD-trees,

    J. L. Blanco and P. K. Rai, “nanoflann: a C++ header-only fork of FLANN, a library for nearest neighbor (NN) with KD-trees,” https:// github.com/jlblancoc/nanoflann, 2014

    2014

  55. [63]

    Enhancement of force exertion capability of a mobile manipulator by kinematic reconfiguration,

    H. Xing, A. Torabi, L. Ding, H. Gao, Z. Deng, and M. Tavakoli, “Enhancement of force exertion capability of a mobile manipulator by kinematic reconfiguration,” IEEE Robotics and Automation Letters , vol. 5, no. 4, pp. 5842–5849, 2020

    2020

  56. [64]

    Dual-user haptic teleoperation of complementary motions of a redundant wheeled mobile manipulator considering task priority,

    H. Xing, L. Ding, H. Gao, W. Li, and M. Tavakoli, “Dual-user haptic teleoperation of complementary motions of a redundant wheeled mobile manipulator considering task priority,” IEEE Transactions on Systems, Man, and Cybernetics: Systems , vol. 52, no. 10, pp. 6283–6295, 2022

    2022

  57. [65]

    Kinematic control of redundant manipulators: Generalizing the task-priority framework to inequality task,

    O. Kanoun, F. Lamiraux, and P.-B. Wieber, “Kinematic control of redundant manipulators: Generalizing the task-priority framework to inequality task,” IEEE Transactions on Robotics , vol. 27, no. 4, pp. 785–792, 2011

    2011

  58. [66]

    3D is here: Point Cloud Library (PCL),

    R. B. Rusu and S. Cousins, “3D is here: Point Cloud Library (PCL),” in 2011 IEEE international conference on robotics and automation. IEEE, 2011, pp. 1–4

    2011

  59. [67]

    Trajectory generation algorithm for safe human-robot collaboration based on multiple depth sensor measurements,

    M. Ragaglia, A. M. Zanchettin, and P. Rocco, “Trajectory generation algorithm for safe human-robot collaboration based on multiple depth sensor measurements,” Mechatronics, vol. 55, pp. 267–281, 2018

    2018

  60. [68]

    KF-RRT: Obstacles tracking and safe dynamic motion planning for robotic manipulators,

    H. Hadzic, D. Osmankovic, and B. Lacevic, “KF-RRT: Obstacles tracking and safe dynamic motion planning for robotic manipulators,” in 29th International Conference on Information, Communication and Automation Technologies (ICAT). IEEE, 2023. Nermin Covic (Student Member, IEEE) ...

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.