Pith. sign in

REVIEW 4 major objections 4 minor 48 references

IANN-MPPI: Interaction-Aware Neural Network-Enhanced Model Predictive Path Integral Approach for Autonomous Driving

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read By feeding each candidate control trajectory from MPPI through a neural interaction predictor, IANN-MPPI lets an autonomous vehicle anticipate that other drivers will yield, enabling efficient merging in dense traffic with 0.10 s per…

desk verdict A practical way to condition MPPI on black-box trajectory predictors, but the evidence doesn't yet prove the predictor generalizes beyond its training distribution. read the letter →

arxiv 2507.11940 v1 pith:KZ6JMUKG submitted 2025-07-16 cs.RO cs.SYeess.SY

Kanghyun Ryu , Minjun Sung , Piyush Gupta , Jovin D'sa , Faizan M. Tariq , David Isele , Sangjae Bae This is my paper · ORCID
classification cs.ROcs.SYeess.SY
keywords ModelPredictivePathIntegralcontrolinteraction-awareplanningtrajectorypredictionautonomousdrivingdensetrafficmergingspline-basedsamplingpriorBiased-MPPIStudent-SGAN
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 show that an autonomous vehicle can plan interactive maneuvers in dense traffic by coupling a sampling-based controller with a black-box neural trajectory predictor. Its central move is to condition the predictor on each control sample the controller evaluates, so that a proposed merge is scored not against static traffic but against predicted reactive yielding by neighboring drivers. In simulation this yields higher merge success than a constant-velocity baseline (67.5% vs 45% under probabilistic yielding) with zero collisions, while a spline-based sampling prior cuts merge time and planning cost. The authors also show the whole loop runs in about 0.10 seconds per planning step on a laptop GPU, making the approach compatible with real-time control. A sympathetic reader would take the paper's claim to be that interaction-aware planning need not sacrifice real-time operation or rely on hand-crafted interaction models.

What carries the argument

The load-bearing mechanism is the ego-conditioned rollout in Eq. (7): instead of calling the predictor once on history, IANN-MPPI feeds each sampled ego trajectory into the predictor step by step, so surrounding-vehicle predictions become a function of the ego's candidate action. In this paper the predictor is Student-SGAN, an RNN-GAN model distilled for speed. The second mechanism is the spline-based prior: cubic Hermite splines are drawn for the left and right lanes, tracked by PID and Stanley controllers to form reference control sequences, and MPPI samples part of its population around those references; because that makes the sampling distribution non-Gaussian, the controller switches to Biased-MPPI importance weights. Together these pieces let the planner search over distinct interaction outcomes, merge versus stay, while remaining parallelizable on a GPU.

What would settle it

Run the same planner against a surrounding-vehicle policy that never yields while keeping the SGAN predictor; if the planner still initiates merges, the predictor is not capturing non-cooperative behavior and the reported safety comes from simulator consistency rather than interaction modeling. A direct check would be to measure the mean displacement error of the Eq. (7) rollouts against the actual simulator responses across the 40 random seeds; large error in the probabilistic-yield zone would falsify the claim that conditioning works.

Watch

Extended reading notes

Core claim

The central claim is that MPPI's rollouts can double as conditioning inputs for a neural interaction predictor, turning cost evaluation into a simulation of how traffic responds to each ego maneuver. Concretely, IANN-MPPI samples control sequences, rolls out ego trajectories with a kinematic bicycle model, feeds each trajectory as the ego's future into Student-SGAN, and rolls the predictor forward according to Eq. (7) to obtain surrounding-vehicle predictions; the cost uses an elliptical Gaussian risk overlap with those predictions. The paper argues this lets the planner nudge into traffic and merge when no gap yet exists, because the predictor foresees yielding, while the same mechanism lets it abort a merge when the predictor judges that another driver will not yield. Reported results are 67.5% success with zero collisions against probabilistic IDM traffic versus 45% for constant-velocity prediction, and 32.5% success with zero collisions against uncooperative IDM, where an always-yield prediction model causes 32.5% collisions. Computation averages 0.10 ± 0.02 s per planning step with 1500 samples.

Load-bearing premise

The whole benefit rests on the learned predictor being truthful about how actual surrounding vehicles would react to each candidate ego trajectory, and the paper does not report prediction accuracy on held-out scenarios or state what data the predictor was trained on; if it was trained on the same simulator used for evaluation, the results are in-distribution and may not transfer.

Editorial extensions

If this is right

  • A planner using interaction-aware neural predictions can initiate a merge before a physical gap exists, because it anticipates the target-lane driver yielding (67.5% success with SGAN vs 45% with constant-velocity prediction in probabilistic IDM traffic).
  • When other drivers do not yield, the same mechanism can abort a merge and return to the lane, avoiding the collisions that an always-yield assumption produces (0% vs 32.5% collisions in uncooperative IDM traffic).
  • Shaping the MPPI sampling distribution with lane splines improves sample diversity: merge time drops from about 31.8 s to 21.4 s and planning cost from 9.49 to 5.90.
  • GPU-parallelized MPPI with a distilled predictor runs in 0.10 ± 0.02 s per planning step, suggesting real-time deployment is feasible.

Reading between the lines

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

  • The same conditioning trick could be applied to any sequence-based predictor, though a joint predictor that directly conditions on the ego's future trajectory would remove the iterative rollout and save computation, an alternative the paper itself hints at.
  • If the predictor inherits a cooperative bias from training data, the planner may over-trust yielding; curating datasets that include non-yielding and near-collision interactions is likely necessary before deployment, a point the paper raises in its limitations.
  • The spline-prior recipe generalizes beyond lane splines: any structured environment, such as intersections or roundabouts, could supply maneuver families as sampling priors for MPPI.
  • A testable next step is to vary the fraction of samples drawn around spline priors versus free Gaussian samples and map the success-versus-collision frontier, since the paper reports only one mixture.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes IANN-MPPI, a sampling-based MPC framework that conditions a neural trajectory predictor (Student-SGAN) on each MPPI control candidate via iterative rollout, then scores candidates with a cost function that includes an ellipsoidal overlap risk term. A cubic-Hermite spline prior is introduced to bias sampling toward lane-changing maneuvers, using Biased-MPPI importance weights. The method is evaluated in a simulated dense-highway merging scenario against constant-velocity (CV) and IDM prediction baselines under three IDM behavior models (probabilistic, uncooperative, cooperative), with 40 Monte Carlo runs per cell. The paper reports improved success rates over CV, zero collisions for the SGAN predictor in all reported runs, reduced merge times with the spline prior, and per-step computation time of 0.10 ± 0.02 s with Student-SGAN.

Significance. If the empirical claims hold, the core idea is a useful and practical way to couple black-box learned predictors with MPPI: it avoids gradient-based optimization, is predictor-agnostic, and leverages MPPI parallelism. The spline prior is a simple mechanism for encoding lane structure in the sampling distribution. The paper is also explicit about limitations (no prediction uncertainty, potential cooperative bias, computational overhead), which is commendable. However, the central evidence as presented is not yet sufficient: the training/evaluation overlap is undisclosed, statistical significance is not established, and the safety/efficiency trade-off against the IDM predictor is mixed.

major comments (4)
  1. [§V-A, Eq. (7), Algorithm 1] The central claim that ego-conditioned NN predictions enable interactive merging rests on Student-SGAN's ability to predict the deployment interaction model, but the paper never reports prediction accuracy on held-out scenarios and never describes the dataset used to train Student-SGAN, instead referring readers to [42]. Since every closed-loop evaluation uses one of three IDM variants as ground truth, it is unknown whether the observed yielding predictions reflect learned interactive behavior or an in-distribution fit to the same simulator. The conclusion's own admission that the predictor 'may be biased toward cooperation' makes this gap concrete. Please report the training-data distribution, evaluate prediction error against held-out interaction data (e.g., displacement error and occupancy metrics for the ego-conditioned rollouts of Eq. (7)), and add at least one out-of-distribution experiment, for example with IDM parameters outside the training range or a different interaction model.
  2. [§V-B, Table II] All comparative claims—such as SGAN 67.5% vs. CV 45.0% success in Probabilistic IDM, or SGAN 32.5% vs. CV 10.0% in Uncooperative IDM—are based on 40 runs per cell with no confidence intervals, standard errors, or significance tests. With 40 Bernoulli draws, the standard error of a success proportion is about 0.08, so it is impossible to tell which observed gaps are meaningful without additional statistics. Please report binomial confidence intervals for success and collision rates, and use paired or appropriate tests for planning cost and merge time. Table III has the same issue and should include the same treatment.
  3. [§V-B, Table II and Fig. 6] The results are mixed with respect to the claim that NN prediction improves efficiency. In Probabilistic IDM, the IDM prediction model achieves higher success (87.5%) and lower planning cost (5.03 ± 2.08) than SGAN (67.5%, 6.09 ± 2.35); in Cooperative IDM all predictors succeed 100% of the time. SGAN's advantage appears mainly in Uncooperative IDM, where it has zero collisions versus 32.5% for IDM prediction, but its success rate (32.5%) is also lower than IDM's (40.0%). The paper should either present a single metric that trades off safety and efficiency, or explicitly characterize the operating regime where the NN predictor is preferable, rather than stating that SGAN 'captures nuanced interactions' as a blanket conclusion.
  4. [§IV-C, Eq. (4)] The safety-risk term in the cost function is summed only from k = t+1 to t+H_pred, with H_pred = 8 (2.4 s), while the planning horizon is H = 17 (5.1 s). This means the planner does not penalize conflicts in the later portion of the planned trajectory. The paper does not justify this truncation; please either extend the risk-evaluation horizon to the full planning horizon, or explain why it is safe to ignore risk beyond H_pred, especially because the reported merge times (8–35 s) are much longer than the prediction horizon.
minor comments (4)
  1. [§III-A] The rollout notation 'xt:t+H = (xt, xt+t+1, . . . , xH)' contains a typo: 'xt+t+1' should be 'xt+1', and 'xH' should be 'xt+H'.
  2. [§III-C, Eq. (4)] The first two cost terms both use λ_goal with indicator functions on x and y separately; it would be clearer to define the local goal as a point and use a single distance term, and to clarify the role of the threshold ε.
  3. [§V-C, Table III] The spline-prior ablation does not specify which prediction model and behavior model are used, the number of runs, or the collision rates; please make the experimental setup explicit.
  4. [§V-D] The sentence 'Across 20 runs, the average and standard deviation of computation time per planning step was 0.10 ± 0.02 seconds' is ambiguous: please state whether these are 20 episodes, 20 planning steps, or 20 independent trials.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the planning algorithm's derivation is self-contained, and the central claim is empirical rather than reduction-by-construction.

full rationale

Walking the derivation chain, IANN-MPPI couples a learned predictor phi with MPPI by rolling out Eq. (7) ego-conditioned predictions and feeding them into the cost in Eq. (4). No step in the paper defines an output quantity in terms of itself, fits a parameter to the evaluation data and then reports it as a prediction, or imports a uniqueness or ansatz result solely through same-author citation. The predictor Student-SGAN is cited to [42] for training details, but that citation is an implementation reference rather than a load-bearing mathematical premise, and the paper's closed-loop experiments provide independent, if limited, evidence for the central claim. The main concern, correctly flagged by the reader, is the undisclosed relationship between Student-SGAN's training distribution and the IDM simulation models used for evaluation; if the predictor were trained on the same IDM behavior, the results could be in-distribution. However, the paper text does not exhibit that reduction, and an unvalidated transfer assumption is a correctness or evidence limitation, not a circularity of the kind defined by this analysis. Therefore, no specific circular step can be quoted and exhibited, and the appropriate finding is no significant circularity.

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

The framework does not derive its success from a single free parameter, but it depends on a learned predictor and several hand-chosen hyperparameters whose values and training data are largely undisclosed. No new physical or conceptual entities are introduced; the spline prior is a sampling distribution and SGAN is an existing prediction model.

free parameters (5)
  • Cost function weights {lambda_goal, lambda_l, lambda_v, lambda_delta, lambda_a, lambda_j, lambda_s, lambda_b… = Not reported
    Positive weights in Eq. (4); the balance between goal, lane, velocity, control, comfort, boundary, and risk terms strongly shapes merging aggressiveness and success, but values are never listed.
  • Sampling variances Sigma and Sigma_spline = Sigma = diag(0.1, 1e-3), Sigma_spline = diag(0.1, 5e-4) (Table I)
    MPPI exploration noise; hand-chosen with no sensitivity analysis, and they directly control whether the planner can discover yielding interactions.
  • Spline-prior sample count M per lane = Not reported
    Algorithm 1 uses M samples per spline lane and K-2M remaining samples; the ratio controls how strongly the prior dominates the search and is not specified.
  • MPPI inverse temperature lambda = 1 (Table I)
    Scales all costs in the importance weights; a standard but hand-picked value that affects how selectively the planner averages samples.
  • Student-SGAN predictor weights = Trained via knowledge distillation, details deferred to [42]
    The learned interaction predictor is load-bearing for the claimed behavior; its weights are fitted to undisclosed data and are not an external given.
assumptions (5)
  • domain assumption Surrounding vehicles' future trajectories depend on the ego's planned trajectory through an unknown interaction function g
    Section III-B, Eq. (3c); this is the defining assumption of interaction-aware planning and is not validated against real driving data in the paper.
  • domain assumption The rolled-out NN predictor in Eq. (7) provides accurate predictions of how surrounding vehicles react to each ego candidate trajectory
    Section IV-A; the entire cost evaluation depends on this. The paper does not report prediction accuracy on held-out scenarios or a train/test split.
  • domain assumption The kinematic bicycle model is an adequate dynamics model for both ego and surrounding vehicles over horizons of 2.4 to 5.1 seconds
    Section III-B; used for rollouts and simulation, with no comparison against a higher-fidelity vehicle model.
  • domain assumption The Gaussian-overlap risk measure rho in Eq. (5) adequately encodes collision safety, and the cost weights adequately encode the planning objective
    Section III-C; the risk term is a smooth proxy, not a collision check, and the optimization can exploit the proxy in ways that may not reflect true collision risk.
  • standard math Biased-MPPI importance weights in Eq. (8) are valid for the non-Gaussian spline mixture distribution
    Section IV-C; taken from Trevisan and Alonso-Mora [35] without reproducing the proof or assessing weight degeneracy with K = 1500 samples.

how reviews work

0 comments
Cite this review

Pith. "Pith review of IANN-MPPI: Interaction-Aware Neural Network-Enhanced Model Predictive Path Integral Approach for Autonomous Driving." pith.science (2026). https://pith.science/paper/KZ6JMUKG

@misc{pith2026250711940,
  author       = {Pith},
  title        = {Pith review of: IANN-MPPI: Interaction-Aware Neural Network-Enhanced Model Predictive Path Integral Approach for Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KZ6JMUKG}},
  note         = {Machine review of arXiv:2507.11940}
}
read the original abstract

Motion planning for autonomous vehicles (AVs) in dense traffic is challenging, often leading to overly conservative behavior and unmet planning objectives. This challenge stems from the AVs' limited ability to anticipate and respond to the interactive behavior of surrounding agents. Traditional decoupled prediction and planning pipelines rely on non-interactive predictions that overlook the fact that agents often adapt their behavior in response to the AV's actions. To address this, we propose Interaction-Aware Neural Network-Enhanced Model Predictive Path Integral (IANN-MPPI) control, which enables interactive trajectory planning by predicting how surrounding agents may react to each control sequence sampled by MPPI. To improve performance in structured lane environments, we introduce a spline-based prior for the MPPI sampling distribution, enabling efficient lane-changing behavior. We evaluate IANN-MPPI in a dense traffic merging scenario, demonstrating its ability to perform efficient merging maneuvers. Our project website is available at https://sites.google.com/berkeley.edu/iann-mppi

Figures

Figures reproduced from arXiv: 2507.11940 by the authors.

Figure 1
Figure 1. Illustrative example of a merging scenario using interaction-aware MPPI. (a) In traditional decoupled prediction–planning frameworks with static predictions, the planner cannot model how surrounding vehicles react to the ego vehicle’s actions, resulting in inefficient behavior. In contrast, interaction-aware MPPI generates predictions conditioned on sampled con￾trol sequences. For instance, (b-1) shows nominal traff… view at source ↗
Figure 2
Figure 2. Illustrative figure for local goal design. The local goal directs the AV to maintain forward progress at a reference velocity while guiding it toward a merge into the target lane. where the state vector x i t = [x i t , yi t , ψi t , vi t ] represents [x-y coordinates, heading angle, speed] and the control input u i t = [δ i t , ai t ] represents [steering angle, acceleration] of the agent i at time t. The parameter… view at source ↗
Figure 3
Figure 3. Rollout of the prediction model for generating ego-conditioned predictions. While the model captures interactions based on historical data, we obtain future-conditioned, interaction-aware predictions by performing multi-step rollouts and updating the ego vehicle’s history with the candidate trajectory. To address this, we use a ML-based trajectory predictor ϕ, trained to jointly predict the interactive future trajec… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Forced yield zone and probabilistic yield zone of surrounding vehicles. When the ego vehicle enters the forced yield zone, the surrounding vehicle must yield to the ego vehicle. In contrast, in the probabilistic yield zone, the surrounding vehicle decides to yield or n…
Figure 5
Figure 5. Figure 5: Comparison of MPPI with non-interaction-aware CV prediction (top) and our IANN-MPPI with SGAN prediction (bottom). SGAN predicts that the surrounding vehicle will slow down (middle column), enabling the ego vehicle to merge via nudging (final column). In contrast, CV f…
Figure 6
Figure 6. Figure 6: Comparison of IANN-MPPI with IDM prediction (top) and SGAN prediction (bottom) in uncooperative scenario. IDM assumes the surrounding vehicle will yield, prompting the ego vehicle to merge (middle column). When this assumption fails and the surrounding vehicle does not…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 45 canonical work pages

  1. [42]

    Towards scalable & efficient interaction-aware planning in autonomous vehicles using knowledge distillation

    Piyush Gupta, David Isele, and Sangjae Bae. Towards scalable & efficient interaction-aware planning in autonomous vehicles using knowledge distillation. In 2024 IEEE Intelligent Vehicles Symposium (IV), pages 2735–2742. IEEE, 2024

    work page 2024

  2. [1]

    Predictive Planner for Autonomous Driving with Consistency Models

    Anjian Li, Sangjae Bae, David Isele, Ryne Beeson, and Faizan M Tariq. Predictive Planner for Autonomous Driving with Consistency Models. In 2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC) . IEEE, 2025

    work page 2025

  3. [2]

    Lane- change in dense traffic with model predictive control and neural net- works

    Sangjae Bae, David Isele, Alireza Nakhaei, Peng Xu, Alexandre Mi- randa A ˜non, Chiho Choi, Kikuo Fujimura, and Scott Moura. Lane- change in dense traffic with model predictive control and neural net- works. IEEE Transactions on Control Systems Technology, 31(2):646– 659, 2022

    work page 2022

  4. [3]

    Cooperation-Aware Lane Change Maneuver in Dense Traffic based on Model Predictive Control with Recurrent Neural Network

    Sangjae Bae, Dhruv Saxena, Alireza Nakhaei, Chiho Choi, Kikuo Fu- jimura, and Scott Moura. Cooperation-Aware Lane Change Maneuver in Dense Traffic based on Model Predictive Control with Recurrent Neural Network. In American Control Conference, 2020

    work page 2020

  5. [4]

    Interaction-aware trajectory planning for autonomous vehicles with analytic integration of neural networks into model predictive control

    Piyush Gupta, David Isele, Donggun Lee, and Sangjae Bae. Interaction-aware trajectory planning for autonomous vehicles with analytic integration of neural networks into model predictive control. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pages 7794–7800. IEEE, 2023

    work page 2023

  6. [5]

    Deep reinforcement learning in a handful of trials using prob- abilistic dynamics models

    Kurtland Chua, Roberto Calandra, Rowan McAllister, and Sergey Levine. Deep reinforcement learning in a handful of trials using prob- abilistic dynamics models. Advances in neural information processing systems, 31, 2018

    work page 2018

  7. [6]

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

    Kanghyun Ryu and Negar Mehr. Integrating predictive motion uncer- tainties with distributionally robust risk-aware control for safe robot navigation in crowds. In 2024 IEEE International Conference on Robotics and Automation (ICRA) , pages 2410–2417. IEEE, 2024

    work page 2024

  8. [7]

    Rehg, Byron Boots, and Evangelos A

    Grady Williams, Nolan Wagener, Brian Goldfain, Paul Drews, James M. Rehg, Byron Boots, and Evangelos A. Theodorou. Informa- tion theoretic MPC for model-based reinforcement learning. In 2017 IEEE International Conference on Robotics and Automation (ICRA) , pages 1714–1721, 2017

    work page 2017

Show all 48 references
  1. [8]

    π-MPPI: A Projection-based Model Predictive Path Integral Scheme for Smooth Optimal Control of Fixed-Wing Aerial Vehicles

    Edvin Martin Andrejev, Amith Manoharan, Karl-Eerik Unt, and Arun Kumar Singh. π-MPPI: A Projection-based Model Predictive Path Integral Scheme for Smooth Optimal Control of Fixed-Wing Aerial Vehicles. IEEE Robotics and Automation Letters , 2025

    2025

  2. [9]

    Quadcopter Trajectory Tracking Based on Model Predictive Path Integral Control and Neural Network

    Yong Li, Qidan Zhu, and Ahsan Elahi. Quadcopter Trajectory Tracking Based on Model Predictive Path Integral Control and Neural Network. Drones (2504-446X), 9(1), 2025

    2025

  3. [10]

    Learning terrain-aware kinodynamic model for autonomous off-road rally driv- ing with model predictive path integral control

    Hojin Lee, Taekyung Kim, Jungwi Mun, and Wonsuk Lee. Learning terrain-aware kinodynamic model for autonomous off-road rally driv- ing with model predictive path integral control. IEEE Robotics and Automation Letters, 8(11):7663–7670, 2023

    2023

  4. [11]

    Active Learning with Dual Model Predictive Path-Integral Control for Interaction-Aware Autonomous Highway On-ramp Merging

    Jacob Knaup, Jovin D’sa, Behdad Chalaki, Tyler Naes, Hos- sein Nourkhiz Mahjoub, Ehsan Moradi-Pari, and Panagiotis Tsiotras. Active Learning with Dual Model Predictive Path-Integral Control for Interaction-Aware Autonomous Highway On-ramp Merging. In 2024 IEEE International Co...

    2024

  5. [12]

    Maximum-entropy multi-agent dynamic games: Forward and inverse solutions

    Negar Mehr, Mingyu Wang, Maulik Bhatt, and Mac Schwager. Maximum-entropy multi-agent dynamic games: Forward and inverse solutions. IEEE transactions on robotics , 39(3):1801–1815, 2023

    2023

  6. [13]

    Hierarchical game-theoretic planning for autonomous vehicles

    Jaime F Fisac, Eli Bronstein, Elis Stefansson, Dorsa Sadigh, S Shankar Sastry, and Anca D Dragan. Hierarchical game-theoretic planning for autonomous vehicles. In 2019 International conference on robotics and automation (ICRA) , pages 9590–9596. IEEE, 2019

    2019

  7. [14]

    Interaction-aware trajectory prediction and planning for au- tonomous vehicles in forced merge scenarios

    Kaiwen Liu, Nan Li, H Eric Tseng, Ilya Kolmanovsky, and Anouck Girard. Interaction-aware trajectory prediction and planning for au- tonomous vehicles in forced merge scenarios. IEEE Transactions on Intelligent Transportation Systems, 24(1):474–488, 2022

    2022

  8. [15]

    Strategic Decision-Making in Multi-Agent Domains: A Weighted Constrained Potential Dynamic Game Approach

    Maulik Bhatt, Yixuan Jia, and Negar Mehr. Strategic Decision-Making in Multi-Agent Domains: A Weighted Constrained Potential Dynamic Game Approach. IEEE Transactions on Robotics , 2025

    2025

  9. [16]

    Efficient constrained multi- agent trajectory optimization using dynamic potential games

    Maulik Bhatt, Yixuan Jia, and Negar Mehr. Efficient constrained multi- agent trajectory optimization using dynamic potential games. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 7303–7310. IEEE, 2023

    2023

  10. [17]

    Motion planning for mobile robots—Focusing on deep reinforcement learning: A systematic review

    Huihui Sun, Weijie Zhang, Runxiang Yu, and Yujie Zhang. Motion planning for mobile robots—Focusing on deep reinforcement learning: A systematic review. IEEE Access, 9:69061–69081, 2021

    2021

  11. [18]

    Towards physically adversarial intelligent networks (PAINs) for safer self- driving

    Piyush Gupta, Demetris Coleman, and Joshua E Siegel. Towards physically adversarial intelligent networks (PAINs) for safer self- driving. IEEE Control Systems Letters , 7:1063–1068, 2022

    2022

  12. [19]

    Motion planner with fixed-horizon constrained reinforcement learning for complex autonomous driving scenarios

    Ke Lin, Yanjie Li, Shiyu Chen, Duantengchuan Li, and Xinyu Wu. Motion planner with fixed-horizon constrained reinforcement learning for complex autonomous driving scenarios. IEEE Transactions on Intelligent Vehicles, 9(1):1577–1588, 2023

    2023

  13. [20]

    Social attention for autonomous decision-making in dense traffic

    Edouard Leurent and Jean Mercat. Social attention for autonomous decision-making in dense traffic. arXiv preprint arXiv:1911.12250 , 2019

    1911 arXiv

  14. [21]

    A review of motion planning for highway au- tonomous driving

    Laurene Claussmann, Marc Revilloud, Dominique Gruyer, and S´ebastien Glaser. A review of motion planning for highway au- tonomous driving. IEEE Transactions on Intelligent Transportation Systems, 21(5):1826–1848, 2019

    2019

  15. [22]

    Slas: Speed and lane advisory system for highway navigation

    Faizan M Tariq, David Isele, John S Baras, and Sangjae Bae. Slas: Speed and lane advisory system for highway navigation. In 2022 IEEE 61st Conference on Decision and Control (CDC) , pages 6979–6986. IEEE, 2022

    2022

  16. [23]

    Rcms: Risk-aware crash mitigation system for autonomous vehicles

    Faizan M Tariq, David Isele, John S Baras, and Sangjae Bae. Rcms: Risk-aware crash mitigation system for autonomous vehicles. In 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC), pages 3950–3957. IEEE, 2023

    2023

  17. [24]

    Delayed-Decision Motion Planning in the Presence of Multiple Predictions

    David Isele, Alexandre Miranda Anon, Faizan M Tariq, Goro Yeh, Avinash Singh, and Sangjae Bae. Delayed-Decision Motion Planning in the Presence of Multiple Predictions. In 2025 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2025

    2025

  18. [25]

    Frenet Corridor Planner: An Optimal Local Path Planning Framework for Autonomous Driving

    Faizan M Tariq, Zheng-Hang Yeh, Avinash Singh, Sangjae Bae, and David Isele. Frenet Corridor Planner: An Optimal Local Path Planning Framework for Autonomous Driving. In 2025 IEEE 36th Intelligent Vehicles Symposium (IV). IEEE, 2025

    2025

  19. [26]

    Autonomous vehicle overtaking in a bidirectional mixed-traffic setting

    Faizan M Tariq, Nilesh Suriyarachchi, Christos Mavridis, and John S Baras. Autonomous vehicle overtaking in a bidirectional mixed-traffic setting. In 2022 American Control Conference (ACC) , pages 3132–

    2022

  20. [27]

    Interaction-aware sampling-based MPC with learned local goal predictions

    Walter Jansma, Elia Trevisan, ´Alvaro Serra-G´omez, and Javier Alonso- Mora. Interaction-aware sampling-based MPC with learned local goal predictions. In 2023 International Symposium on Multi-Robot and Multi-Agent Systems (MRS) , pages 15–21. IEEE, 2023

    2023

  21. [28]

    Rehg, and Evangelos A

    Grady Williams, Paul Drews, Brian Goldfain, James M. Rehg, and Evangelos A. Theodorou. Aggressive driving with model predictive path integral control. In 2016 IEEE International Conference on Robotics and Automation (ICRA) , pages 1433–1440, 2016

    2016

  22. [29]

    Approximate inverse reinforcement learning from vision-based imitation learning

    Keuntaek Lee, Bogdan Vlahov, Jason Gibson, James M Rehg, and Evangelos A Theodorou. Approximate inverse reinforcement learning from vision-based imitation learning. In 2021 IEEE International Conference on Robotics and Automation (ICRA) , pages 10793–10799. IEEE, 2021

    2021

  23. [30]

    Blending MPC & Value Function Approximation for Efficient Reinforcement Learning

    Mohak Bhardwaj, Sanjiban Choudhury, and Byron Boots. Blending MPC & Value Function Approximation for Efficient Reinforcement Learning. In International Conference on Learning Representations , 2021

    2021

  24. [31]

    CoVO-MPC: Theoretical analysis of sampling-based MPC and optimal covariance design

    Zeji Yi, Chaoyi Pan, Guanqi He, Guannan Qu, and Guanya Shi. CoVO-MPC: Theoretical analysis of sampling-based MPC and optimal covariance design. In 6th Annual Learning for Dynamics & Control Conference, pages 1122–1135. PMLR, 2024

    2024

  25. [32]

    Trajectory distribution control for model predictive path integral control using covariance steering

    Ji Yin, Zhiyuan Zhang, Evangelos Theodorou, and Panagiotis Tsio- tras. Trajectory distribution control for model predictive path integral control using covariance steering. In 2022 International Conference on Robotics and Automation (ICRA) , pages 1478–1484. IEEE, 2022

    2022

  26. [33]

    Stein variational guided model predictive path integral control: Proposal and experi- ments with fast maneuvering vehicles

    Kohei Honda, Naoki Akai, Kosuke Suzuki, Mizuho Aoki, Hirotaka Hosogaya, Hiroyuki Okuda, and Tatsuya Suzuki. Stein variational guided model predictive path integral control: Proposal and experi- ments with fast maneuvering vehicles. In 2024 IEEE International Conference on Robo...

    2024

  27. [34]

    Variational Inference MPC using Normalizing Flows and Out-of-Distribution Projection

    Thomas Power and Dmitry Berenson. Variational Inference MPC using Normalizing Flows and Out-of-Distribution Projection. In Robotics: Science and Systems , 2022

    2022

  28. [35]

    Biased-MPPI: Informing Sampling-Based Model Predictive Control by Fusing Ancillary Con- trollers

    Elia Trevisan and Javier Alonso-Mora. Biased-MPPI: Informing Sampling-Based Model Predictive Control by Fusing Ancillary Con- trollers. IEEE Robotics and Automation Letters , 2024

    2024

  29. [36]

    Information-theoretic model predictive control: Theory and applications to autonomous driving

    Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, and Evangelos A Theodorou. Information-theoretic model predictive control: Theory and applications to autonomous driving. IEEE Trans- actions on Robotics , 34(6):1603–1622, 2018

    2018

  30. [37]

    Kinematic and dynamic vehicle models for autonomous driving control design

    Jason Kong, Mark Pfeiffer, Georg Schildbach, and Francesco Borrelli. Kinematic and dynamic vehicle models for autonomous driving control design. In IEEE Intelligent Vehicles Symposium , pages 1094–1099, 2015

    2015

  31. [38]

    Predicting parameters for modeling traffic participants

    Ahmadreza Moradipari, Sangjae Bae, Mahnoosh Alizadeh, Ehsan Moradi Pari, and David Isele. Predicting parameters for modeling traffic participants. In 2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC) , pages 703–708. IEEE, 2022

    2022

  32. [39]

    Social gan: Socially acceptable trajectories with generative adversarial networks

    Agrim Gupta, Justin Johnson, Li Fei-Fei, Silvio Savarese, and Alexan- dre Alahi. Social gan: Socially acceptable trajectories with generative adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 2255–2264, 2018

    2018

  33. [40]

    Efficient and interaction-aware trajectory planning for autonomous vehicles with particle swarm optimization

    Lin Song, David Isele, Naira Hovakimyan, and Sangjae Bae. Efficient and interaction-aware trajectory planning for autonomous vehicles with particle swarm optimization. In IEEE Intelligent Vehicles Symposium, pages 2070–2077, 2024

    2024

  34. [41]

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

    Tim Salzmann, Boris Ivanovic, Punarjay Chakravarty, and Marco Pavone. Trajectron++: Dynamically-feasible trajectory forecasting with heterogeneous data. In Computer Vision–ECCV 2020: 16th Eu- ropean Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XVIII 16, pages...

    2020

  35. [43]

    Spline interpolation techniques

    KAY A Erdogan. Spline interpolation techniques. Journal of Technical Science and Technologies, pages 47–52, 2013

    2013

  36. [44]

    A path-tracking algorithm using predictive Stanley lateral controller

    Ahmed AbdElmoniem, Ahmed Osama, Mohamed Abdelaziz, and Shady A Maged. A path-tracking algorithm using predictive Stanley lateral controller. International Journal of Advanced Robotic Systems , 17(6):1729881420974852, 2020

    2020

  37. [45]

    Congested traffic states in empirical observations and microscopic simulations

    Martin Treiber, Ansgar Hennecke, and Dirk Helbing. Congested traffic states in empirical observations and microscopic simulations. Physical review E, 62(2):1805, 2000

    2000

  38. [46]

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

    Christoph Sch ¨oller, Vincent Aravantinos, Florian Lay, and Alois Knoll. What the constant velocity model can teach us about pedestrian motion prediction. IEEE Robotics and Automation Letters , 5(2):1696–1703, 2020

    2020

  39. [47]

    Motion transformer with global intention localization and local movement refinement

    Shaoshuai Shi, Li Jiang, Dengxin Dai, and Bernt Schiele. Motion transformer with global intention localization and local movement refinement. Advances in Neural Information Processing Systems , 35:6531–6543, 2022

    2022

  40. [48]

    Bridging active exploration and uncertainty-aware deployment using probabilistic ensemble neural network dynamics

    Taekyung Kim, Jungwi Mun, Junwon Seo, Beomsu Kim, and Seongil Hong. Bridging active exploration and uncertainty-aware deployment using probabilistic ensemble neural network dynamics. arXiv preprint arXiv:2305.12240, 2023

    2023 arXiv

Pith tools

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