Pith. sign in

REVIEW 5 major objections 6 minor 44 references

Free Energy-Inspired Cognitive Risk Integration for AV Navigation in Pedestrian-Rich Environments

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper sets out to establish that a single cognitive-risk signal, derived from the Free Energy Principle, should govern both pedestrian reactions and an autonomous vehicle's policy in shared spaces.

desk verdict A well-structured integration of risk-weighted social forces and graph-based SAC, but the cognitive-uncertainty mechanism depends on an unspecified predictor, so the reported gains are currently uninterpretable. read the letter →

arxiv 2507.20850 v1 pith:2LKUZWUW submitted 2025-07-28 cs.RO cs.AI

Meiting Dang , Yanping Wu , Yafei Wang , Dezong Zhao , David Flynn , Chongfeng Wei This is my paper · ORCID
classification cs.ROcs.AI
keywords AutonomousvehiclesPedestrianbehaviormodelingFreeenergyprincipleCognitiveuncertaintySocialforcemodelGraphneuralnetworksSoftactor-criticSharedspacenavigation
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 sets out to establish that a single cognitive-risk signal, derived from the Free Energy Principle, can make both simulated pedestrians and an autonomous vehicle's policy behave more realistically and safely in shared spaces. It proposes the Cognitive-Risk Social Force Model (CR-SFM) for pedestrians and a graph-enhanced Soft Actor-Critic agent (G-SAC-Cog) for the vehicle, with both sides using the same fused measure of physical risk and cognitive uncertainty. In simulations built on the Hamburg Bergedorf Station dataset, the full framework reaches its destination in 0.94 of episodes, collides in only 0.06, averages 4.470 m/s, and keeps jerk at 0.524 m/s³, beating the UAW-PCG baseline and both ablations. If the claim is right, cognitive uncertainty is a usable shared currency for modeling bidirectional human-vehicle interaction, not just an extra reward term.

What carries the argument

The load-bearing object is the fused risk measure: the physical risk ψ computed from a virtual distance that inflates or shrinks the geometric distance based on an agent's velocity, acceleration, and approach direction, multiplied by the cognitive term $(1 + \lambda u)$, where $u$ is the KL-divergence uncertainty from the predict–observe–update loop. This product appears twice: as the force-weight coefficients $w^{\text{goal}}$, $w^{\text{veh}}$, $w^{\text{ped}}$ in CR-SFM (Eqs. 16–18), and as the adjacency entry $A_t(i,j) = \psi_t(i,j)(1 + \lambda u_t(i,j))$ in the GCN (Eq. 25). The argument is that routing this one scalar signal through both the pedestrian model and the AV policy is what produces adaptive, bidirectional interaction rather than fixed-rule behavior.

What would settle it

Replace the undefined internal prediction model in Algorithm 1 with a trivial constant-velocity predictor and rerun the full simulation battery; if G-SAC-Cog and CR-SFM keep their margins over the ablations, the reported safety gain does not depend on the free-energy-inspired mechanism, whereas if the margins collapse, the gain is tied to that specific predictor.

Watch

Extended reading notes

Core claim

The central claim is that cognitive uncertainty — quantified as the KL divergence between an agent's predicted and observed velocity distributions — can be fused with physical risk into a single weight that makes both pedestrian trajectories and AV policies safer and more human-like. For pedestrians, the Cognitive-Risk Social Force Model multiplies goal, vehicle, and pedestrian repulsion forces by risk-based weights, so a nearby but receding agent produces little repulsion while an uncertain approaching one pulls the pedestrian off course. For the AV, the same fused risk defines the dynamic adjacency matrix of a GCN embedded in SAC's actor and critic, so the policy focuses its graph aggregation on high-risk agents. The quantitative claim is that G-SAC-Cog reaches a 0.94 success rate and 0.06 collision rate in the simulated shared-space scenarios, with the CR-SFM pedestrian model recording zero collisions and lower displacement errors than the SFM and RA-SFM baselines.

Load-bearing premise

The cognitive uncertainty $u_t$ is computed from a predicted velocity provided by an 'internal prediction model' that is never defined in the paper; if this predictor is arbitrary or unrealistic, the KL-divergence signal and every risk weight built on it lose their grounding.

Editorial extensions

If this is right

  • The CR-SFM pedestrian model produces zero collisions and lower ADE and FDE than SFM and RA-SFM, showing that the fused-risk weighting makes simulated pedestrians both more accurate and safer.
  • G-SAC-Cog outperforms UAW-PCG and both ablations on success rate, collision rate, and average speed, indicating that a risk-encoded adjacency matrix helps the policy perceive which interactions matter.
  • The jerk and maximum-acceleration results show that acceleration-based control with risk-weighted graph features yields smoother motion than velocity-based baselines, so safety does not come at the cost of comfort.
  • If the framework is correct, the same cognitive-risk signal can serve as a common interface between pedestrian simulation and AV decision-making, removing the need for fixed-behavior pedestrian models in training.

Reading between the lines

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

  • The paper's dependence on an unspecified 'internal prediction model' in Algorithm 1 means the reported gains could partly reflect that choice; a controlled swap of predictor (constant-velocity versus learned) would reveal how much of the effect is cognitive uncertainty rather than prediction quality.
  • The same adjacency-matrix design could be extended to variable pedestrian counts by letting nodes enter and leave the graph online, a setting the paper lists as future work but does not test.
  • Because the risk signal is shared between pedestrian and vehicle, the framework offers a direct way to inject the same uncertainty model into real-world traffic simulators or digital twins for scenario-based testing.
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

5 major / 6 minor

Summary. The paper proposes a framework for AV navigation in pedestrian-rich shared spaces, built around a Free-Energy-Principle-style measure of cognitive uncertainty. The uncertainty is defined in Eq. (5) as the KL divergence between a predicted and an observed Gaussian velocity distribution, and it is used in two places: the pedestrian Cognitive-Risk Social Force Model (CR-SFM), where it modulates goal and repulsive force weights (Eqs. 12-18), and the AV decision-making model G-SAC-Cog, where it weights a dynamic adjacency matrix for a GCN integrated into Soft Actor-Critic (Eq. 25). Experiments are carried out on scenarios extracted from the HBS dataset. Table I reports ADE/FDE/collision rate for CR-SFM against CV, SFM, and RA-SFM, and Table II reports success rate, collision rate, speed, and jerk for G-SAC-Cog against UAW-PCG, S-SAC, G-SAC-NoCog, and a human driver. The central claim is that the cognitive-uncertainty signal improves both pedestrian trajectory realism and AV safety/efficiency/smoothness.

Significance. If the mechanism were fully specified, the paper would address a real gap: interactive, responsive pedestrian simulation for AV decision-making in shared spaces, with a risk-aware graph representation inside a state-of-the-art RL algorithm. The combination of a social-force pedestrian model with a GCN-SAC controller is sensible, and the qualitative examples in Figs. 4-5 are informative. However, the current manuscript cannot support its central attribution claim, because the core cognitive-uncertainty signal is produced by an internal prediction model that is never defined, and because the experimental reporting relies on single-point metrics without variance or a clear train/test separation. The claimed gains therefore cannot be reproduced or traced to the proposed FEP-inspired mechanism as written.

major comments (5)
  1. [Section IV-A, Algorithm 1, Eq. (5)] The internal prediction model is listed as an input to Algorithm 1 but is never specified. The predicted velocity \hat{v}_i^t determines the prior p_t(v), and therefore the cognitive uncertainty u_t in Eq. (5). This u_t enters every risk weight in CR-SFM (Eqs. 16-18) and the GCN adjacency matrix in G-SAC-Cog (Eq. 25). Different plausible choices, such as constant-velocity extrapolation, a learned social-force predictor, or an oracle using the observed velocity, will in general produce different force weights and graph edges, and hence different numbers in Tables I and II. The paper must specify the internal prediction model in full, including how it is initialized, updated, and combined with the posterior from the previous step, and should report sensitivity to this choice.
  2. [Section V-B, Table I] The pedestrian model parameters are calibrated using Bayesian optimization on trajectories extracted from the training dataset, but the paper does not state whether the ADE/FDE/CR values in Table I are computed on a disjoint test set. If the reported values are on the calibration set, they partly reflect fitting quality rather than predictive fidelity, undermining the comparison between CR-SFM and the baselines. The authors should clearly specify the train/test split, the number of scenarios used for evaluation, and whether the reported metrics are averaged over that held-out set.
  3. [Section VI-B, Table II] All AV performance numbers are single-point estimates with no standard deviations, confidence intervals, number of evaluation episodes, or multiple seeds. Given that the difference between G-SAC-Cog (success 0.94, collision 0.06) and UAW-PCG (success 0.83, collision 0.17) is based on what appears to be a limited number of HBS interaction scenarios, the reported superiority is not statistically supported. The authors should report per-episode results across several random seeds and provide significance measures or confidence intervals.
  4. [Section IV-A, Eq. (5)] The KL divergence D_KL(N(\mu_p,\sigma_p^2)||N(\mu_o,\sigma_o^2)) has a closed form in which, for fixed variances, u_t is a monotone function of the squared prediction error (\mu_p - \mu_o)^2. It is therefore unclear whether the proposed quantity captures anything beyond the predictor's raw error, and the FEP-inspired Bayesian update in Eqs. (6)-(7) does not by itself add information. The authors should state whether this is intended, and should justify why this particular proxy constitutes 'cognitive uncertainty' rather than simply prediction error, especially since the qualitative claim in Section VI-A that CR-SFM 'detects latent threats' relies entirely on this quantity.
  5. [Section V-A and Section IV] The hyperparameters \gamma_1, \gamma_2, \lambda_1, \lambda_2, \lambda_3, \lambda, \sigma_o, and \sigma_p are not reported anywhere in the manuscript. All risk weights and adjacency values depend on these parameters, and without them the experiments are not reproducible. The authors should provide the full parameter set used, including the ranges and fitted values from the Bayesian optimization procedure for the pedestrian models.
minor comments (6)
  1. [Section VI-A.2] The text says 'FED' where it should say 'FDE' in the sentence describing SFM results; this typo should be corrected.
  2. [Section III] There is a typo in 'statest' that should read 'state'.
  3. [Section IV-A, Algorithm 1] Algorithm 1 says 'Use internal prediction model to predict velocity', but the phrase 'internal prediction model' is not defined in the algorithm or in the surrounding text; this is a clarity issue even beyond the substantive concern in Major Comment 1.
  4. [Section V-B and Table II] The row labeled 'Human driver' reports '1' for success and collision rates with no indication of how many human-driving episodes were recorded; this should be stated for the comparison to be interpretable.
  5. [References] Reference [38] duplicates reference [15] (the same Helbing and Molnar social force paper); the duplicate should be removed or converted into a proper repeated citation.
  6. [Section IV-B, Eq. (13)] The definition of \phi in Eq. (13) is stated in prose as 'the angle between the agent's velocity vector and the vector pointing from the agent to pedestrian i', but it would help to draw the geometric configuration or provide an explicit vector formula, since the sign of cos\phi affects the risk computation.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the core claims are evaluated against external benchmarks and a train/test split; self-citations are minor and not load-bearing.

full rationale

The derivation chain is not circular. The cognitive uncertainty u_t = D_KL(p_t(v)||o_t(v)) is a modeling assumption, not a fit to the reported outcomes; the AV policy and pedestrian model are evaluated in a closed-loop simulator, which is a standard experimental design rather than a logical reduction. The pedestrian CR-SFM is calibrated via Bayesian optimization on the HBS training trajectories and then evaluated on held-out test episodes (Eq. 26), so the ADE/FDE numbers are not fits to the evaluation data. The AV model is compared against the external UAW-PCG baseline [10] and human driving data, providing an external benchmark. The paper cites prior work by the same authors for the virtual-distance risk formula (Eq. 13, ref [40]) and the evaluation protocol (ref [45]), but these are used for established components, not to justify the central claim that cognitive uncertainty improves navigation. The most serious issue is an omission, not circularity: Algorithm 1 (Section IV-A) relies on an 'internal prediction model' that is never defined, making the cognitive uncertainty signal underdetermined and the results non-reproducible. This is a significant correctness limitation, but it does not make the reported improvements equivalent to their inputs by construction. Overall circularity score: 1.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The framework rests on several hand-designed risk formulas and an unspecified internal prediction model. The KL-based uncertainty proxy, the virtual distance risk, and the exponential goal-weight decay are ad hoc constructions whose parameters are fit to the training data. The standard SFM and SAC/GCN machinery are assumed from prior literature.

free parameters (8)
  • gamma1 = not reported (Bayesian optimized or hand-set)
    Scales the velocity/acceleration term in the virtual distance formula (Eq. 13), controlling how strongly kinematic features inflate physical distance.
  • gamma2 = not reported
    Scales the virtual distance in the physical risk sigmoid (Eq. 15), setting how steeply risk decays with distance.
  • lambda1 = not reported
    Weights cognitive uncertainty into the vehicle-repulsion force weight (Eq. 16).
  • lambda2 = not reported
    Weights cognitive uncertainty into the pedestrian-repulsion force weight (Eq. 17).
  • lambda3 = not reported
    Controls how quickly goal force decays with perceived interaction risk (Eq. 18).
  • lambda (adjacency) = not reported
    Controls the influence of cognitive uncertainty on GCN adjacency edge weights (Eq. 25).
  • sigma_o (observation variance) = not reported
    Fixed variance of the observed velocity distribution in Eq. (4), required for the KL divergence and Bayesian update.
  • initial sigma_p (prior variance) = not reported
    Initial variance of the Gaussian prior in Eq. (3); not specified, yet needed to seed the recursive belief update.
assumptions (6)
  • ad hoc to paper KL divergence between predicted and observed velocity distributions is a valid proxy for cognitive uncertainty.
    Adopted in Section IV-A without empirical justification; it is the central modeling assumption connecting FEP to risk weights.
  • ad hoc to paper Each agent has a functioning internal prediction model that produces velocity predictions for surrounding agents.
    Algorithm 1 assumes such a model as input, but it is never defined or evaluated.
  • standard math Gaussian prior-likelihood conjugacy permits closed-form Bayesian update (Eqs. 6-7).
    Standard result used without proof.
  • domain assumption The Social Force Model (Eq. 11) adequately describes pedestrian dynamics before cognitive weighting.
    Taken from prior literature (Helbing and Molnar); not derived or validated here.
  • ad hoc to paper Physical risk from virtual distance (Eqs. 13-15) captures perceived threat.
    Hand-designed tanh/sigmoid combination; no behavioral validation.
  • domain assumption The MDP with SAC and GCN (Eqs. 19-24) is a suitable model for AV decision-making.
    Standard deep RL framework; assumed to hold for this problem.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Free Energy-Inspired Cognitive Risk Integration for AV Navigation in Pedestrian-Rich Environments." pith.science (2026). https://pith.science/paper/2LKUZWUW

@misc{pith2026250720850,
  author       = {Pith},
  title        = {Pith review of: Free Energy-Inspired Cognitive Risk Integration for AV Navigation in Pedestrian-Rich Environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2LKUZWUW}},
  note         = {Machine review of arXiv:2507.20850}
}
read the original abstract

Recent advances in autonomous vehicle (AV) behavior planning have shown impressive social interaction capabilities when interacting with other road users. However, achieving human-like prediction and decision-making in interactions with vulnerable road users remains a key challenge in complex multi-agent interactive environments. Existing research focuses primarily on crowd navigation for small mobile robots, which cannot be directly applied to AVs due to inherent differences in their decision-making strategies and dynamic boundaries. Moreover, pedestrians in these multi-agent simulations follow fixed behavior patterns that cannot dynamically respond to AV actions. To overcome these limitations, this paper proposes a novel framework for modeling interactions between the AV and multiple pedestrians. In this framework, a cognitive process modeling approach inspired by the Free Energy Principle is integrated into both the AV and pedestrian models to simulate more realistic interaction dynamics. Specifically, the proposed pedestrian Cognitive-Risk Social Force Model adjusts goal-directed and repulsive forces using a fused measure of cognitive uncertainty and physical risk to produce human-like trajectories. Meanwhile, the AV leverages this fused risk to construct a dynamic, risk-aware adjacency matrix for a Graph Convolutional Network within a Soft Actor-Critic architecture, allowing it to make more reasonable and informed decisions. Simulation results indicate that our proposed framework effectively improves safety, efficiency, and smoothness of AV navigation compared to the state-of-the-art method.

Figures

Figures reproduced from arXiv: 2507.20850 by the authors.

Figure 1
Figure 1. The proposed framework of AV-pedestrian interaction in the urban shared space. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Comparison of the SFM and CR-SFM in pedestrian-vehicle interac [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Qualitative examples of pedestrian trajectory simulation under different [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparison of the interaction process and dynamic state evolution between the proposed model and real driving data for Case 1. (a) Interaction [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Comparison of the interaction process and dynamic state evolution between the proposed model and real driving data for Case 2. (a) Interaction [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 43 canonical work pages

  1. [1]

    Impacts of connected and autonomous vehicles on urban transportation and environment: A comprehensive review,

    M. M. Rahman and J.-C. Thill, “Impacts of connected and autonomous vehicles on urban transportation and environment: A comprehensive review,” Sustainable Cities and Society , vol. 96, p. 104649, 2023

    work page 2023

  2. [2]

    Advanced learning technologies for intelligent transportation systems: Prospects and challenges,

    R. A. Khalil, Z. Safelnasr, N. Yemane, M. Kedir, A. Shafiqurrahman, and N. Saeed, “Advanced learning technologies for intelligent transportation systems: Prospects and challenges,” IEEE Open Journal of Vehicular Technology, 2024

    work page 2024

  3. [3]

    Interactions and behaviors of pedestrians with autonomous vehicles: A synthesis,

    S. Rezwana and N. Lownes, “Interactions and behaviors of pedestrians with autonomous vehicles: A synthesis,” Future Transportation, vol. 4, no. 3, pp. 722–745, 2024

    work page 2024

  4. [4]

    Does another pedestrian matter? a virtual reality study on the interaction between multiple pedes- trians and autonomous vehicles in shared space,

    Y . Feng, Z. Xu, H. Farah, and B. Van Arem, “Does another pedestrian matter? a virtual reality study on the interaction between multiple pedes- trians and autonomous vehicles in shared space,” IEEE Transactions on Intelligent Transportation Systems, 2024

    work page 2024

  5. [5]

    Deconstructing pedestrian crossing decisions in interactions with continuous traffic: An anthropomorphic model,

    K. Tian, G. Markkula, C. Wei, Y . M. Lee, R. Madigan, T. Hirose, N. Merat, and R. Romano, “Deconstructing pedestrian crossing decisions in interactions with continuous traffic: An anthropomorphic model,” IEEE Transactions on Intelligent Transportation Systems, vol. 25, no. 3, pp. 2466–2478, 2023

    work page 2023

  6. [6]

    Integrating predictive motion uncertainties with distributionally robust risk-aware control for safe robot navigation in crowds,

    K. Ryu and N. Mehr, “Integrating predictive motion uncertainties with distributionally robust risk-aware control for safe robot navigation in crowds,” in 2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2024, pp. 2410–2417

    work page 2024

  7. [7]

    Sicnav: Safe and interactive crowd navigation using model predictive control and bilevel optimization,

    S. Samavi, J. R. Han, F. Shkurti, and A. P. Schoellig, “Sicnav: Safe and interactive crowd navigation using model predictive control and bilevel optimization,” IEEE Transactions on Robotics , 2024

    work page 2024

  8. [8]

    Security considerations in ai-robotics: A survey of current methods,

    S. Neupane, S. Mitra, I. Fernandez, S. Saha, S. Mittal, J. Chen, N. Pillai, and S. Rahimi, “Security considerations in ai-robotics: A survey of current methods,” Challenges, and Opportunities , 2023

    work page 2023

Show all 44 references
  1. [9]

    What the constant velocity model can teach us about pedestrian motion prediction,

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

    2020

  2. [10]

    Uncertainty-aware drl for autonomous vehicle crowd navigation in shared space,

    M. Golchoubian, M. Ghafurian, K. Dautenhahn, and N. L. Azad, “Uncertainty-aware drl for autonomous vehicle crowd navigation in shared space,” IEEE Transactions on Intelligent Vehicles , 2024

    2024

  3. [11]

    Towards proactive naviga- tion: A pedestrian-vehicle cooperation based behavioral model,

    M. Kabtoul, A. Spalanzani, and P. Martinet, “Towards proactive naviga- tion: A pedestrian-vehicle cooperation based behavioral model,” in 2020 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2020, pp. 6958–6964

    2020

  4. [12]

    Game theory- based simultaneous prediction and planning for autonomous vehicle navigation in crowded environments,

    K. Li, Y . Chen, M. Shan, J. Li, S. Worrall, and E. Nebot, “Game theory- based simultaneous prediction and planning for autonomous vehicle navigation in crowded environments,” in 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC) . IEEE, 2023,...

    2023

  5. [13]

    Reciprocal n- body collision avoidance,

    J. Van Den Berg, S. J. Guy, M. Lin, and D. Manocha, “Reciprocal n- body collision avoidance,” in Robotics Research: The 14th International Symposium ISRR. Springer, 2011, pp. 3–19

    2011

  6. [14]

    Crowd-robot interaction: Crowd-aware robot navigation with attention-based deep reinforcement 13 learning,

    C. Chen, Y . Liu, S. Kreiss, and A. Alahi, “Crowd-robot interaction: Crowd-aware robot navigation with attention-based deep reinforcement 13 learning,” in 2019 international conference on robotics and automation (ICRA). IEEE, 2019, pp. 6015–6022

    2019

  7. [16]

    Simulation of pedestrian interaction with autonomous vehicles via social force model,

    M. M. Rashid, M. Seyedi, and S. Jung, “Simulation of pedestrian interaction with autonomous vehicles via social force model,”Simulation modelling practice and theory , vol. 132, p. 102901, 2024

    2024

  8. [17]

    How to evaluate the navigation of autonomous vehicles around pedestrians?

    M. Kabtoul, M. Pr ´edhumeau, A. Spalanzani, J. Dugdale, and P. Mar- tinet, “How to evaluate the navigation of autonomous vehicles around pedestrians?” IEEE Transactions on Intelligent Transportation Systems , vol. 25, no. 3, pp. 2311–2321, 2023

    2023

  9. [18]

    Human- robot co-navigation using anticipatory indicators of human walking motion,

    V . V . Unhelkar, C. P´erez-D’Arpino, L. Stirling, and J. A. Shah, “Human- robot co-navigation using anticipatory indicators of human walking motion,” in 2015 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2015, pp. 6183–6190

    2015

  10. [19]

    Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation,

    P. Trautman, J. Ma, R. M. Murray, and A. Krause, “Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation,” The International Journal of Robotics Research, vol. 34, no. 3, pp. 335–356, 2015

    2015

  11. [20]

    Moving in a crowd: Safe and efficient navigation among heterogeneous agents

    J. Godoy, I. Karamouzas, S. J. Guy, and M. L. Gini, “Moving in a crowd: Safe and efficient navigation among heterogeneous agents.” in IJCAI, 2016, pp. 294–300

    2016

  12. [21]

    Move beyond trajectories: Distribution space coupling for crowd navigation,

    M. Sun, F. Baldini, P. Trautman, and T. Murphey, “Move beyond trajectories: Distribution space coupling for crowd navigation,” arXiv preprint arXiv:2106.13667, 2021

    2021 arXiv

  13. [22]

    Path-following navigation in crowds with deep reinforcement learning,

    H. Fu, Q. Wang, and H. He, “Path-following navigation in crowds with deep reinforcement learning,” IEEE Internet of Things Journal , 2024

    2024

  14. [23]

    A spatiotemporal graphical attention navigation algorithm based on limited state information,

    T. Ma, Z. Liu, T. Liu, Y . Zhao, and Y . Chai, “A spatiotemporal graphical attention navigation algorithm based on limited state information,” IEEE Transactions on Computational Social Systems , 2024

    2024

  15. [24]

    Motion planning among dynamic, decision-making agents with deep reinforcement learning,

    M. Everett, Y . F. Chen, and J. P. How, “Motion planning among dynamic, decision-making agents with deep reinforcement learning,” in 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018, pp. 3052–3059

    2018

  16. [25]

    Group- aware robot navigation in crowds using spatio-temporal graph attention network with deep reinforcement learning,

    X. Lu, A. Faragasso, Y . Wang, A. Yamashita, and H. Asama, “Group- aware robot navigation in crowds using spatio-temporal graph attention network with deep reinforcement learning,” IEEE Robotics and Automa- tion Letters, 2025

    2025

  17. [26]

    Her-drl: Heterogeneous relational deep reinforcement learning for single-robot and multi-robot crowd navigation,

    X. Zhou, S. Piao, W. Chi, L. Chen, and W. Li, “Her-drl: Heterogeneous relational deep reinforcement learning for single-robot and multi-robot crowd navigation,” IEEE Robotics and Automation Letters , 2025

    2025

  18. [27]

    Learning hierarchical graph-based policy for goal-reaching in unknown environments,

    Y . Cui, S. Ye, X. Xu, H. Sha, C. Wang, L. Lin, Z. Liu, R. Xiong, and Y . Wang, “Learning hierarchical graph-based policy for goal-reaching in unknown environments,” IEEE Robotics and Automation Letters , 2024

    2024

  19. [28]

    Multi-robot social- aware cooperative planning in pedestrian environments using attention- based actor-critic,

    L. Dong, Z. He, C. Song, X. Yuan, and H. Zhang, “Multi-robot social- aware cooperative planning in pedestrian environments using attention- based actor-critic,” Artificial Intelligence Review, vol. 57, no. 4, p. 108, 2024

    2024

  20. [29]

    Enhanced spatial attention graph for motion planning in crowded, partially observable environments,

    W. Shi, Y . Zhou, X. Zeng, S. Li, and M. Bennewitz, “Enhanced spatial attention graph for motion planning in crowded, partially observable environments,” in 2022 International Conference on Robotics and Au- tomation (ICRA). IEEE, 2022, pp. 4750–4756

    2022

  21. [30]

    Learning crowd behaviors in navigation with attention-based spatial-temporal graphs,

    Y . Zhou and J. Garcke, “Learning crowd behaviors in navigation with attention-based spatial-temporal graphs,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 5485–5491

    2024

  22. [31]

    Uncertainty, neuromodulation, and atten- tion,

    J. Y . Angela and P. Dayan, “Uncertainty, neuromodulation, and atten- tion,” Neuron, vol. 46, no. 4, pp. 681–692, 2005

    2005

  23. [32]

    Cognitive uncertainty,

    B. Enke and T. Graeber, “Cognitive uncertainty,” The Quarterly Journal of Economics, vol. 138, no. 4, pp. 2021–2067, 2023

    2021

  24. [33]

    Confidence and certainty: distinct probabilistic quantities for different goals,

    A. Pouget, J. Drugowitsch, and A. Kepecs, “Confidence and certainty: distinct probabilistic quantities for different goals,” Nature neuroscience, vol. 19, no. 3, pp. 366–374, 2016

    2016

  25. [34]

    An overview of the free energy principle and related research,

    Z. Zhang and F. Xu, “An overview of the free energy principle and related research,” Neural Computation , vol. 36, no. 5, pp. 963–1021, 2024

    2024

  26. [35]

    The free-energy principle: a unified brain theory?

    K. Friston, “The free-energy principle: a unified brain theory?” Nature reviews neuroscience, vol. 11, no. 2, pp. 127–138, 2010

    2010

  27. [36]

    Free energy risk metrics for systemically safe ai: Gatekeeping multi-agent study,

    M. Walters, R. Kaufmann, J. Sefas, and T. Kopinski, “Free energy risk metrics for systemically safe ai: Gatekeeping multi-agent study,” arXiv preprint arXiv:2502.04249, 2025

    2025 arXiv

  28. [37]

    Active vision for robot manipulators using the free energy principle,

    T. Van de Maele, T. Verbelen, O. C ¸ atal, C. De Boom, and B. Dhoedt, “Active vision for robot manipulators using the free energy principle,” Frontiers in neurorobotics, vol. 15, p. 642780, 2021

    2021

  29. [38]

    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

  30. [39]

    Interaction-aware decision-making for automated vehicles using social value orientation,

    L. Crosato, H. P. Shum, E. S. Ho, and C. Wei, “Interaction-aware decision-making for automated vehicles using social value orientation,” IEEE Transactions on Intelligent Vehicles, vol. 8, no. 2, pp. 1339–1349, 2022

    2022

  31. [40]

    Coupling intention and actions of vehicle–pedestrian interaction: A virtual reality experiment study,

    M. Dang, Y . Jin, P. Hang, L. Crosato, Y . Sun, and C. Wei, “Coupling intention and actions of vehicle–pedestrian interaction: A virtual reality experiment study,” Accident Analysis & Prevention, vol. 203, p. 107639, 2024

    2024

  32. [41]

    Soft actor-critic: Off- policy maximum entropy deep reinforcement learning with a stochastic actor,

    T. Haarnoja, A. Zhou, P. Abbeel, and S. Levine, “Soft actor-critic: Off- policy maximum entropy deep reinforcement learning with a stochastic actor,” in International conference on machine learning . Pmlr, 2018, pp. 1861–1870

    2018

  33. [42]

    Semi-supervised learning with graph learning-convolutional networks,

    B. Jiang, Z. Zhang, D. Lin, J. Tang, and B. Luo, “Semi-supervised learning with graph learning-convolutional networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 11 313–11 320

    2019

  34. [43]

    A discrete choice model for solving conflict situations between pedestrians and vehicles in shared space,

    F. Pascucci, N. Rinke, C. Schiermeyer, V . Berkhahn, and B. Friedrich, “A discrete choice model for solving conflict situations between pedestrians and vehicles in shared space,” arXiv preprint arXiv:1709.09412 , 2017

    2017 arXiv

  35. [44]

    Evaluating pedestrian tra- jectory prediction methods with respect to autonomous driving,

    N. Uhlemann, F. Fent, and M. Lienkamp, “Evaluating pedestrian tra- jectory prediction methods with respect to autonomous driving,” IEEE Transactions on Intelligent Transportation Systems , 2024

    2024

  36. [45]

    Dynamic game-theoretical decision-making framework for vehicle-pedestrian interaction with hu- man bounded rationality,

    M. Dang, D. Zhao, Y . Wang, and C. Wei, “Dynamic game-theoretical decision-making framework for vehicle-pedestrian interaction with hu- man bounded rationality,” IEEE Transactions on Intelligent Transporta- tion Systems, 2025. Meiting Dang received the B.S. and the M.S. degree...

    2002

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.