Pith. sign in

REVIEW 4 major objections 7 minor 28 references

DKMGP: A Gaussian Process Approach to Multi-Task and Multi-Step Vehicle Dynamics Modeling in Autonomous Racing

T0 review · 4 major / 7 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 single multi-task Gaussian process can replace three separate GP models for race-car state correction, running about 1750 times faster while staying within a few percent of the one-step baseline's accuracy.

desk verdict Useful, measured speedup for GP-based dynamics correction, but the multi-step accuracy claim rests on an unstated Markov assumption and a weak baseline comparison. read the letter →

arxiv 2411.13755 v1 pith:HPYU47L4 submitted 2024-11-20 cs.RO

Jingyun Ning , Madhur Behl This is my paper · ORCID
classification cs.RO
keywords autonomousracingGaussianprocessdeepkernellearningmulti-taskmulti-steppredictionvehicledynamicsadaptivecorrectionhorizonIndyChallenge
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 single multi-task deep-kernel Gaussian process (DKMGP) can replace three separate single-step GP models for predicting race-car state residuals, making GP-based dynamics correction fast enough for real-time closed-loop control in autonomous racing. Trained on data from a full-scale Indy car at speeds over 230 km/h, the model predicts the residuals of longitudinal velocity, lateral velocity, and yaw rate in one forward pass, with an adaptive correction horizon that lengthens in mild driving conditions and shortens in aggressive ones. The reported result is up to 99% prediction accuracy compared to the one-step DKL-SKIP baseline, while running roughly 500 to 1750 times faster, depending on the chosen horizon. If correct, this removes the latency barrier that previously kept GP corrections out of high-speed model-based control loops.

What carries the argument

The load-bearing object is the DKMGP model itself: a deep kernel learning feature extractor (a two-layer DNN that reduces inputs from dimension 9 to dimension 5) feeding a multi-task variational Gaussian process built on a linear model of coregionalization. The LMC couples the three residual tasks through a small number of shared latent functions, and variational inference keeps the computational cost tractable. The second piece is the adaptive correction horizon (ACH) algorithm, a lookup rule that classifies driving conditions as Cruising, Controlled, Pushing, or Aggressive and sets the correction horizon to 15, 10, 5, or 3 steps respectively. Together these let one model produce corrected predictions over a 43-step horizon at inference rates of 22 to 75 Hz on the paper's hardware.

What would settle it

Take a test lap where the car transitions between regimes mid-horizon (for example, hard braking after a straight) and compare DKMGP's predicted residual e_{t+n} against the true residual when intermediate states are available. If the prediction error grows systematically in such transition segments, the Markov assumption on which the multi-step training is based is violated.

Watch

Extended reading notes

Core claim

The central claim is that the error between a simplified kinematic model and the true dynamics of a race car can be corrected by one multi-task Gaussian process with deep kernel learning, rather than by three independent models. The paper builds on the observation that the residuals of the base states $v_x, v_y, \omega$ are physically correlated, so a linear model of coregionalization can share latent structure across tasks. A deep neural network maps the 9-dimensional state-input vector into a 5-dimensional feature space, and a variational multi-task GP predicts all three residuals at once. To support multi-step prediction, the paper trains the model to directly output the residual after $n-1$ uncorrected propagations of the E-kin model, with the horizon $n$ chosen by a rule-based adaptive algorithm based on speed, acceleration, and steering angle. The authors report prediction accuracy comparable to the single-step DKL-SKIP baseline, with a 1752-fold speedup in inference over a 43-step horizon.

Load-bearing premise

The load-bearing premise is that the residual of the base states after n-1 uncorrected model propagations is determined by the initial state and input sequence alone, so the model can be trained to predict e_{t+n} without seeing intermediate states.

Editorial extensions

If this is right

  • Closed-loop model predictive control could use DKMGP's corrected predictions at controller rates, since the inference throughput of 22 to 75 Hz matches typical MPC update frequencies in racing.
  • A single correction model eliminates the need to train and calibrate three separate GP models, shrinking the deployment effort for learning-based vehicle dynamics.
  • The adaptive correction horizon provides a direct trade-off knob: longer horizons in cruising conditions save computation, while shorter horizons in aggressive driving preserve accuracy where the dynamics are hardest.
  • The GP's uncertainty estimates, already part of the variational framework, could feed into risk-aware planning even though the current paper does not exercise them in closed loop.

Reading between the lines

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

  • The same architecture could be applied to other multi-output dynamical systems where a cheap physics model leaves correlated residuals, such as aircraft or marine vehicles, not just race cars.
  • The rule-based ACH thresholds could be learned from data, for example by treating the horizon as a decision variable that minimizes a cost combining prediction error and computation.
  • The paper's accuracy claims rest on a single-lap comparison with DKL-SKIP; a direct test over the full multi-lap test set would clarify whether the 99% accuracy figure holds consistently.
  • The Markov assumption, if relaxed by feeding predicted intermediate states back into the feature extractor, would turn DKMGP into a recurrent model and likely improve accuracy at long horizons, at some computational cost.
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 / 7 minor

Summary. The paper proposes DKMGP, a deep-kernel multi-task variational Gaussian process that learns residual corrections for an extended kinematic (E-kin) vehicle model, predicting errors in longitudinal velocity, lateral velocity, and yaw rate with a single model. A rule-based adaptive correction horizon (ACH) algorithm chooses how many E-kin propagation steps to take between corrections. The method is evaluated on high-speed data from a full-scale Indy Autonomous Challenge car, comparing inference rate and prediction accuracy against the earlier DKL-SKIP approach and a nonlinear single-track model. The paper claims up to 99% of DKL-SKIP's accuracy while being 522-1752x faster, with better or comparable performance than a well-tuned single-track model.

Significance. If the claims hold, the paper makes a practically useful contribution: replacing three separate single-step GP models with one multi-task model, enabling real-time multi-step residual correction at inference rates above 20 Hz, and validating on real full-scale autonomous racing data at speeds exceeding 230 km/h. The inference-rate results in Table 3 are direct measurements and provide credible support for the efficiency claim. The multi-step formulation and the ACH algorithm are also novel relative to the prior DKL-SKIP work. However, the central accuracy claim is currently under-supported: the evaluation is not matched between methods, no uncertainty quantification is reported, and the multi-step training target is not identifiable from the stated model input unless a Markov/policy assumption is made. The accuracy advantage of DKL-SKIP on lateral velocity and yaw rate is also understated in the abstract and text.

major comments (4)
  1. [Section 4, Eq. (3), Figure 3] The multi-step training target is not identifiable from the stated model input. The DKMGP is trained to predict e_{t+n}, the residual after n-1 uncorrected E-kin propagations using the full recorded input sequence {u_t,...,u_{t+n-1}}, but the model input is only the initial pair (s_t, u_t). Unless the future input sequence is a deterministic function of s_t and u_t (for example, a known closed-loop control policy), no function of the stated inputs can determine e_{t+n}; the GP can at best fit the average residual over input trajectories. The paper neither states nor validates such a Markov/policy assumption. This issue is load-bearing for the multi-step prediction contribution, because the inference loop also feeds back corrected predicted states as initial states, while the training procedure never generates corrected states as starting points. I would ask the authors to state the assumption explicitly, verify it on the recorded data (for example, by testing whether future inputs are predictable from the initial state), and compare against a model that receives the planned input sequence as additional input features.
  2. [Section 5.3, Table 4] The accuracy comparison between DKMGP and DKL-SKIP is not matched. The text states that DKL-SKIP validation was restricted to a single lap because its inference rate was too slow, but it does not state whether the DKMGP row in Table 4 was computed on the same lap, on the full test set, or with the ACH algorithm or a fixed horizon. No error bars or repeated evaluations are reported, so the relative differences in Table 4 cannot be interpreted. Moreover, the abstract's claim of "up to 99% prediction accuracy compared to one-step DKL-SKIP" is not supported by Table 4: DKMGP MAE is about 3% higher for vx, 77% higher for vy, and 94% higher for omega (0.1351 vs 0.1311, 0.0943 vs 0.0533, and 0.0424 vs 0.0218). Calling this a "minor accuracy gain" understates the gap on the lateral states. The authors should report matched evaluation on identical laps, specify the DKMGP configuration, and provide uncertainty or variance estimates.
  3. [Section 4, Algorithm 1, Figure 5] The ACH algorithm is described only qualitatively. The thresholds in Algorithm 1 and the horizon values 3, 5, 10, and 15 are hand-set on the authors' data, with no sensitivity analysis or evidence that they generalize to other tracks, speeds, or vehicle configurations. Figure 5 is a qualitative plot and is said to show that "ACH consistently achieves the smallest CTE," but no aggregate cross-track error values, error bars, or statistical comparison against the fixed-horizon baselines are provided. Since ACH is listed as a main contribution, it needs quantitative validation, including a statement of which DKMGP configuration is used for the accuracy comparison in Table 4.
  4. [Section 5.4, Figure 6] The comparison with the nonlinear single-track model is reported through sectional average cross-track error over a single plot, without quantitative summary values, error bars, or details on how many laps and which track sections are included. The claim that DKMGP "can outperform the single-track model at low to moderate speeds" is therefore not quantitatively established. This is a secondary claim, but it should be backed by numbers.
minor comments (7)
  1. [Abstract and Figure 3 caption] There are typos in the abstract ("upto") and in Figure 3 ("DKGMP" instead of "DKMGP"); Section 5.4 also has "Singe-track" in the heading.
  2. [Section 2.1] The text describing lateral dynamics says "yaw rate (ψ)" twice; the first occurrence should presumably be "heading angle (ψ)" since yaw rate is denoted by ω.
  3. [Section 2.3, Eq. (1)] Equation (1) has an argument mismatch: the left side is f_corr(s_{t+1}, u_{t+1}) while the right side uses f_Ekin(s_t, u_t) plus e_{t+1}; the notation should be made consistent.
  4. [Section 3.2, Eqs. (5) and covariance expression] The variational objective is written as minimizing KL(p(epsilon_tau|f) || q(f)), which is backwards; the ELBO in Eq. (5) is the correct expression. The covariance sum "sum_{q=1}^{j}" appears to have the wrong upper limit and should be over Q latent functions.
  5. [Table 4] The label "E-kin ST" is ambiguous; the text refers to this row as the uncorrected E-kin model, so the label should be clarified to avoid confusion with the single-track model.
  6. [Section 5.3, Table 3] The paper should define precisely what one "inference" event is: whether the reported Hz values correspond to completing one full 43-step horizon, one correction step, or one model call. This is needed for reproducibility and for interpreting the 1752x speedup.
  7. [Section 3.1 and Table 2] Section 3.1 says the DKL maps each input from R^9 to R^4, but Table 2 lists the optimal "Number of Features" as 5; this inconsistency should be resolved.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the correction targets are measured residuals on held-out laps, and ACH is a hand-set scheduling rule.

full rationale

The central prediction chain is not circular. In Section 4, the multi-step residual e_{t+n} is constructed as the difference between a real measured state s_{t+n} and n-step E-kin propagation from recorded inputs, and DKMGP is trained to regress this measured residual from the initial state/input. That is a standard supervised correction model; the test evaluation is on a disjoint set of 17 laps, so the reported accuracy is not the training objective by construction. The ACH rule in Algorithm 1 is a hand-specified mapping from velocity, acceleration, and steering thresholds to fixed correction horizons; it is parameter selection on the authors' operating regime, not a fitted quantity renamed as a prediction, and it is not defined in terms of DKMGP's own output. Self-citations (Ning and Behl 2023a, 2023b, 2024) motivate DKL and supply the DKL-SKIP baseline, but the DKL machinery is also attributed to the external Wilson et al. (2016) reference and the empirical comparison is against held-out Indy Autonomous Challenge data, so the self-citations are not load-bearing. The skeptic's concern that e_{t+n} depends on the full future input sequence while the model sees only s_t is a real identifiability/Markov-assumption risk and a correctness threat, but it is not a definitional equivalence: the paper nowhere defines s_t in terms of e_{t+n} or claims the GP output is e_{t+n} by construction. Under the requested circularity standard, there is no quoted reduction of the claimed result to its inputs.

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

The central claims rest on standard GP/VI machinery, a Markovian residual assumption, the accuracy of EKF state estimates, the representativeness of a single race event, and hand-set ACH thresholds. No new physical entities are introduced.

free parameters (5)
  • ACH speed thresholds = 40, 50, 60 m/s
    Algorithm 1 partitions vx into Cruising/Controlled/Pushing/Aggressive; boundaries are chosen by hand, no tuning procedure shown.
  • ACH acceleration thresholds = 0.5, 1.0, 3.0 m/s^2
    Algorithm 1 uses these |ax| boundaries to classify driving condition; they are hand-set.
  • ACH steering thresholds = 4.5, 7.5, 11.5 degrees
    Algorithm 1 uses these |delta_w| boundaries; no justification or sensitivity analysis is given.
  • ACH correction horizon values = 15, 10, 5, 3 steps
    Mapped to the four driving conditions in Algorithm 1; these integers balance accuracy and inference rate but are set by the authors without sensitivity analysis.
  • DKMGP hyperparameters = batch 144, epochs 1140, RBF kernel, hidden 256/64, lr 0.0064, features 5
    Selected via Optuna grid search on training data (Section 5.2, Table 2); they affect the reported accuracy and speed.
assumptions (4)
  • domain assumption Residual et+n depends only on the initial measured state st
    Section 4 Training Phase: input features are st; this Markov assumption is not stated explicitly and not validated.
  • domain assumption EKF state estimates at 100 Hz are accurate ground truth
    Section 5.1: residuals and metrics are computed against EKF estimates; any EKF bias transfers to the learned corrections.
  • domain assumption The LVMS/CES 2024 dataset represents racing conditions for which the ACH rules will be used
    Section 5.1: data from one event; no cross-track generalization test.
  • standard math Variational GP and deep kernel learning machinery is correct
    Sections 3.1-3.2 invoke standard results (Rasmussen, Alvarez, Wilson) without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DKMGP: A Gaussian Process Approach to Multi-Task and Multi-Step Vehicle Dynamics Modeling in Autonomous Racing." pith.science (2026). https://pith.science/paper/HPYU47L4

@misc{pith2026241113755,
  author       = {Pith},
  title        = {Pith review of: DKMGP: A Gaussian Process Approach to Multi-Task and Multi-Step Vehicle Dynamics Modeling in Autonomous Racing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HPYU47L4}},
  note         = {Machine review of arXiv:2411.13755}
}
read the original abstract

Autonomous racing is gaining attention for its potential to advance autonomous vehicle technologies. Accurate race car dynamics modeling is essential for capturing and predicting future states like position, orientation, and velocity. However, accurately modeling complex subsystems such as tires and suspension poses significant challenges. In this paper, we introduce the Deep Kernel-based Multi-task Gaussian Process (DKMGP), which leverages the structure of a variational multi-task and multi-step Gaussian process model enhanced with deep kernel learning for vehicle dynamics modeling. Unlike existing single-step methods, DKMGP performs multi-step corrections with an adaptive correction horizon (ACH) algorithm that dynamically adjusts to varying driving conditions. To validate and evaluate the proposed DKMGP method, we compare the model performance with DKL-SKIP and a well-tuned single-track model, using high-speed dynamics data (exceeding 230kmph) collected from a full-scale Indy race car during the Indy Autonomous Challenge held at the Las Vegas Motor Speedway at CES 2024. The results demonstrate that DKMGP achieves upto 99% prediction accuracy compared to one-step DKL-SKIP, while improving real-time computational efficiency by 1752x. Our results show that DKMGP is a scalable and efficient solution for vehicle dynamics modeling making it suitable for high-speed autonomous racing control.

Figures

Figures reproduced from arXiv: 2411.13755 by the authors.

Figure 1
Figure 1. The single-track vehicle model of the racecar, with all three base states (vx.vy, ω) highlighted [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The architecture of the DKMGP model. Vehicle States & Inputs DKMGP corrected Prediction n-Step DKGMP E-Kin Prediction n steps E-Kin {s Prediction t, ut} et+n={ϵvx , ϵvy , ϵω} m-Step Prediction Horizon {st+1 , ut+1} {st+n , ut+n} ^ ^ {st+n+et+n} ^ [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. (A) Data from overtaking competition, LVMS CES 2024. (B) LVMS track layout. (C) LiDAR point cloud data. (D) AV-21 sensor setup. This section outlines the experimental setup and validation results for the DKMGP-integrated E￾kin model, focusing on its prediction accuracy and real-time performance. 5.1. Real-world Data Collection We evaluate the performance of the DKMGP model using real-world autonomous racing data col… view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Comparison of DKMGP models across different fixed correction horizons. DKMGP prediction accuracy We first compare the ACH-based DKMGP model to DKMGP mod￾els with fixed-length correction horizons [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Comparison of sectional cross-track errors between the single-track model and the DKMGP (E-kin) model. The E-kin model consistently achieves lower sCTE below 40 m s−1 , showing better prediction performance over the single-track model under these conditions. This advan…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 21 canonical work pages

  1. [1]

    Optuna: A next-generation hyperparameter optimization framework

    Takuya Akiba, Shotaro Sano, Toshihiko Yanase, Takeru Ohta, and Masanori Koyama. Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pages 2623--2631, 2019

    2019

  2. [2]

    Commonroad: Composable benchmarks for motion planning on roads

    Matthias Althoff, Markus Koschi, and Stefanie Manzinger. Commonroad: Composable benchmarks for motion planning on roads. In 2017 IEEE Intelligent Vehicles Symposium (IV), pages 719--726. IEEE, 2017

    work page 2017

  3. [3]

    Kernels for vector-valued functions: A review

    Mauricio A Alvarez, Lorenzo Rosasco, Neil D Lawrence, et al. Kernels for vector-valued functions: A review. Foundations and Trends in Machine Learning , 4 0 (3): 0 195--266, 2012

    2012

  4. [4]

    Varundev Suresh Babu and Madhur Behl. f1tenth. dev-an open-source ros based f1/10 autonomous racing simulator. In 2020 IEEE 16th International Conference on Automation Science and Engineering (CASE), pages 1614--1620. IEEE, 2020

    work page 2020

  5. [5]

    Autonomous vehicles on the edge: A survey on autonomous vehicle racing

    Johannes Betz, Hongrui Zheng, Alexander Liniger, Ugo Rosolia, Phillip Karle, Madhur Behl, Venkat Krovi, and Rahul Mangharam. Autonomous vehicles on the edge: A survey on autonomous vehicle racing. IEEE Open Journal of Intelligent Transportation Systems, 3: 0 458--488, 2022

    work page 2022

  6. [6]

    Experiment tracking with weights and biases, 2020

    Lukas Biewald. Experiment tracking with weights and biases, 2020. URL https://www.wandb.com/. Software available from wandb.com

    2020

  7. [7]

    Efficient implementation of randomized mpc for miniature race cars

    Jes \'u s Velasco Carrau, Alexander Liniger, Xiaojing Zhang, and John Lygeros. Efficient implementation of randomized mpc for miniature race cars. In 2016 European Control Conference (ECC), pages 957--962. IEEE, 2016

    work page 2016

  8. [8]

    Deep dynamics: Vehicle dynamics modeling with a physics-constrained neural network for autonomous racing

    John Chrosniak, Jingyun Ning, and Madhur Behl. Deep dynamics: Vehicle dynamics modeling with a physics-constrained neural network for autonomous racing. IEEE Robotics and Automation Letters, 2024

    work page 2024

Show all 28 references
  1. [9]

    Modelling longitudinal vehicle dynamics with neural networks

    Mauro Da Lio, Daniele Bortoluzzi, and Gastone Pietro Rosati Papini. Modelling longitudinal vehicle dynamics with neural networks. Vehicle System Dynamics, 58 0 (11): 0 1675--1693, 2020

    2020

  2. [10]

    Autonomous head-to-head racing in the indy autonomous challenge simulation race

    Gabriel Hartmann, Zvi Shiller, and Amos Azaria. Autonomous head-to-head racing in the indy autonomous challenge simulation race. arXiv preprint arXiv:2109.05455, 2021

    2021 arXiv

  3. [11]

    End-to-end neural network for vehicle dynamics modeling

    Leonhard Hermansdorfer, Rainer Trauth, Johannes Betz, and Markus Lienkamp. End-to-end neural network for vehicle dynamics modeling. In 2020 6th IEEE Congress on Information Science and Technology (CiSt), pages 407--412, 2020. doi:10.1109/CiSt49399.2021.9357196

    2020

  4. [12]

    Bayesrace: Learning to race autonomously using prior experience

    Achin Jain, Matthew O'Kelly, Pratik Chaudhari, and Manfred Morari. Bayesrace: Learning to race autonomously using prior experience. In Conference on Robot Learning, pages 1918--1929. PMLR, 2021

    1918

  5. [13]

    Learning-based model predictive control for autonomous racing

    Juraj Kabzan, Lukas Hewing, Alexander Liniger, and Melanie N Zeilinger. Learning-based model predictive control for autonomous racing. IEEE Robotics and Automation Letters, 4 0 (4): 0 3363--3370, 2019

    2019

  6. [14]

    Racecar-the dataset for high-speed autonomous racing

    Amar Kulkarni, John Chrosniak, Emory Ducote, Florian Sauerbeck, Andrew Saba, Utkarsh Chirimar, John Link, Madhur Behl, and Marcello Cellina. Racecar-the dataset for high-speed autonomous racing. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)...

    2023

  7. [15]

    Optimization-based autonomous racing of 1: 43 scale rc cars

    Alexander Liniger, Alexander Domahidi, and Manfred Morari. Optimization-based autonomous racing of 1: 43 scale rc cars. Optimal Control Applications and Methods, 36 0 (5): 0 628--647, 2015

    2015

  8. [16]

    Scalable deep kernel gaussian process for vehicle dynamics in autonomous racing

    Jingyun Ning and Madhur Behl. Scalable deep kernel gaussian process for vehicle dynamics in autonomous racing. In Conference on Robot Learning, pages 761--773. PMLR, 2023 a

    2023

  9. [17]

    Vehicle dynamics modeling for autonomous racing using gaussian processes

    Jingyun Ning and Madhur Behl. Vehicle dynamics modeling for autonomous racing using gaussian processes. arXiv preprint arXiv:2306.03405, 2023 b

    2023 arXiv

  10. [18]

    Gaussian processes for vehicle dynamics learning in autonomous racing

    Jingyun Ning and Madhur Behl. Gaussian processes for vehicle dynamics learning in autonomous racing. SAE International Journal of Vehicle Dynamics, Stability, and NVH, 8 0 (10-08-03-0019), 2024

    2024

  11. [19]

    F1/10: An open-source autonomous cyber-physical platform

    Matthew O'Kelly, Varundev Sukhil, Houssam Abbas, Jack Harkins, Chris Kao, Yash Vardhan Pant, Rahul Mangharam, Dipshil Agarwal, Madhur Behl, Paolo Burgio, et al. F1/10: An open-source autonomous cyber-physical platform. arXiv preprint arXiv:1901.08567, 2019

    1901 arXiv

  12. [20]

    Tire and vehicle dynamics

    Hans Pacejka. Tire and vehicle dynamics. Elsevier, 2005

    2005

  13. [21]

    Data-driven vehicle modeling of longitudinal dynamics based on a multibody model and deep neural networks

    Yongjun Pan, Xiaobo Nie, Zhixiong Li, and Shuitao Gu. Data-driven vehicle modeling of longitudinal dynamics based on a multibody model and deep neural networks. Measurement, 180: 0 109541, 2021

    2021

  14. [22]

    Motion planning and control for multi vehicle autonomous racing at high speeds

    Ayoub Raji, Alexander Liniger, Andrea Giove, Alessandro Toschi, Nicola Musiu, Daniele Morra, Micaela Verucchi, Danilo Caporale, and Marko Bertogna. Motion planning and control for multi vehicle autonomous racing at high speeds. In 2022 IEEE 25th International Conference on Int...

    2022

  15. [23]

    Gaussian processes in machine learning

    Carl Edward Rasmussen. Gaussian processes in machine learning. In Summer school on machine learning, pages 63--71. Springer, 2003

    2003

  16. [24]

    Online constrained model-based reinforcement learning

    Benjamin Van Niekerk, Andreas Damianou, and Benjamin Rosman. Online constrained model-based reinforcement learning. Conf. Uncertainty in Artificial Intelligence, 2017

    2017

  17. [25]

    Graphical models, exponential families, and variational inference

    Martin J Wainwright, Michael I Jordan, et al. Graphical models, exponential families, and variational inference. Foundations and Trends in Machine Learning , 1 0 (1--2): 0 1--305, 2008

    2008

  18. [26]

    Deepracing: Parameterized trajectories for autonomous racing

    Trent Weiss and Madhur Behl. Deepracing: Parameterized trajectories for autonomous racing. arXiv preprint arXiv:2005.05178, 2020

    2005 arXiv

  19. [27]

    Deep kernel learning

    Andrew Gordon Wilson, Zhiting Hu, Ruslan Salakhutdinov, and Eric P Xing. Deep kernel learning. In Artificial intelligence and statistics, pages 370--378. PMLR, 2016

    2016

  20. [28]

    Indy autonomous challenge-autonomous race cars at the handling limits

    Alexander Wischnewski, Maximilian Geisslinger, Johannes Betz, Tobias Betz, Felix Fent, Alexander Heilmeier, Leonhard Hermansdorfer, Thomas Herrmann, Sebastian Huch, Phillip Karle, et al. Indy autonomous challenge-autonomous race cars at the handling limits. In 12th Internation...

    2021

Pith tools

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