Swarm-Gen: Fast Generation of Diverse Feasible Swarm Behaviors

Arun Kumar Singh; B.Bhanu Teja; K.Madhava Krishna; Simon Idoko

REVIEW 4 major objections 5 minor 25 references

Swarm-Gen: Fast Generation of Diverse Feasible Swarm Behaviors

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Sampling a learned generative model and projecting the samples through a differentiable safety filter generates diverse, collision-free swarm trajectories in tens of milliseconds.

desk verdict A GPU-batched safety filter with learned warm-start is the real contribution; the multi-modal diversity claim is not supported by the single-trajectory dataset. read the letter →

arxiv 2501.19042 v1 pith:YW5TNM72 submitted 2025-01-31 cs.RO cs.AI

Simon Idoko , B.Bhanu Teja , K.Madhava Krishna , Arun Kumar Singh This is my paper · ORCID

classification cs.ROcs.AI

keywords robotswarmsmulti-modaltrajectorygenerationsafetyfilterVQ-VAECVAEfixed-pointsolverlearnedinitializationGPU-acceleratedoptimization

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

This paper claims that a practical way to generate many different, collision-free trajectories for a robot swarm is to sample from a learned generative model and then project each sample onto the space of feasible motions with a safety filter. The authors argue that this decoupling, imagination first and constraint satisfaction second, is what makes multi-modal behavior generation fast enough for real-time use. They compare a conditional VAE and a vector-quantized VAE, showing the former is faster while the latter produces more diverse behaviors, and both yield large fractions of feasible trajectories from 50 samples per start-goal pair. If the claim is correct, robotics simulators and multi-robot teams gain a way to obtain alternative coordination strategies between the same endpoints in tens of milliseconds on commodity GPUs, without tuning cost weights.

What carries the argument

The load-bearing object is the safety filter, written as $\min_\xi \tfrac{1}{2}\|\xi-\hat{\xi}\|_2^2$ subject to $A\xi=b$ and $g(\xi)\le 0$, where $\xi$ stacks polynomial coefficients for all robots. The paper rewrites the non-convex inter-robot collision and workspace ellipsoid constraints in spherical coordinates, so that each step of an alternating-minimization scheme, involving angles, distances, Lagrange multipliers, and coefficient updates, has a closed-form update and the whole loop becomes a fixed-point iteration $_{k+1}\xi,\,_{k+1}\lambda = f_{FP}({}_k\xi,\,{}_{k+1}\lambda)$ that uses only matrix-matrix products once a precomputed factorization is stored. That differentiability is what lets the authors train an initialization network, a CNN over start-goal states feeding an MLP, by unrolling $f_{FP}$ for $K$ iterations and minimizing fixed-point residuals plus displacement from the generative sample. The generative models, a VQ-VAE with a PixelCNN prior and a CVAE, supply the distribution over trajectories, while this fixed-point filter supplies feasibility; together they form the claimed real-time generator.

What would settle it

Run the pipeline on a fixed start-goal pair with a large batch size, cluster the safety-filtered trajectories by a behavioral signature such as pairwise orderings of agent positions over time, and plot the number of clusters against batch size; if the number saturates at one or two clusters even for VQ-VAE, the central claim of multi-modal diverse behavior fails, whereas continued appearance of new clusters as batch size grows would confirm it.

Watch

Extended reading notes

Core claim

The central discovery is that combining a generative prior over optimal trajectories with a learned, GPU-accelerated safety filter turns the hard problem of finding diverse feasible swarm behaviors into a two-stage sampling-and-project pipeline. Trajectories from a CVAE or VQ-VAE are polynomials parameterized by coefficients, and a differentiable quadratic-programming layer enforces boundary conditions at the decoder output. The safety filter then solves $\min_\xi \tfrac{1}{2}\|\xi-\hat{\xi}\|_2^2$ subject to $A\xi=b$ and $g(\xi)\le 0$ by an alternating-minimization fixed-point iteration that uses a spherical-coordinate reformulation of the collision and workspace constraints, so every update is closed-form, differentiable, and batchable over GPUs. Because the solver is differentiable, an initialization network can be trained end-to-end through unrolled solver iterations to warm-start the fixed point, and the paper reports that this learned initialization reduces primal residuals faster than zero or raw generative-model initializations. The paper's claim is that this architecture produces a large set of multi-modal, feasible trajectories in a few tens of milliseconds, with VQ-VAE delivering higher diversity and CVAE delivering higher speed and more feasible solutions.

Load-bearing premise

The load-bearing premise is that the training data contain the full range of multi-modal swarm behaviors, even though each sampled start-goal pair in the dataset is assigned just one optimal trajectory; if the observed variety is only sampling noise rather than distinct coordination strategies, the multi-modality claim weakens.

Editorial extensions

If this is right

A robotics simulator could generate alternative coordination behaviors on the fly, making data-driven training of navigation policies cheaper and richer without a separate optimization run per scene.
For up to eight agents, both pipelines run at interactive rates on a commodity GPU, so an operator could steer a team and receive several safe alternatives in real time.
The speed-versus-diversity trade-off is explicit: use VQ-VAE when distinct behaviors matter and batch sizes are small, and use CVAE when throughput matters or the swarm is large.
Because the initialization network is trained through the solver, it can be retrained whenever the solver or constraint set changes, keeping the speedup tied to the actual downstream optimizer.

Reading between the lines

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

A testable consequence the paper does not pursue is that its 'multi-modality' should be measured by clustering the safety-filtered trajectories for a fixed start-goal pair; if all samples fall into one or two behavioral patterns for both models, the observed diversity is mostly sampling noise rather than distinct coordination strategies.
The single-optimal-trajectory-per-pair dataset suggests the diversity is manufactured at inference time by latent sampling plus the non-convex safety-filter projection; a richer expert set containing multiple true modes per pair might change the comparison between the two generative models.
The same safety-filter-plus-generative-model recipe could extend to other constraint types, such as road geometry or fixed obstacle maps, by adding a similarly structured reformulation of those constraints.

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.

Desk Editor's Note

read the letter

Short version: the solid piece here is a GPU-batched safety filter for multi-robot trajectory generation with a learned, solver-aware warm start; the flashy diversity claim is not actually supported by the training data. If you read it for the solver, you'll be fine; if you read it for the multi-modality, you'll be disappointed.

The authors combine a CVAE or VQ-VAE with a fixed-point safety filter that projects sampled polynomial trajectories onto collision-free, workspace-satisfying, boundary-constrained sets. The filter is a batched, closed-form alternating-minimization solver, and they train a small PointNet+MLP to initialize the iterations by backpropagating through the unrolled solver. That is a real extension of [10] and [18], and the experiments show the learned initialization converges faster than zero init or raw generative samples. The CVAE vs VQ-VAE runtime/diversity trade-off is also a useful data point.

The soft spot is the central claim. The dataset is one optimal trajectory per start-goal pair. That means the learned models cannot encode multiple coordination modes for a given scenario; they can only perturb around a single expert solution. The observed diversity—low cosine similarity after SF—could just be sampling noise plus the non-convex SF projection mapping nearby initial guesses to different local optima. They never cluster the outputs or show that distinct behavior modes exist; low average pairwise similarity is consistent with one continuous blob. Calling this 'multi-modal' is an overreach given how the data was collected. The fix is not trivial: they need to generate multiple distinct expert solutions per scenario, or at least demonstrate that the modes are identifiable and consistent across seeds.

Other issues: no comparison to direct optimization with random restarts or to a diffusion baseline, no error bars or significance tests, and the SF has no convergence guarantee for the non-convex problem (though the appendix derivation is plausible). These are all fixable in revision.

Who should read it: people working on data-driven swarm simulation, and anyone interested in learning to warm-start non-convex solvers. The solver contribution alone is worth a serious look.

Recommendation: send it to review. The evaluation deficiencies are substantial but not fatal; a competent referee could push the authors to add baselines and address the mode claim, and the core solver work deserves to see the light.

Referee Report

4 major / 5 minor

Summary. The paper proposes a pipeline, Swarm-Gen, for generating diverse and feasible multi-robot swarm trajectories. A CVAE or VQ-VAE is trained on optimal trajectories produced by an existing trajectory optimizer, and at inference time many latent-space samples are drawn and then projected onto the set of collision- and workspace-feasible trajectories by a custom batched safety filter (SF). The SF is formulated as a fixed-point iteration, and a neural network is trained, in a self-supervised manner through the unrolled solver, to provide good initializations for the SF. The paper reports that the approach generates several feasible, diverse swarm behaviors in tens of milliseconds on an RTX 3090, that VQ-VAE yields more diverse outputs while CVAE is faster, and that the learned initialization accelerates SF convergence relative to zero and raw generative-sample initializations.

Significance. If the claims hold, the paper addresses a useful and relatively underexplored problem: generating multiple feasible, collision-free swarm trajectories in real time for data-driven simulation and downstream planning. The paper has clear strengths: it provides open-source code, the safety-filter derivation in the appendix is plausible, the GPU-batched fixed-point formulation is a practical contribution, and the learned warm-start idea, already validated in single-agent settings, is sensibly extended to a non-convex multi-robot setting. The comparison between CVAE and VQ-VAE is also potentially useful for practitioners. However, the significance of the central claim depends on whether the generated outputs are genuinely multi-modal in a behavioral sense and whether the feasibility and speed claims are supported by rigorous, comparative empirical evidence; on both points the current manuscript is not yet convincing.

major comments (4)

[Abstract, Section I, Section V-A, Fig. 6] The central claim of generating 'multi-modal' swarm behaviors is not established. The training data are generated by running an improved version of [10] between sampled start and goal pairs (Section V-A), which produces one deterministic optimal trajectory per pair. The learned CVAE/VQ-VAE can therefore only encode perturbations around a single expert mode per conditioning input. The diversity observed after the safety filter could arise from latent sampling noise and from the non-convex SF projection mapping nearby samples to different local optima, rather than from learned, distinct coordination modes. The paper should provide a direct mode-identification test: cluster the feasible outputs (for example, by robot-ordering or by pairwise trajectory distance) and show well-separated clusters; additionally, compare against a baseline that samples noise around the single expert trajectory and then applies the SF. Without such evidence, the abstract's 'multi-modal' claim should be weakened to 'diverse'.
[Section V-B, Fig. 6(a), Eq. (25)] The feasibility claim is not quantitatively supported. Fig. 6(a) reports the 'fraction of feasible solutions' after 200 SF iterations, but the paper does not state the feasibility tolerance, the maximum constraint violation, or whether the primal residual (25) is driven below any threshold. Since the SF solves a non-convex problem through a fixed-point iteration, 200 iterations may satisfy constraints only approximately. The paper should report the distribution of final constraint violations (collision and workspace) over the full test set, or state an explicit tolerance and the percentage of outputs meeting it.
[Section IV, Section V-B, Figs. 6-7] The empirical evaluation lacks baseline comparisons and statistical measures. The paper contrasts with diffusion-based [21] only qualitatively, and the speed/diversity claims are not compared against any prior generative method, against the authors' own single-trajectory optimizer [10], or against a simple baseline of sampling around the expert trajectory plus SF. Moreover, Figs. 6 and 7 show point estimates without error bars or significance tests; it is therefore unclear whether the reported CVAE-versus-VQ-VAE differences in feasibility, diversity, and runtime would replicate across seeds and test scenes. Please add error bars over multiple random seeds and a baseline comparison on the same hardware.
[Section V-C, Fig. 8] The claim that the learned initialization provides faster convergence is based on a single residual plot without quantitative metrics. Fig. 8 shows primal-residual curves, but no iteration count to a target residual, no wall-clock time, and no error bars over multiple scenes. Since 'faster convergence' is one of the two stated contributions, the paper should define a convergence metric (for example, iterations or time to reach a specified residual) and report it over the same benchmark used in the diversity experiments.

minor comments (5)

[Eq. (1)] The boundary condition at time H is written as (p_{i|H}, \ddot p_{i|H}, \ddot p_{i|H}) = b_H; the second entry should presumably be \dot p_{i|H}.
[Eq. (11)] The fixed-point iteration is written as (k+1 ξ, k+1 λ) = f_FP(k ξ, k+1 λ); the right-hand side should use k λ (or the notation should be clarified), since the update uses the current multiplier.
[Eq. (12)] The summation limits are written as 'k−1 X k=0', which appears to be a typo for K−1; please correct the index.
[Eqs. (22)-(24)] The Lagrange multiplier λ in the augmented Lagrangian (22) is written as if it multiplies ξ, while the update in (23d) is for the residual Fξ − e. This can be made consistent by defining λ as F^T times the multiplier of the equality constraint, but the paper should state this redefinition explicitly.
[Abstract, Section V-B] Minor typos: 'thecinitialization' in the abstract, 'files closer' in the caption of Fig. 1, and 'is ran' in the caption of Fig. 7. Also, the caption of Fig. 6 does not mention what the box plots or error bars (if any) represent.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: generative models, safety filter, and initialization network are fitted independently, with no prediction reducing to its inputs by construction.

full rationale

The paper's derivation chain is not circular. The generative models (CVAE and VQ-VAE) are trained only on reconstruction, codebook, commitment, or KL losses (Eqs. 7 and 9), while the diversity and feasibility metrics (Fig. 6, Eq. 25) are measured on post-hoc outputs and are never used as training targets. The safety filter is a standalone projection (Eq. 10) onto independently stated boundary, workspace, and collision constraints (Eqs. 1-3), with the fixed-point solver derived in the Appendix (Eqs. 23-24). Although the text says the solver is 'an improved and batched version of the optimizer presented in [10]' and the expert dataset is produced by 'an improved version of [10]', the present paper provides the actual fixed-point equations rather than importing the solver as a black-box theorem; the self-citations to [10], [24], and [25] are normal author-continuation and are not used to forbid alternatives or to establish uniqueness. The nearest concern is that Section V-A collects only one expert trajectory per sampled start-goal pair, which may weaken the multi-modality claim; however, this is an empirical limitation, not a circular reduction, because diversity could still arise from latent-space sampling and from the non-convex safety-filter projection, and no equation equates the claimed diversity with any fitted value. The absence of external benchmarks affects generalizability, not circularity. Score 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a small number of hand-chosen hyperparameters (penalty weight, iteration count, latent sizes) and on the unproved convergence of a non-convex fixed-point projection. No new physical entities are introduced; the initialization network and batched solver are computational components, not invented entities.

free parameters (5)

Augmented Lagrangian penalty rho = not reported
Constant in Eq. (22) that weights constraint violation in the SF objective; chosen by hand, affects convergence and feasibility of the fixed-point solver.
Number of SF iterations = 200
The paper runs the SF for 200 iterations in experiments (Section V-B); feasibility is only empirically verified at this iteration count, with no convergence guarantee.
VQ-VAE codebook size N = 512
Architecture choice in Section V-A that determines the discrete latent capacity and hence diversity; set by hand.
VQ-VAE latent vector size L = 25 (4/8 agents), 100 (16 agents)
Set per swarm size in Section V-A; influences diversity and compute.
VQ-VAE commitment loss weight beta = not reported
Hyperparameter in Eq. (7) balancing codebook and commitment losses; not specified.

assumptions (4)

standard math The spherical reformulation (13)-(14) exactly represents all spheroid collision-avoidance and workspace constraints.
Used to derive the SF in the Appendix; the transformation is a change of variables and is exact for spheroids.
domain assumption The alternating-minimization fixed-point iteration (11) converges to a feasible solution within 200 iterations for all sampled inputs.
The SF problem is non-convex; no convergence proof is given. The paper relies on empirical residual plots (Fig. 8) and the statement that 200 iterations sufficed.
domain assumption The learned generative models approximate the true distribution of optimal swarm trajectories, including its multi-modality.
The VQ-VAE/CVAE are trained on expert trajectories; whether the latent space captures distinct modes depends on dataset diversity, which is not established.
domain assumption The self-supervised loss (12), minimized through unrolled fixed-point iterations, produces an initialization that accelerates the SF solver without biasing the feasible set.
The training objective includes a term to stay close to the generative prediction, which could bias solutions; the paper assumes this does not harm feasibility or diversity.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Swarm-Gen: Fast Generation of Diverse Feasible Swarm Behaviors." pith.science (2026). https://pith.science/paper/YW5TNM72

@misc{pith2026250119042,
  author       = {Pith},
  title        = {Pith review of: Swarm-Gen: Fast Generation of Diverse Feasible Swarm Behaviors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YW5TNM72}},
  note         = {Machine review of arXiv:2501.19042}
}

read the original abstract

Coordination behavior in robot swarms is inherently multi-modal in nature. That is, there are numerous ways in which a swarm of robots can avoid inter-agent collisions and reach their respective goals. However, the problem of generating diverse and feasible swarm behaviors in a scalable manner remains largely unaddressed. In this paper, we fill this gap by combining generative models with a safety-filter (SF). Specifically, we sample diverse trajectories from a learned generative model which is subsequently projected onto the feasible set using the SF. We experiment with two choices for generative models, namely: Conditional Variational Autoencoder (CVAE) and Vector-Quantized Variational Autoencoder (VQ-VAE). We highlight the trade-offs these two models provide in terms of computation time and trajectory diversity. We develop a custom solver for our SF and equip it with a neural network that predicts context-specific initialization. Thecinitialization network is trained in a self-supervised manner, taking advantage of the differentiability of the SF solver. We provide two sets of empirical results. First, we demonstrate that we can generate a large set of multi-modal, feasible trajectories, simulating diverse swarm behaviors, within a few tens of milliseconds. Second, we show that our initialization network provides faster convergence of our SF solver vis-a-vis other alternative heuristics.

Figures

Figures reproduced from arXiv: 2501.19042 by the authors.

**Figure 1.** The first column shows the trajectories sampled from CVAE and VQ-VAE. These are then passed through the SF. The plots from second column onward show diverse feasible swarm coordination trajectories obtained after SF operations. The start and goal positions are marked with cross and red-dot respectively. Abstract— Coordination behavior in robot swarms is inherently multi-modal in nature. That is, there are numerous … view at source ↗

**Figure 2.** Architecture of our VQ-VAE pipeline. A CNN encoder compresses optimal trajectories (time-stamped waypoints) into a discrete latent space Zq. The decoder reconstructs polynomial coefficients ξ based on Zq, which is then converted to trajectories. The differentiable QP block ensures that the reconstructed trajectories satisfy the boundary conditions on start and goal states. QP layer to obtain the coefficient ξ used t… view at source ↗

**Figure 3.** Architecture of PixelCNN used to sample from the VQ-VAE. Its output is a multinomial distribution over hq which can be used to generate different samples of it. These are then used to generate different Zq and consequently different multi-robot trajectories. B. CVAE Prior Over Optimal Trajectories The auto-regressive nature of PixelCNN implies that sampling from a trained VQ-VAE would be computationally expensive, … view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: Training pipeline for learning context specific initialization for our SF solver. It consists of learnable layers [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 5.** Figure 5: Our pipeline to fit CVAE to the dataset of expert multi-robot trajectories. The differentiable QP block is simply (10) without the inequality constraints. C. Safety-Filter(SF) Our SF is defined by the following optimization problem. min ξ 1 2 [PITH_FULL_IMAGE:figures…

**Figure 6.** Figure 6: Fig.(a) represents the fraction of the trajectories sampled from VQ-VAE/CVAE that becomes feasible after applying SF for 200 iterations. Fig.(b) compares the diversity of CVAE and VQ-VAE models through the cosine similarity metric computed over the feasible trajectorie…

**Figure 7.** Figure 7: Fig(a)-(b): Computation time with respect to batch size (number of sampled trajectories from VQ-VAE/CVAE) when the SF is ran for 200 iterations. Fig.(c)-(d): Computation time with respect to the number of SF iterations for a given batch size of 10. Computation Time: Fi…

**Figure 8.** Figure 8: Primal residuals of the SF solver with different initialization strategies. VI. CONCLUSION We tackled an important but relatively unaddressed problem of generating multiple feasible and diverse trajectories for robot swarms. At an abstract level, this is similar to id…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 20 canonical work pages

[10]

Gpu accelerated convex approximations for fast multi- agent trajectory optimization,

F. Rastgar, H. Masnavi, J. Shrestha, K. Kruusam ¨ae, A. Aabloo, and A. K. Singh, “Gpu accelerated convex approximations for fast multi- agent trajectory optimization,” IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3303–3310, 2021

work page 2021
[18]

Learning to warm-start fixed-point optimization algorithms,

R. Sambharya, G. Hall, B. Amos, and B. Stellato, “Learning to warm-start fixed-point optimization algorithms,” Journal of Machine Learning Research, vol. 25, no. 166, pp. 1–46, 2024

work page 2024
[21]

Motiondiffuser: Controllable multi-agent motion prediction using diffusion,

C. Jiang, A. Cornman, C. Park, B. Sapp, Y . Zhou, D. Anguelov et al. , “Motiondiffuser: Controllable multi-agent motion prediction using diffusion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 9644–9653

work page 2023
[1]

Heter- sim: Heterogeneous multi-agent systems simulation by interactive data-driven optimization,

J. Ren, W. Xiang, Y . Xiao, R. Yang, D. Manocha, and X. Jin, “Heter- sim: Heterogeneous multi-agent systems simulation by interactive data-driven optimization,” IEEE transactions on visualization and computer graphics, vol. 27, no. 3, pp. 1953–1966, 2019

work page 1953
[2]

B-gap: Behavior-rich simulation and navigation for autonomous driving,

A. Mavrogiannis, R. Chandra, and D. Manocha, “B-gap: Behavior-rich simulation and navigation for autonomous driving,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 4718–4725, 2022

work page 2022
[3]

Gpudrive: Data-driven, multi-agent driving simulation at 1 mil- lion fps,

S. Kazemkhani, A. Pandya, D. Cornelisse, B. Shacklett, and E. Vinit- sky, “Gpudrive: Data-driven, multi-agent driving simulation at 1 mil- lion fps,” arXiv preprint arXiv:2408.01584 , 2024

arXiv 2024
[4]

Target-visible polynomial trajectory generation within an mav team,

Y . Lee, J. Park, B. Jeon, and H. J. Kim, “Target-visible polynomial trajectory generation within an mav team,” in 2021 IEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems (IROS) . IEEE, 2021, pp. 1982–1989

work page 2021
[5]

Communication through motion: Legibility of multi-robot systems,

B. Capelli, C. Secchi, and L. Sabattini, “Communication through motion: Legibility of multi-robot systems,” in 2019 International Symposium on Multi-Robot and Multi-Agent Systems (MRS) . IEEE, 2019, pp. 126–132

work page 2019

Show all 25 references

[6]

Amswarm: An alternating minimization approach for safe motion planning of quadrotor swarms in cluttered environments,

V . K. Adajania, S. Zhou, A. K. Singh, and A. P. Schoellig, “Amswarm: An alternating minimization approach for safe motion planning of quadrotor swarms in cluttered environments,” in 2023 IEEE Interna- tional Conference on Robotics and Automation (ICRA) . IEEE, 2023, pp. 1421–1427

2023
[7]

Amswarmx: Safe swarm coordination in complex environments via implicit non-convex decomposition of the obstacle-free space,

——, “Amswarmx: Safe swarm coordination in complex environments via implicit non-convex decomposition of the obstacle-free space,” in 2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2024, pp. 14 555–14 561

2024
[8]

Distributed predictive drone swarms in cluttered environments,

E. Soria, F. Schiano, and D. Floreano, “Distributed predictive drone swarms in cluttered environments,” IEEE Robotics and Automation Letters, vol. 7, no. 1, pp. 73–80, 2021

2021
[9]

Generation of collision-free trajectories for a quadrocopter fleet: A sequential convex programming approach,

F. Augugliaro, A. P. Schoellig, and R. D’Andrea, “Generation of collision-free trajectories for a quadrocopter fleet: A sequential convex programming approach,” in 2012 IEEE/RSJ international conference on Intelligent Robots and Systems . IEEE, 2012, pp. 1917–1922

2012
[11]

Smart: Simultane- ous multi-agent recurrent trajectory prediction,

N. Sriram, B. Liu, F. Pittaluga, and M. Chandraker, “Smart: Simultane- ous multi-agent recurrent trajectory prediction,” in Computer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXVII 16 . Springer, 2020, pp. 463–479

2020
[12]

Continual multi- agent interaction behavior prediction with conditional generative mem- ory,

H. Ma, Y . Sun, J. Li, M. Tomizuka, and C. Choi, “Continual multi- agent interaction behavior prediction with conditional generative mem- ory,” IEEE Robotics and Automation Letters , vol. 6, no. 4, pp. 8410– 8417, 2021

2021
[13]

Congestion- aware multi-agent trajectory prediction for collision avoidance,

X. Xie, C. Zhang, Y . Zhu, Y . N. Wu, and S.-C. Zhu, “Congestion- aware multi-agent trajectory prediction for collision avoidance,” in 2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 13 693–13 700

2021
[14]

Tra- jectron++: Dynamically-feasible trajectory forecasting with heteroge- neous data,

T. Salzmann, B. Ivanovic, P. Chakravarty, and M. Pavone, “Tra- jectron++: Dynamically-feasible trajectory forecasting with heteroge- neous data,” in Computer Vision–ECCV 2020: 16th European Confer- ence, Glasgow, UK, August 23–28, 2020, Proceedings, Part XVIII 16 . Springer, 2...

2020
[15]

Neural discrete representation learning,

A. v. d. Oord, O. Vinyals, and K. Kavukcuoglu, “Neural discrete representation learning,” arXiv preprint arXiv:1711.00937 , 2017

2017 arXiv
[16]

Conditional image generation with pixelcnn decoders,

A. Van den Oord, N. Kalchbrenner, L. Espeholt, O. Vinyals, A. Graves et al. , “Conditional image generation with pixelcnn decoders,” Ad- vances in neural information processing systems , vol. 29, 2016

2016
[17]

Learning structured output represen- tation using deep conditional generative models,

K. Sohn, H. Lee, and X. Yan, “Learning structured output represen- tation using deep conditional generative models,” Advances in neural information processing systems , vol. 28, 2015

2015
[19]

Multi-agent tensor fusion for contextual trajectory prediction,

T. Zhao, Y . Xu, M. Monfort, W. Choi, C. Baker, Y . Zhao, Y . Wang, and Y . N. Wu, “Multi-agent tensor fusion for contextual trajectory prediction,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 12 126–12 134

2019
[20]

Agentformer: Agent- aware transformers for socio-temporal multi-agent forecasting,

Y . Yuan, X. Weng, Y . Ou, and K. M. Kitani, “Agentformer: Agent- aware transformers for socio-temporal multi-agent forecasting,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9813–9823

2021
[22]

Transformer-based model predictive control: Trajectory optimization via sequence modeling,

D. Celestini, D. Gammelli, T. Guffanti, S. D’Amico, E. Capello, and M. Pavone, “Transformer-based model predictive control: Trajectory optimization via sequence modeling,” IEEE Robotics and Automation Letters, 2024

2024
[23]

Pilot: Efficient planning by imitation learning and optimisation for safe autonomous driving,

H. Pulver, F. Eiras, L. Carozza, M. Hawasly, S. V . Albrecht, and S. Ramamoorthy, “Pilot: Efficient planning by imitation learning and optimisation for safe autonomous driving,” in 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2021, pp...

2021
[24]

End-to-end learning of behavioural inputs for autonomous driving in dense traffic,

J. Shrestha, S. Idoko, B. Sharma, and A. K. Singh, “End-to-end learning of behavioural inputs for autonomous driving in dense traffic,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2023, pp. 10 020–10 027

2023
[25]

Learning sampling distribution and safety filter for autonomous driving with vq-vae and differentiable optimization,

S. Idoko, B. Sharma, and A. K. Singh, “Learning sampling distribution and safety filter for autonomous driving with vq-vae and differentiable optimization,” arXiv preprint arXiv:2403.19461 , 2024

2024 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [10]

Gpu accelerated convex approximations for fast multi- agent trajectory optimization,

F. Rastgar, H. Masnavi, J. Shrestha, K. Kruusam ¨ae, A. Aabloo, and A. K. Singh, “Gpu accelerated convex approximations for fast multi- agent trajectory optimization,” IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 3303–3310, 2021

work page 2021

[2] [18]

Learning to warm-start fixed-point optimization algorithms,

R. Sambharya, G. Hall, B. Amos, and B. Stellato, “Learning to warm-start fixed-point optimization algorithms,” Journal of Machine Learning Research, vol. 25, no. 166, pp. 1–46, 2024

work page 2024

[3] [21]

Motiondiffuser: Controllable multi-agent motion prediction using diffusion,

C. Jiang, A. Cornman, C. Park, B. Sapp, Y . Zhou, D. Anguelov et al. , “Motiondiffuser: Controllable multi-agent motion prediction using diffusion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 9644–9653

work page 2023

[4] [1]

Heter- sim: Heterogeneous multi-agent systems simulation by interactive data-driven optimization,

J. Ren, W. Xiang, Y . Xiao, R. Yang, D. Manocha, and X. Jin, “Heter- sim: Heterogeneous multi-agent systems simulation by interactive data-driven optimization,” IEEE transactions on visualization and computer graphics, vol. 27, no. 3, pp. 1953–1966, 2019

work page 1953

[5] [2]

B-gap: Behavior-rich simulation and navigation for autonomous driving,

A. Mavrogiannis, R. Chandra, and D. Manocha, “B-gap: Behavior-rich simulation and navigation for autonomous driving,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 4718–4725, 2022

work page 2022

[6] [3]

Gpudrive: Data-driven, multi-agent driving simulation at 1 mil- lion fps,

S. Kazemkhani, A. Pandya, D. Cornelisse, B. Shacklett, and E. Vinit- sky, “Gpudrive: Data-driven, multi-agent driving simulation at 1 mil- lion fps,” arXiv preprint arXiv:2408.01584 , 2024

arXiv 2024

[7] [4]

Target-visible polynomial trajectory generation within an mav team,

Y . Lee, J. Park, B. Jeon, and H. J. Kim, “Target-visible polynomial trajectory generation within an mav team,” in 2021 IEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems (IROS) . IEEE, 2021, pp. 1982–1989

work page 2021

[8] [5]

Communication through motion: Legibility of multi-robot systems,

B. Capelli, C. Secchi, and L. Sabattini, “Communication through motion: Legibility of multi-robot systems,” in 2019 International Symposium on Multi-Robot and Multi-Agent Systems (MRS) . IEEE, 2019, pp. 126–132

work page 2019