Future-Oriented Navigation: Dynamic Obstacle Avoidance with One-Shot Energy-Based Multimodal Motion Prediction

Emmanuel Dean; Georg Hess; Junjie Hu; Knut {\AA}kesson; Lennart Svensson; Ze Zhang

REVIEW 4 major objections 7 minor 34 references

Future-Oriented Navigation: Dynamic Obstacle Avoidance with One-Shot Energy-Based Multimodal Motion Prediction

T0 review · 4 major / 7 minor · reviewed 2026-08-16 · 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 one-shot energy-based motion predictor feeding an MPC controller achieves higher collision-free success rates for mobile robots in dynamic warehouse and hospital settings than existing obstacle-avoidance approaches.

desk verdict A useful ENLL/PELU variant for MPC-based navigation, but the evaluation overclaims; fix the loss equations, re-run or relabel baselines, and add error bars. read the letter →

arxiv 2505.00237 v3 pith:QDOPEFKK submitted 2025-05-01 cs.RO cs.LGcs.SYeess.SY

Ze Zhang , Georg Hess , Junjie Hu , Emmanuel Dean , Lennart Svensson , Knut {AA}kesson This is my paper · ORCID

classification cs.ROcs.LGcs.SYeess.SY

keywords human-awaremotionplanningcollisionavoidancedeeplearningmethodsenergy-basedmultimodalpredictionmodelpredictivecontroldynamicobstacleone-shot

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 tries to show that a mobile robot can avoid moving humans and vehicles more reliably by replacing iterative per-obstacle prediction with a one-shot energy-based predictor coupled to a model predictive controller. The predictor outputs a full stack of per-timestep probability maps in a single forward pass, and the controller turns those maps into elliptical obstacle constraints after clustering and Gaussian fitting. The authors argue that their energy-based negative-log-likelihood training loss, stabilized by a positive exponential linear unit, produces sharper predictions than BCE or KLD losses and avoids the over-conservative regions that freeze planning. In simulated warehouse and hospital scenarios, the integrated MPC-ENLL controller achieves the highest success rates among the compared methods, with and without cooperative pedestrians. If the results transfer, the approach would give industrial mobile robots a practical, real-time way to act on multimodal forecasts.

What carries the argument

The load-bearing object is the ENLL loss: a pixel-level negative log-likelihood over the predictor's energy map, with the ground-truth pixel replaced by a Gaussian mask so that nearby high-probability pixels are pulled down instead of only one pixel. A Positive Exponential Linear Unit (PELU) output layer clips the exploding exponential by switching to a linear regime, which keeps training finite on large energy maps. The U-Net emits stacked energy maps for all future time steps in one forward pass; Monte Carlo sampling plus clustering-and-Gaussian-fitting (CGF) turns the maps into elliptical predicted-obstacle regions; and the MPC uses these regions both as soft costs over the full horizon and as hard constraints over a short critical horizon. That chain, stable energy training, one-shot maps, geometric clustering, and predictive MPC, is what carries the claimed collision-avoidance improvement.

What would settle it

Run the same MPC-ENLL controller in a real warehouse or hospital with a live tracking system and count collision-free runs: if its success rate drops to the level of constant-velocity or WTA baselines, or the learned maps stop concentrating on actual futures, the central claim is refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that treating multi-step multimodal motion prediction as a large-scale classification over discrete energy maps, trained with a Gaussian-weighted negative log-likelihood and a PELU output layer, yields concentrated probability maps that are stable to train and fast to run. These maps, after sample clustering and Gaussian fitting, become geometric constraints inside an MPC that already handles static obstacles, reference tracking, and fleet collision avoidance. The paper claims this combined pipeline, MPC-ENLL, outperforms the compared obstacle-avoidance methods: it reaches 96% success in a blind-corner scenario, 91% in an abrupt pedestrian-turn scenario, and 90% in a longer hospital scenario, against a 94% and 81% WTA baseline and low-scoring BCE/KLD variants in the hospital case. The claim is specifically that the energy-based predictor's concentration, not just the controller, is what makes the difference.

Load-bearing premise

The load-bearing premise is that the simulated humans and vehicles, modelled as omnidirectional pedestrians on predefined paths with random velocity noise, move realistically enough for the reported 90-96% success rates to transfer to a real deployment.

Editorial extensions

If this is right

Because prediction runs one-shot, the motion predictor takes about 3 ms per object, roughly 19 times faster than the iterative WTA predictor, so the integrated controller can operate within a real-time solver budget.
Because the ENLL loss sharpens the predicted probability maps, it reduces the over-conservative occupied areas produced by BCE/KLD training, lessening the Freezing Robot Problem.
Because nearby predicted futures are grouped by CGF before being sent to MPC, non-convex obstructions from individual predictions are merged into a single convex region, which keeps the robot from freezing.
In the reported simulations, MPC-ENLL reaches 96%, 91%, and 90% success in the three main scenarios, with higher success than the MPC-WTA, DWA-CVM, RPP, TEB-MPC, MPC-nCVM, and MPC-SGAN comparisons.
The formulation also includes distributed fleet collision avoidance, so the same controller coordinates multiple AMRs near pedestrians, as shown in the intersection and long-term robustness tests.

Reading between the lines

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

An implied consequence of the training recipe: the PELU-stabilized pixel-level NLL loss should transfer to other high-resolution prediction tasks that currently suffer from unstable energy-based training, since the instability it fixes is not specific to pedestrian motion.
The Gibbs-Boltzmann inverse-temperature parameter, which the paper mentions but does not tune, provides a testable calibration knob: scaling energies up or down should shift the predictor between concentrated and conservative forecasts, directly affecting robot clearance and success rate.
Because the pipeline assumes a bird's-eye-view tracker as input, its real-world behavior is an upper bound set by perception quality; the paper lists vision-and-tracking integration as future work, so the next decisive test is end-to-end operation with real detections.

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.

Desk Editor's Note

read the letter

Colleague,

Worth a read if you care about learned predictors for MPC in dynamic environments. The new piece is the ENLL loss with the PELU output layer: it replaces the usual BCE/KLD training of occupancy maps with an energy-based loss that produces more concentrated predictions, and the paper shows—both on a synthetic crossroad and in a warehouse simulator—that concentrated maps lead to better downstream collision avoidance than over-spread maps. That is a genuine variant, not a parameter scan, and the code and video are real artifacts. Credit where due: the authors also show their CGF grouping helps avoid freezing when multiple pedestrians approach together.

The soft spots are in the evaluation. The headline 'outperforms other popular approaches' is broader than the evidence. The external baselines MPC-WTA* and DWA-CVM* are lifted from their earlier paper [11], not re-run in this codebase, so the comparison is not controlled. Scenarios 3 and 4 contain no external baselines at all. And under the cooperative pedestrian model in Scenario 2, the paper itself states all methods achieve perfect collision avoidance—so the 'with or without cooperative pedestrian model' part of the conclusion is directly contradicted by their own table. No confidence intervals are given on the success percentages, though with 100 runs per scenario the binomial error bars would be easy to compute. Also, Eq. (16) and (17) don't line up: the denominator in (16) sums e^{-E}, but the loss in (17) is the log of a weighted sum of \bar A e^{-E}, which is not the NLL of the distribution defined in (16). The idea is clear, but an exact reproduction from the text would trip on this.

The synthetic evaluation with an omnidirectional pedestrian model moving on predefined paths is another limit; the authors acknowledge simulation-only in their own limitation statement, so that is not a hidden flaw, but it does mean the 90-96% numbers are not evidence about real deployment.

My recommendation: send it to peer review—the core idea is worth refereeing and the artifacts support it—but the authors should be required to fix the loss equations, clearly separate borrowed baselines from re-run ones, report variance, and narrow the comparative claim to what the experiments actually show. As it stands, it is a conditional accept, not a clean one.

Referee Report

4 major / 7 minor

Summary. The paper proposes an integrated navigation pipeline for mobile robots in dynamic environments: a U-Net-based energy-based motion predictor trained with a novel ENLL loss and PELU output layer, a clustering/Gaussian fitting (CGF) module that converts predicted probability maps into multimodal elliptical obstacles, and an MPC formulation with soft and hard avoidance terms plus fleet collision avoidance. The method is evaluated in ROS2/Gazebo simulations of warehouse and hospital scenarios, with success rates and additional metrics compared against baselines such as MPC-WTA*, DWA-CVM*, RPP, TEB-MPC, MPC-nCVM, MPC-SGAN, and self-variants trained with BCE and KLD losses. The paper claims that the combined MPC-ENLL approach outperforms other popular obstacle avoidance approaches with and without cooperative pedestrian models.

Significance. If the comparative claim held, this would be a practical contribution: it offers an off-the-shelf integration of one-shot multimodal motion prediction into MPC, and the energy-based ENLL objective is a plausible way to obtain more concentrated predictions that are easier for downstream planners to use. The paper's strengths include a public code repository and simulation environments, a clean internal comparison of loss formulations (BCE vs KLD vs ENLL), an explicit inference-time measurement, and a qualitative demonstration that CGF grouping mitigates the freezing-robot problem. The main weakness is that the evaluation as reported does not yet establish the headline outperformance claim over external methods, and one part of the cooperative-pedestrian conclusion is contradicted by the paper's own text.

major comments (4)

[§VII-B, Table II] The central claim that MPC-ENLL outperforms other popular obstacle avoidance approaches is not supported by the reported controlled comparisons. The external baselines MPC-WTA* and DWA-CVM* are taken from Ref. [11] rather than re-run in the current framework; Scenario 3 has no external baseline at all; and Scenario 4 only compares the BCE/KLD/ENLL self-variants. In addition, the success percentages in Table II are reported without confidence intervals or significance tests. With 100 runs per scenario, the Scenario 1 difference of 96% versus 94% is well within binomial sampling noise, and even the larger Scenario 2 difference (91% versus 81–84%) is not shown to be statistically reliable. The paper should re-run all external baselines under identical conditions, add baselines to Scenarios 3 and 4, and report confidence intervals or significance tests before claiming outperformance.
[§VII-B, §VIII] The cooperative-pedestrian part of the conclusion is contradicted by the paper's own text. Section VII-B states that under the cooperative Social Force model, 'in Scenario 2, all methods achieve perfect collision avoidance.' If success rate is the criterion, MPC-ENLL does not outperform the other methods in that setting. If the outperformance claim is meant to rest on other metrics such as smoothness, clearance, or solving time, those numbers are not reported for the cooperative runs. The conclusion in Section VIII should be qualified or supported with the corresponding cooperative-pedestrian metrics.
[§V-A, Eqs. (16)–(17)] There is a mathematical inconsistency in the derivation of the NLL loss. Eq. (16) defines \bar P_{w,h} as a per-pixel normalized probability, but Eq. (17) then defines L_NLL as -ln \sum_{w',h'} \bar A_{w',h'} e^{-E_{w',h'}} + ln \sum_{w',h'} e^{-E_{w',h'}}, which is not -ln \bar P as written. The subsequent ENLL loss in Eq. (20) is a different expression again. The authors should state explicitly what probabilistic model the ENLL loss corresponds to—for example, a weighted log-sum-exp objective over all pixels—and correct or re-derive Eqs. (16)–(17). This is load-bearing because the training loss is the paper's main methodological novelty.
[§VI-A, §VII-B, §VIII] The evaluation is entirely in a synthetic simulator whose pedestrians are omnidirectional and move along predefined paths with random velocity noise. The paper itself acknowledges at the end of Section VIII that only simulated environments were used. Since the success-rate claim is the principal evidence for the practical advantage of MPC-ENLL, the authors should either temper the practical conclusion or provide evidence of transferability, for example by testing with more realistic human-motion models, real-world trajectory data, or a sensitivity analysis over simulator parameters.

minor comments (7)

[Abstract] The abstract contains a typo: 'performance is accessed' should read 'performance is assessed.'
[§VII-E] The heading 'In this session' should be 'In this section.'
[Table II] The caption note 'Scenario 3 shows worse data in two robots' is unclear; please specify whether success is counted per robot or per episode and clarify what 'worse data' means.
[§IV-A, Eq. (2)] The mode-dependent weight \alpha_m is introduced in Eq. (2), but Section VI-B says it is 'currently not used and set to 1.' Either use this parameter or remove it from the formulation to avoid confusion.
[§VII-B, Table II] The claim that 'MPC-ENLL maintains a large clearance from dynamic obstacles' is not supported by Scenario 1 of Table II, where MPC-WTA* reports a larger dynamic clearance (0.786 m versus 0.704 m); please clarify or adjust this claim.
[Table II] The table lists 'MPC-SGAN [9]' while the text in Section VII-B refers to Social-GAN [16]; please reconcile the citation so the reader knows which implementation was used.
[§VI-A] The variance of the Gaussian mask (10 pixels) and the PELU offset \epsilon are important hyperparameters for conservativeness and training stability; they should be reported in the paper rather than only in the configuration file of the repository.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ENLL training objective is a standard supervised loss with a PELU stabilization, and the downstream comparisons are either external baselines or ablations; self-citations are structural rather than load-bearing.

full rationale

The paper's central predictive contribution is the ENLL loss in Eq. (20), obtained from the Gibbs-Boltzmann form in Eq. (12), a discretized normalization in Eq. (13), a masked NLL in Eq. (17), and a PELU output layer in Eq. (18). This is a training objective, not a metric that contains the conclusion; the model is optimized on supervised ground-truth positions and then evaluated. The prediction comparison in Table I uses a GMM-fitted ground-truth distribution on a synthetic crossroad, which is an external evaluation target rather than a quantity derived from the ENLL model itself. In the obstacle-avoidance evaluation, Table II includes non-self baselines RPP, TEB, MPC-nCVM, and MPC-SGAN, plus ablations MPC-BCE and MPC-KLD that share the same MPC but differ in loss. The starred MPC-WTA* and DWA-CVM* numbers are borrowed from the authors' prior paper [11], but that is a published, independently reported baseline, not a definitional identity; no equation in the present paper forces MPC-ENLL's success rate to equal those prior numbers. The MPC formulation and CGF grouping are inherited from [11], [23], and [24], but this is normal building on prior work rather than a circular reduction, since the claimed novelty and the measured improvements come from the new energy-based predictor and its integration. The conclusion's cooperative-pedestrian claim is somewhat overstated relative to the reported 'all methods achieve perfect collision avoidance' in Scenario 2, but that is an evidence and wording concern, not a circularity. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors, and no ansatz is smuggled in via citation.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The method rests on several tunable parameters, dominated by the Gaussian mask variance, MPC weights, horizons, and the PELU offset. There are no newly postulated physical entities. The main conceptual assumptions are that image-resolution summation approximates the partition function, that the synthetic Gaussian mask is a valid label proxy, that synthetic pedestrian motion transfers to reality, and that truncated MPC solves remain safe.

free parameters (8)

Gaussian mask variance = 10 pixels
Used to construct the relaxed label masks A in BCE, KLD, and ENLL losses (Section VI-A). It controls how concentrated the supervised target is and affects overestimation behavior.
Prediction horizon N = 20 steps (4 s)
Defines the number of stacked future probability maps and the MPC horizon.
Critical horizon N_crit = 5
Hard dynamic-obstacle avoidance is enforced only for the first 5 steps; larger values would risk infeasibility.
Sampling time = 0.2 s
Discrete-time step for MPC and the motion model (Section VI-B).
Time weight beta_tk = predefined, decreasing with step
Scales the soft dynamic obstacle cost in Eq. (2); the exact schedule is only in the config file.
Mode weight alpha_m = 1 (unused)
The paper explicitly states the model weight is currently not used, so predicted mode probabilities do not enter Eq. (2).
MPC weights (Qs, Qu, Qa, QD, Qf) and margins (dfleet, r_extra) = not given, see repo config
Set the trade-off between tracking, smoothness, obstacle clearance, and fleet separation.
PELU epsilon = small positive offset
Prevents the PELU output from reaching zero so the logarithm in ENLL stays finite; the exact value is not stated.

assumptions (5)

domain assumption The partition function in Eq. (12) is accurately approximated by the pixel summation in Eq. (13) at image resolution.
No error analysis is given; the approximation relies on the grid resolution being fine enough.
domain assumption The Gaussian mask A is a valid proxy for the unobservable true future-position distribution.
The paper itself says this 'typically cannot be validated in the real world' (Section V-A).
domain assumption Synthetic omnidirectional pedestrian trajectories with random velocity noise represent real dynamic obstacles.
Section VI-A uses 580-660 synthetic trajectories for warehouse and hospital scenes; the conclusion admits all evaluation is simulated.
domain assumption A 0.1 s solver time limit does not invalidate MPC solutions.
Section VII-B admits that premature solutions without the collision-free guarantee are possible.
domain assumption Monte Carlo sampling from probability maps followed by GMM or ellipse fitting captures the relevant future modes.
CGF grouping in Section V-B assumes the sampled clusters are sufficient statistics for planning.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Future-Oriented Navigation: Dynamic Obstacle Avoidance with One-Shot Energy-Based Multimodal Motion Prediction." pith.science (2026). https://pith.science/paper/QDOPEFKK

@misc{pith2026250500237,
  author       = {Pith},
  title        = {Pith review of: Future-Oriented Navigation: Dynamic Obstacle Avoidance with One-Shot Energy-Based Multimodal Motion Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QDOPEFKK}},
  note         = {Machine review of arXiv:2505.00237}
}

read the original abstract

This paper proposes an integrated approach for the safe and efficient control of mobile robots in dynamic and uncertain environments. The approach consists of two key steps: one-shot multimodal motion prediction to anticipate motions of dynamic obstacles and model predictive control to incorporate these predictions into the motion planning process. Motion prediction is driven by an energy-based neural network that generates high-resolution, multi-step predictions in a single operation. The prediction outcomes are further utilized to create geometric shapes formulated as mathematical constraints. Instead of treating each dynamic obstacle individually, predicted obstacles are grouped by proximity in an unsupervised way to improve performance and efficiency. The overall collision-free navigation is handled by model predictive control with a specific design for proactive dynamic obstacle avoidance. The proposed approach allows mobile robots to navigate effectively in dynamic environments. Its performance is accessed across various scenarios that represent typical warehouse settings. The results demonstrate that the proposed approach outperforms other existing dynamic obstacle avoidance methods.

Figures

Figures reproduced from arXiv: 2505.00237 by the authors.

**Figure 2.** Training results on the Stanford Drone Dataset [25] with different [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Comparison of training results on the synthetic crossroad dataset from different learning strategies. The first column shows the scenario (top) with [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Scenario 3 in obstacle avoidance evaluation. Four pedestrians and two AMRs are crossing an intersection, which causes a busy area and is challenging [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Success rates of different methods with and without the cooperative [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: Comparison between grouped prediction (left) and individual predic [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

**Figure 7.** Figure 7: Gazebo simulation of Scenarios 3 (up) and 4 (down). Red circles and [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 29 canonical work pages

[11]

Prescient collision- free navigation of mobile robots with iterative multimodal motion prediction of dynamic obstacles,

Z. Zhang, H. Hajieghrary, E. Dean, and K. ˚Akesson, “Prescient collision- free navigation of mobile robots with iterative multimodal motion prediction of dynamic obstacles,”Robotics and Automation Letters, vol. 8, no. 9, pp. 5488–5495, 2023

work page 2023
[1]

Automated guided vehicle systems, state-of-the-art control algorithms and techniques,

M. De Ryck, M. Versteyhe, and F. Debrouwere, “Automated guided vehicle systems, state-of-the-art control algorithms and techniques,” Journal of Manufacturing Systems, vol. 54, pp. 152–173, 2020

work page 2020
[2]

Unfreezing the robot: Navigation in dense, interacting crowds,

P. Trautman and A. Krause, “Unfreezing the robot: Navigation in dense, interacting crowds,” inIROS, 2010, pp. 797–803

work page 2010
[3]

Fro- zone: Freezing-free, pedestrian-friendly navigation in human crowds,

A. J. Sathyamoorthy, U. Patel, T. Guan, and D. Manocha, “Fro- zone: Freezing-free, pedestrian-friendly navigation in human crowds,” Robotics and Automation Letters, vol. 5, no. 3, pp. 4352–4359, 2020

work page 2020
[4]

Intention aware robot crowd navigation with attention-based interaction graph,

S. Liu, P. Chang, Z. Huang, N. Chakraborty, K. Hong, W. Liang, D. L. McPherson, J. Geng, and K. Driggs-Campbell, “Intention aware robot crowd navigation with attention-based interaction graph,” inICRA, 2023, pp. 12 015–12 021

work page 2023
[5]

Core challenges of social robot navigation: A survey,

C. Mavrogiannis, F. Baldini, A. Wang, D. Zhao, P. Trautman, A. Stein- feld, and J. Oh, “Core challenges of social robot navigation: A survey,” ACM Transactions on Human-Robot Interaction, vol. 12, no. 3, 2023

work page 2023
[6]

Dynamic channel: A planning framework for crowd navigation,

C. Cao, P. Trautman, and S. Iba, “Dynamic channel: A planning framework for crowd navigation,” inICRA, 2019, pp. 5551–5557

2019
[7]

Avoiding dense and dynamic obstacles in enclosed spaces: Application to moving in crowds,

L. Huber, J.-J. Slotine, and A. Billard, “Avoiding dense and dynamic obstacles in enclosed spaces: Application to moving in crowds,”IEEE Transactions on Robotics, vol. 38, no. 5, pp. 3113–3132, 2022

2022

Show all 34 references

[8]

Robot motion planning in dynamic, uncertain environments,

N. E. Du Toit and J. W. Burdick, “Robot motion planning in dynamic, uncertain environments,”IEEE Transactions on Robotics, vol. 28, no. 1, pp. 101–115, 2012

2012
[9]

Proactive model predictive control with multi-modal human motion prediction in cluttered dynamic environments,

L. Heuer, L. Palmieri, A. Rudenko, A. Mannucci, M. Magnusson, and K. O. Arras, “Proactive model predictive control with multi-modal human motion prediction in cluttered dynamic environments,” inIROS, 2023, pp. 229–236

2023
[10]

Group-based motion prediction for navigation in crowded environments,

A. Wang, C. Mavrogiannis, and A. Steinfeld, “Group-based motion prediction for navigation in crowded environments,” inConference on Robot Learning (CoRL), 2022, pp. 871–882

2022
[12]

What the constant ve- locity model can teach us about pedestrian motion prediction,

C. Sch ¨oller, V . Aravantinos, F. Lay, and A. Knoll, “What the constant ve- locity model can teach us about pedestrian motion prediction,”Robotics and Automation Letters, vol. 5, no. 2, pp. 1696–1703, 2020

2020
[13]

Social force model for pedestrian dynamics,

D. Helbing and P. Molnar, “Social force model for pedestrian dynamics,” Physical review E, vol. 51, no. 5, p. 4282, 1995

1995
[14]

Reciprocal velocity obstacles for real-time multi-agent navigation,

J. van den Berg, M. Lin, and D. Manocha, “Reciprocal velocity obstacles for real-time multi-agent navigation,” inICRA, 2008, pp. 1928–1935

2008
[15]

Social LSTM: Human trajectory prediction in crowded spaces,

A. Alahi, K. Goel, V . Ramanathan, A. Robicquet, L. Fei-Fei, and S. Savarese, “Social LSTM: Human trajectory prediction in crowded spaces,” inCVPR, June 2016

2016
[16]

Social GAN: Socially acceptable trajectories with generative adversarial net- works,

A. Gupta, J. Johnson, L. Fei-Fei, S. Savarese, and A. Alahi, “Social GAN: Socially acceptable trajectories with generative adversarial net- works,” inCVPR, June 2018

2018
[17]

Overcoming limitations of mixture density networks: A sampling and fitting framework for multimodal future prediction,

O. Makansi, E. Ilg, O. Cicek, and T. Brox, “Overcoming limitations of mixture density networks: A sampling and fitting framework for multimodal future prediction,” inCVPR, 2019

2019
[18]

From goals, waypoints & paths to long term human trajectory forecasting,

K. Mangalam, Y . An, H. Girase, and J. Malik, “From goals, waypoints & paths to long term human trajectory forecasting,” inICCV, 2021, pp. 15 213–15 222

2021
[19]

Scene compliant trajectory forecast with agent-centric spatio-temporal grids,

D. Ridel, N. Deo, D. Wolf, and M. Trivedi, “Scene compliant trajectory forecast with agent-centric spatio-temporal grids,”Robotics and Automa- tion Letters, vol. 5, no. 2, pp. 2816–2823, 2020

2020
[20]

Trajectron++: Dynamically-feasible trajectory forecasting with heterogeneous data,

T. Salzmann, B. Ivanovic, P. Chakravarty, and M. Pavone, “Trajectron++: Dynamically-feasible trajectory forecasting with heterogeneous data,” in European Conference on Computer Vision (ECCV). Springer, 2020, pp. 683–700

2020
[21]

End-to-end trajectory distribution predic- tion based on occupancy grid maps,

K. Guo, W. Liu, and J. Pan, “End-to-end trajectory distribution predic- tion based on occupancy grid maps,” inCVPR, 2022, pp. 2232–2241

2022
[22]

Bird’s-eye-view trajectory plan- ning of multiple robots using continuous deep reinforcement learning and model predictive control,

K. Ceder, Z. Zhang, A. Burman, I. Kuangaliyev, K. Mattsson, G. Nyman, A. Peters´en, L. Wisell, and K. ˚Akesson, “Bird’s-eye-view trajectory plan- ning of multiple robots using continuous deep reinforcement learning and model predictive control,” inIROS, 2024, pp. 8002–8008

2024
[23]

Motion predic- tion based on multiple futures for dynamic obstacle avoidance of mobile robots,

Z. Zhang, E. Dean, Y . Karayiannidis, and K. ˚Akesson, “Motion predic- tion based on multiple futures for dynamic obstacle avoidance of mobile robots,” inCASE, 2021, pp. 475–481

2021
[24]

Centralized versus distributed nonlinear model predictive control for online robot fleet trajectory planning,

F. Bertilsson, M. Gordon, J. Hansson, D. M ¨oller, D. S ¨oderberg, Z. Zhang, and K. ˚Akesson, “Centralized versus distributed nonlinear model predictive control for online robot fleet trajectory planning,” in CASE, 2022, pp. 701–706

2022
[25]

Learning social etiquette: Human trajectory understanding in crowded scenes,

A. Robicquet, A. Sadeghian, A. Alahi, and S. Savarese, “Learning social etiquette: Human trajectory understanding in crowded scenes,” inEuropean Conference on Computer Vision (ECCV). Springer, 2016

2016
[26]

A tutorial on energy-based learning,

Y . LeCun, S. Chopra, R. Hadsell, M. Ranzato, F. Huang,et al., “A tutorial on energy-based learning,”Predicting structured data, vol. 1, no. 0, 2006

2006
[27]

Energy-based models for deep probabilistic regression,

F. K. Gustafsson, M. Danelljan, G. Bhat, and T. B. Sch ¨on, “Energy-based models for deep probabilistic regression,” inEuropean Conference on Computer Vision (ECCV). Springer, 2020, pp. 325–343

2020
[28]

How to train your energy-based models,

Y . Song and D. P. Kingma, “How to train your energy-based models,” arXiv preprint arXiv:2101.03288, 2021

2021 arXiv
[29]

Fast and accurate deep network learning by exponential linear units (ELUs),

S. H. Djork-Arn ´e Clevert, Thomas Unterthiner, “Fast and accurate deep network learning by exponential linear units (ELUs),” inInternational Conference on Learning Representations (ICLR), 2016

2016
[30]

U-Net: Convolutional net- works for biomedical image segmentation,

O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional net- works for biomedical image segmentation,” inMedical Image Comput- ing and Computer-Assisted Intervention (MICCAI). Springer, 2015, pp. 234–241

2015
[31]

Sliced and radon wasserstein barycenters of measures,

N. Bonneel, J. Rabin, G. Peyr ´e, and H. Pfister, “Sliced and radon wasserstein barycenters of measures,”Journal of Mathematical Imaging and Vision, vol. 51, pp. 22–45, 2015

2015
[32]

Regulated pure pursuit for robot path tracking,

S. Macenski, S. Singh, F. Mart ´ın, and J. Gin ´es, “Regulated pure pursuit for robot path tracking,”Autonomous Robots, vol. 47, p. 685–694, 2023

2023
[33]

Timed-elastic-bands for time-optimal point-to-point nonlinear model predictive control,

C. R ¨osmann, F. Hoffmann, and T. Bertram, “Timed-elastic-bands for time-optimal point-to-point nonlinear model predictive control,” inECC, 2015, pp. 3352–3357

2015
[34]

Predictive collision avoidance for the dynamic window approach,

M. Missura and M. Bennewitz, “Predictive collision avoidance for the dynamic window approach,” inICRA, 2019, pp. 8620–8626

2019

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [11]

Prescient collision- free navigation of mobile robots with iterative multimodal motion prediction of dynamic obstacles,

Z. Zhang, H. Hajieghrary, E. Dean, and K. ˚Akesson, “Prescient collision- free navigation of mobile robots with iterative multimodal motion prediction of dynamic obstacles,”Robotics and Automation Letters, vol. 8, no. 9, pp. 5488–5495, 2023

work page 2023

[2] [1]

Automated guided vehicle systems, state-of-the-art control algorithms and techniques,

M. De Ryck, M. Versteyhe, and F. Debrouwere, “Automated guided vehicle systems, state-of-the-art control algorithms and techniques,” Journal of Manufacturing Systems, vol. 54, pp. 152–173, 2020

work page 2020

[3] [2]

Unfreezing the robot: Navigation in dense, interacting crowds,

P. Trautman and A. Krause, “Unfreezing the robot: Navigation in dense, interacting crowds,” inIROS, 2010, pp. 797–803

work page 2010

[4] [3]

Fro- zone: Freezing-free, pedestrian-friendly navigation in human crowds,

A. J. Sathyamoorthy, U. Patel, T. Guan, and D. Manocha, “Fro- zone: Freezing-free, pedestrian-friendly navigation in human crowds,” Robotics and Automation Letters, vol. 5, no. 3, pp. 4352–4359, 2020

work page 2020

[5] [4]

Intention aware robot crowd navigation with attention-based interaction graph,

S. Liu, P. Chang, Z. Huang, N. Chakraborty, K. Hong, W. Liang, D. L. McPherson, J. Geng, and K. Driggs-Campbell, “Intention aware robot crowd navigation with attention-based interaction graph,” inICRA, 2023, pp. 12 015–12 021

work page 2023

[6] [5]

Core challenges of social robot navigation: A survey,

C. Mavrogiannis, F. Baldini, A. Wang, D. Zhao, P. Trautman, A. Stein- feld, and J. Oh, “Core challenges of social robot navigation: A survey,” ACM Transactions on Human-Robot Interaction, vol. 12, no. 3, 2023

work page 2023

[7] [6]

Dynamic channel: A planning framework for crowd navigation,

C. Cao, P. Trautman, and S. Iba, “Dynamic channel: A planning framework for crowd navigation,” inICRA, 2019, pp. 5551–5557

2019

[8] [7]

Avoiding dense and dynamic obstacles in enclosed spaces: Application to moving in crowds,

L. Huber, J.-J. Slotine, and A. Billard, “Avoiding dense and dynamic obstacles in enclosed spaces: Application to moving in crowds,”IEEE Transactions on Robotics, vol. 38, no. 5, pp. 3113–3132, 2022

2022