Pith. sign in

REVIEW 5 major objections 6 minor 31 references

Landing Trajectory Prediction for UAS Based on Generative Adversarial Network

T0 review · 5 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A GAN that reads ten past positions predicts drone landing trajectories with lower error than the GMR baseline on real pilot-flown data.

desk verdict A direct Social-GAN port to UAS landing with a genuinely useful new dataset and a plausible real-data result, but the abstract overclaims and the statistics are too thin. read the letter →

arxiv 2411.14403 v1 pith:LOFJHWS2 submitted 2024-11-21 cs.RO cs.AI

Jun Xiang , Drake Essick , Luiz Gonzalez Bautista , Junfei Xie , Jun Chen This is my paper · ORCID
classification cs.ROcs.AI
keywords generativeadversarialnetworktrajectorypredictionunmannedaircraftsystemslandingmanagementLSTMGaussianmixtureregressionaveragedisplacementerrorvertiport
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 tries to show that a generative adversarial network can predict the landing trajectory of an unmanned aircraft directly from its recent positions, without an explicit flight-dynamics model or control inputs. The proposed LSTM-based generator reads the previous ten position samples and outputs the next ten; a second LSTM discriminator scores whether a trajectory looks real. Comparing against Gaussian Mixture Regression on two simulated datasets and a newly collected dataset of more than 2,600 real pilot-flown drone landings, the authors report lower average displacement error for the GAN on the real data at all ten predicted points, and at the first four points on the simulated data. The paper also claims the trained discriminator can distinguish true from generated trajectories, making it usable as a learned evaluator. The authors note accuracy degrades as the prediction horizon lengthens, so the method is presented as a short-horizon predictor.

What carries the argument

The load-bearing mechanism is the adversarial training loop between an LSTM-based generator and an LSTM-based discriminator. The generator follows the Social-GAN structure: an LSTM encoder converts each embedded position into hidden states, a pooling module combines the hidden states with environment factors, and an LSTM decoder produces the ten predicted positions. The discriminator maps a trajectory to a scalar score, and the generator's loss is driven by that score. The identity carrying the argument is $T_{t_1,t_2}=G(T_{t_0,t_1})$, which replaces the ordinary differential equation $x_{t+1}=f(x_t,a_t(E))$ with a learned map from past positions alone.

What would settle it

A decisive test would be to retrain both models on the real dataset with a held-out pilot: if the GAN's all-ten-point ADE advantage over GMR disappears when the test trajectories come from a pilot whose landings were excluded from training, the claim that the model captures general pilot-control regularities would be refuted.

Watch

Extended reading notes

Core claim

The central claim is that a GAN whose generator is an LSTM encoder-pooling-decoder can learn the mapping $T_{t_1,t_2}=G(T_{t_0,t_1})$ from a ten-point past trajectory to the next ten positions, and that this learned mapping is more accurate than the GMR baseline on real pilot-controlled landings. On the real DJI Tello dataset, the GAN beats GMR at all ten predicted points; on the vertical and linear simulated datasets, it beats GMR only at the first four points, and GMR is better afterwards. The authors interpret the real-data advantage as the GAN capturing subtle control regularities that the Gaussian model misses, and they present the discriminator's higher scores for true trajectories as evidence that it can serve as a learned judge of prediction quality. The same tables show a limitation stated in the conclusion: error increases with prediction length, so the method is not claimed to work for long-term prediction.

Load-bearing premise

The load-bearing premise is that a ten-point history of positions alone contains enough information to predict the next ten positions, so the flight-dynamics model, control inputs, and environment can be omitted from the predictor.

Editorial extensions

If this is right

  • Near-vertiport conflict detection and avoidance can use a GAN-based predictor that produces a ten-step landing forecast faster than a dynamics-based model.
  • On real pilot-flown data, the GAN's advantage over GMR persists across all ten forecast points, suggesting learned models are better suited than Gaussian regression to irregular human control styles.
  • On the simulated datasets, the GAN only improves on GMR for the first four points; users needing longer-horizon simulated forecasts should keep GMR as the baseline.
  • The trained discriminator can be reused to score new candidate trajectories, offering a learned evaluation signal without requiring a physics model.

Reading between the lines

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

  • Because the model ingests only raw positions, it is likely sensitive to training distribution; a natural next test is whether the GAN retains its real-data advantage when wind, payload, or pilot identity shifts between training and test flights.
  • The simulated-data result (advantage only for four points) suggests the GAN-versus-GMR comparison is entangled with data regularity; a simulated dataset with pilot-like stochastic control would separate the model-class effect from the data effect.
  • The discriminator score could be repurposed as an online confidence monitor that flags predictions whose score drops, without retraining.
  • A straightforward extension is autoregressive rollouts, feeding predicted points back as input to extend the horizon, which the paper's stated short-horizon limitation implies.
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

5 major / 6 minor

Summary. The paper proposes an LSTM-based generative adversarial network for predicting UAS landing trajectories. The generator observes the previous 10 position samples and outputs 10 future positions, while a discriminator is trained to score trajectory quality. The method is compared with Gaussian Mixture Regression (GMR) on two simulated datasets and a real drone-landing dataset collected by the authors. The paper claims that the GAN outperforms GMR in accuracy across datasets and also claims that the trained discriminator can be used to evaluate other predictions.

Significance. If the real-data advantage holds under rigorous statistical testing and the dataset is released, the paper makes a useful empirical contribution: a GAN-based trajectory predictor and a real pilot-controlled UAS landing dataset. The idea of reusing the trained discriminator as a score function is also interesting, though it is not validated beyond the narrow setting of distinguishing true versus GAN-generated trajectories. The paper's main weakness is that the evidence is currently insufficient: the abstract overstates the results, the simulated-dataset rows of Table 1 contradict the 'various datasets' claim, the statistical basis for the real-dataset advantage is missing, and the GMR baseline configuration is undocumented. No code or dataset link is provided, which limits reproducibility.

major comments (5)
  1. [Abstract and Section IV.A, Table 1] The abstract states that the proposed model 'can output more accurate predictions than the baseline method (GMR) in various datasets.' This is contradicted by Table 1: for both the vertical and linear simulated datasets, the GAN has higher mean ADE than GMR for points 5 through 10 (e.g., point 10 vertical: GAN 14.88±8.41 vs GMR 9.0±3.69; point 10 linear: GAN 17.01±8.97 vs GMR 10.93±5.72). The real dataset is the only one where the GAN is consistently better. The claim must be narrowed to the real dataset, or the discrepancy must be analyzed and explained.
  2. [Section IV.A, Table 1] The table reports values as mean ± spread but never states whether the spread is a standard deviation or a standard error, and no significance tests are provided. For the real dataset, the intervals overlap at later horizons (point 9: GAN 0.14±0.07 vs GMR 0.21±0.16; point 10: GAN 0.18±0.08 vs GMR 0.24±0.21). The paper should report per-trajectory paired errors, state the exact definition of the spread, and apply a paired statistical test (e.g., Wilcoxon signed-rank test) or bootstrap confidence intervals to demonstrate that the real-data advantage is not sampling noise.
  3. [Section IV.A, GMR baseline] The GMR baseline is not documented sufficiently for the comparison to be reproducible. The number of Gaussian components K in Eq. (2) is not given, the input and output dimensional setup is not described (e.g., whether GMR also uses 10 past positions to predict 10 future positions), and there is no statement about hyperparameter selection or cross-validation. Without these details, the reader cannot rule out that the GAN's advantage comes from a poorly tuned baseline. The authors should specify the GMR configuration and ideally include a sensitivity analysis over K.
  4. [Section III.A, Eq. (10)] The modeling premise that 'environment factors and previous status are only variables required to derive a trajectory' and the resulting mapping T_{t1,t2} = G(T_{t0,t1}) assume that ten consecutive position samples contain all information needed to forecast the next ten positions. This excludes explicit flight dynamics, control inputs, and environmental context. The assumption is load-bearing because if the input representation is insufficient, the model cannot generalize beyond its training distribution. The paper should explicitly acknowledge this as a limitation, justify it empirically (e.g., by an ablation over input length or comparison with a physics-based predictor), and state the conditions under which the assumption holds.
  5. [Section IV.B, Table 2] The contribution that the trained discriminator 'can be used to evaluate other predictions' is not supported by the presented evidence. Table 2 reports scores for true and GAN-generated trajectories from the same datasets used in training. For the real dataset, the score intervals overlap substantially (true 2.32±2.76 vs fake 0.21±0.86), and no significance test is reported. The discriminator is trained to distinguish real trajectories from those generated by this specific GAN, not to score arbitrary trajectory quality. To support the claimed use as a general evaluator, additional experiments on out-of-distribution trajectories or a clear operational definition of 'good' trajectory are needed.
minor comments (6)
  1. [Abstract and Section IV] The number of trajectories is inconsistent: the abstract says 'more than 2600 trajectories,' while Section IV reports 2430 training trajectories and 100 evaluation trajectories, which sum to 2530. Please clarify the actual total.
  2. [Section III.C, Eqs. (13)-(15)] The notation is unclear: the pooling function P, the embedding function phi, and the activation gamma are not defined. Eq. (14) introduces et = phi(x_{t-1}; W_em) after et was already defined in Eq. (11). Please define all symbols and remove the redundancy.
  3. [Abstract] The word 'experiences' should be 'experiments' in the sentence 'According to the results of the experiences.'
  4. [Table 1] The units of ADE are not stated. Please specify the units (e.g., meters) and the time interval between consecutive trajectory points.
  5. [Section V and Figure 6] The conclusion says 'The prediction error is lower than the baseline method' without adding the qualifier that this holds for the real dataset at all points and for the simulated datasets only for the first four points. Also, the caption of Figure 6 reads 'GMM prediction' but the baseline is GMR; please use consistent terminology.
  6. [Figure 8] The axes and the plotted quantity (per-point mean displacement, standard deviation, or something else) are not labeled. Please clarify the content of the figure.

Circularity Check

1 steps flagged · score 2.0 of 10

Main GAN-vs-GMR comparison is an independent empirical benchmark; only the discriminator-as-evaluator contribution is mildly self-referential.

  1. fitted input called prediction [Section I contribution bullet and Section IV.B, Table 2]
    "After GANs are well-trained, the discriminator part can be used to evaluate other predictions. ... We show the discriminator 100 true trajectories and 100 fake trajectories from each dataset and let the discriminator generates the scores for those trajectories. ... the discriminator gives higher scores to the true trajectories. It proves that the discriminator can distinguish true and fake trajectories."

    The discriminator's output is optimized by Eq. (1) to give high scores to real trajectories and low scores to generated ones. Table 2 demonstrates exactly this behavior on trajectories from the same datasets used to train the GAN. The claimed ability to 'evaluate other predictions' is therefore supported only by an in-sample confirmation of the discriminator's own training objective, not by an independent measure of trajectory quality. This is a fitted input being presented as a general evaluation capability, though it does not affect the ADE comparison in Table 1.

full rationale

The paper's central claim is an empirical comparison: GAN and GMR are trained and evaluated on held-out trajectory sets, with ADE computed against ground truth (Table 1). This is a standard supervised benchmark, not a derivation, so the main result does not reduce to its inputs. The GMR baseline is an external method (refs. 30-31), and the self-citations in refs. 3 and 11 are background references rather than load-bearing support for the prediction result. The only mildly self-referential element is the third contribution, where the trained discriminator is proposed as a general trajectory evaluator and then validated by showing that it scores true trajectories higher than fake ones; since that behavior is precisely what the adversarial objective trains it to do, the evaluator claim is a sanity check of the training objective rather than an independent proof. This does not undermine the main comparison, so the overall circularity score is low.

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

The central comparison rests on the modeling premise that past positions alone determine future positions, on an unspecified LSTM-GAN configuration, and on an unvalidated discriminator quality score. The GMR baseline's configuration is also undisclosed. These choices, not the GAN concept itself, determine whether the reported ADE numbers are meaningful.

free parameters (4)
  • GAN training hyperparameters = not reported
    Learning rate, batch size, epochs, optimizer, LSTM hidden size, embedding dimension, pooling dimension, and latent noise dimension are never stated, yet they determine the reported ADE values.
  • GMR number of Gaussian components K = not reported
    The GMR baseline in equation (2) requires a choice of K, but the paper does not state it, so the comparison may be unfair or unreproducible.
  • Trajectory history and prediction length = 10 points
    The paper fixes both the observed and predicted sequences to 10 points without justification; the ADE results degrade with horizon, so this choice materially affects the conclusions.
  • Environment factor E and random input z = unspecified
    Equation (13) defines pooling as P(E, z), but the paper never specifies what E and z are for the UAS landing task, leaving a free modeling choice that is not documented.
assumptions (4)
  • domain assumption Past trajectory plus environment factors are sufficient to predict future positions
    Section III.A states that flight dynamics and control inputs can be bypassed because only environment and previous status are required to derive a trajectory, giving T_{t1,t2} = G(T_{t0,t1}).
  • domain assumption LSTM can encode trajectory sequences and generate plausible continuations
    Section III.C adopts the LSTM encoder-decoder architecture from Social-GAN without empirical justification for this domain or comparison to other sequence models.
  • domain assumption GAN training reaches a useful equilibrium on this dataset
    The min-max objective in equation (1) is stated, but no convergence criteria, mode-collapse checks, or training stability analysis are provided.
  • ad hoc to paper The discriminator's score is a valid trajectory quality measure
    Section IV.B claims the trained discriminator can evaluate predictions, but the only evidence is higher scores for training-distribution real trajectories versus generated ones, which does not establish validity for unseen predictions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Landing Trajectory Prediction for UAS Based on Generative Adversarial Network." pith.science (2026). https://pith.science/paper/LOFJHWS2

@misc{pith2026241114403,
  author       = {Pith},
  title        = {Pith review of: Landing Trajectory Prediction for UAS Based on Generative Adversarial Network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LOFJHWS2}},
  note         = {Machine review of arXiv:2411.14403}
}
read the original abstract

Models for trajectory prediction are an essential component of many advanced air mobility studies. These models help aircraft detect conflict and plan avoidance maneuvers, which is especially important in Unmanned Aircraft systems (UAS) landing management due to the congested airspace near vertiports. In this paper, we propose a landing trajectory prediction model for UAS based on Generative Adversarial Network (GAN). The GAN is a prestigious neural network that has been developed for many years. In previous research, GAN has achieved many state-of-the-art results in many generation tasks. The GAN consists of one neural network generator and a neural network discriminator. Because of the learning capacity of the neural networks, the generator is capable to understand the features of the sample trajectory. The generator takes the previous trajectory as input and outputs some random status of a flight. According to the results of the experiences, the proposed model can output more accurate predictions than the baseline method(GMR) in various datasets. To evaluate the proposed model, we also create a real UAV landing dataset that includes more than 2600 trajectories of drone control manually by real pilots.

Figures

Figures reproduced from arXiv: 2411.14403 by the authors.

Figure 1
Figure 1. Vertical Landing Trajectory [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Linear Landing Trajectory 4 of 10 American Institute of Aeronautics and Astronautics [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Real Landing Trajectory [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: DJI Tello Drone C. LSTM-Pooling based Generator The trajectory can be treated as a kind of data sequence, therefore, LSTM, which is one of the best neural network structures to understand data sequences, can be used to understand and generate the trajectory. Social-GAN…
Figure 5
Figure 5. Figure 5: Training Process D. Discriminator Instead of a traditional mathematical model, we use a neural network as a discriminator to evaluate the generated trajectory. Based on the author’s knowledge, there is no traditional mathematical model that is completely able to identi…
Figure 6
Figure 6. Figure 6: GMM prediction [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: GANs prediction [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: prediction displacement in each axis(real dataset) [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 27 canonical work pages

  1. [1]

    The Federal Aviation Administration (FAA) Aerospace Forecast Fiscal Years (FY) 2020-2040,

    FAA, “The Federal Aviation Administration (FAA) Aerospace Forecast Fiscal Years (FY) 2020-2040,” 2020

    work page 2020

  2. [2]

    Prediction of MUA V flight behavior from active and passive imaging in complex environment,

    Laurenzis, M., Rebert, M., Schertzer, S., Bacher, E., and Christnacher, F., “Prediction of MUA V flight behavior from active and passive imaging in complex environment,” Laser Radar Technology and Applications XXV , V ol. 11410, SPIE, 2020, pp. 10–17

    work page 2020

  3. [3]

    Optimized landing of drones in the context of congested air traffic and limited vertiports,

    Zhou, Z., Chen, J., and Liu, Y ., “Optimized landing of drones in the context of congested air traffic and limited vertiports,” IEEE Transactions on Intelligent Transportation Systems, V ol. 22, No. 9, 2020, pp. 6007–6017

    work page 2020

  4. [4]

    UA V path prediction for CD&R to manned aircraft in a confined airspace for cooperative mission,

    Lin, C. E. and Lai, Y .-H., “UA V path prediction for CD&R to manned aircraft in a confined airspace for cooperative mission,” International Journal of Aerospace Engineering , V ol. 2018, 2018

    work page 2018

  5. [5]

    Safety Assured Online Guidance With Airborne Separation for Urban Air Mobility Operations in Uncertain Environments,

    Wu, P., Yang, X., Wei, P., and Chen, J., “Safety Assured Online Guidance With Airborne Separation for Urban Air Mobility Operations in Uncertain Environments,” IEEE Transactions on Intelligent Transportation Systems , V ol. 23, No. 10, 2022, pp. 19413 – 19427

    work page 2022

  6. [6]

    Model-based path prediction for fixed-wing unmanned aircraft using pose estimates,

    Kang, C. and Woolsey, C. A., “Model-based path prediction for fixed-wing unmanned aircraft using pose estimates,” Aerospace Science and Technology, V ol. 105, 2020, pp. 106030

    work page 2020

  7. [7]

    Efficient and robust online trajectory prediction for non-cooperative unmanned aerial vehicles,

    Xie, G. and Chen, X., “Efficient and robust online trajectory prediction for non-cooperative unmanned aerial vehicles,”Journal of Aerospace Information Systems, V ol. 19, No. 2, 2022, pp. 143–153

    work page 2022

  8. [8]

    A recurrent neural network approach for aircraft trajectory prediction with weather features from sherlock,

    Pang, Y ., Yao, H., Hu, J., and Liu, Y ., “A recurrent neural network approach for aircraft trajectory prediction with weather features from sherlock,” AIAA Aviation 2019 F orum, 2019, p. 3413

    work page 2019

Show all 31 references
  1. [9]

    Aircraft trajectory prediction made easy with predictive analytics,

    Ayhan, S. and Samet, H., “Aircraft trajectory prediction made easy with predictive analytics,” Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 2016, pp. 21–30

    2016

  2. [10]

    Predicting aircraft trajectories: a deep generative convolutional recurrent neural networks approach,

    Liu, Y . and Hansen, M., “Predicting aircraft trajectories: a deep generative convolutional recurrent neural networks approach,” arXiv preprint arXiv:1812.11670, 2018

    2018 arXiv

  3. [11]

    3-d trajectory modeling for unmanned aerial vehicles,

    Wang, B., Xie, J., Wan, Y ., Guijarro Reyes, G. A., and Garcia Carrillo, L. R., “3-d trajectory modeling for unmanned aerial vehicles,” AIAA Scitech 2019 F orum, 2019, p. 1061

    2019

  4. [12]

    Short-term trajectory prediction methods,

    Chatterji, G., “Short-term trajectory prediction methods,” Guidance, Navigation, and Control Conference and Exhibit , 1999, p. 4233

    1999

  5. [13]

    Airspace encounter models for estimating collision risk,

    Kochenderfer, M. J., Edwards, M. W., Espindle, L. P., Kuchar, J. K., and Griffith, J. D., “Airspace encounter models for estimating collision risk,” Journal of Guidance, Control, and Dynamics , V ol. 33, No. 2, 2010, pp. 487–499

    2010

  6. [14]

    Learning probabilistic trajectory models of aircraft in terminal airspace from position data,

    Barratt, S. T., Kochenderfer, M. J., and Boyd, S. P., “Learning probabilistic trajectory models of aircraft in terminal airspace from position data,” IEEE Transactions on Intelligent Transportation Systems, V ol. 20, No. 9, 2018, pp. 3536–3545

    2018

  7. [15]

    Attention is all you need,

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., and Polosukhin, I., “Attention is all you need,” Advances in neural information processing systems , 2017, pp. 5998–6008

    2017

  8. [16]

    Bert: Pre-training of deep bidirectional transformers for language understanding,

    Devlin, J., Chang, M.-W., Lee, K., and Toutanova, K., “Bert: Pre-training of deep bidirectional transformers for language understanding,” arXiv preprint arXiv:1810.04805, 2018

    2018 arXiv

  9. [17]

    Graph attention networks,

    Veli ˇckovi´c, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., and Bengio, Y ., “Graph attention networks,” arXiv preprint arXiv:1710.10903, 2017

    2017 arXiv

  10. [18]

    Self-attention generative adversarial networks,

    Zhang, H., Goodfellow, I., Metaxas, D., and Odena, A., “Self-attention generative adversarial networks,” International con- ference on machine learning, PMLR, 2019, pp. 7354–7363

    2019

  11. [19]

    Vitgan: Training gans with vision transformers,

    Lee, K., Chang, H., Jiang, L., Zhang, H., Tu, Z., and Liu, C., “Vitgan: Training gans with vision transformers,” arXiv preprint arXiv:2107.04589, 2021

    2021 arXiv

  12. [20]

    Fully convolutional networks for semantic segmentation,

    Long, J., Shelhamer, E., and Darrell, T., “Fully convolutional networks for semantic segmentation,” Proceedings of the IEEE conference on computer vision and pattern recognition , 2015, pp. 3431–3440

    2015

  13. [21]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    Ren, S., He, K., Girshick, R., and Sun, J., “Faster r-cnn: Towards real-time object detection with region proposal networks,” Advances in neural information processing systems , V ol. 28, 2015, pp. 91–99

    2015

  14. [22]

    Deep residual learning for image recognition,

    He, K., Zhang, X., Ren, S., and Sun, J., “Deep residual learning for image recognition,” Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  15. [23]

    Densely connected convolutional networks,

    Huang, G., Liu, Z., Van Der Maaten, L., and Weinberger, K. Q., “Densely connected convolutional networks,” Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 4700–4708

    2017

  16. [24]

    Social gan: Socially acceptable trajectories with generative adversarial networks,

    Gupta, A., Johnson, J., Fei-Fei, L., Savarese, S., and Alahi, A., “Social gan: Socially acceptable trajectories with generative adversarial networks,” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2018, pp. 2255– 2264. 9 of 10 American Institu...

    2018

  17. [25]

    Sophie: An attentive gan for predicting paths compliant to social and physical constraints,

    Sadeghian, A., Kosaraju, V ., Sadeghian, A., Hirose, N., Rezatofighi, H., and Savarese, S., “Sophie: An attentive gan for predicting paths compliant to social and physical constraints,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp...

    2019

  18. [26]

    Conditional generative adversarial networks (CGAN) for aircraft trajectory prediction considering weather effects,

    Pang, Y . and Liu, Y ., “Conditional generative adversarial networks (CGAN) for aircraft trajectory prediction considering weather effects,” AIAA Scitech 2020 F orum, 2020, p. 1853

    2020

  19. [27]

    Learning a probabilistic latent space of object shapes via 3d generative-adversarial modeling,

    Wu, J., Zhang, C., Xue, T., Freeman, W. T., and Tenenbaum, J. B., “Learning a probabilistic latent space of object shapes via 3d generative-adversarial modeling,”Proceedings of the 30th International Conference on Neural Information Processing Systems, 2016, pp. 82–90

    2016

  20. [28]

    Long short-term memory,

    Hochreiter, S. and Schmidhuber, J., “Long short-term memory,” Neural computation, V ol. 9, No. 8, 1997, pp. 1735–1780

    1997

  21. [29]

    Generative adversarial nets,

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y ., “Generative adversarial nets,” Advances in neural information processing systems , V ol. 27, 2014

    2014

  22. [30]

    On learning, representing, and generalizing a task in a humanoid robot,

    Calinon, S., Guenter, F., and Billard, A., “On learning, representing, and generalizing a task in a humanoid robot,” IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics) , V ol. 37, No. 2, 2007, pp. 286–298

    2007

  23. [31]

    gmr: Gaussian Mixture Regression,

    Fabisch, A., “gmr: Gaussian Mixture Regression,” Journal of Open Source Software, V ol. 6, No. 62, 2021, pp. 3054. 10 of 10 American Institute of Aeronautics and Astronautics

    2021

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.