Safe Reinforcement Learning with a Predictive Safety Filter for Motion Planning and Control: A Drifting Vehicle Example

Baha Zarrouki; Bei Zhou; Cheng Hu; Johannes Betz; Lei Xie; Mattia Piccinini

REVIEW 2 major objections 4 minor 33 references

Safe Reinforcement Learning with a Predictive Safety Filter for Motion Planning and Control: A Drifting Vehicle Example

T0 review · 2 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A predictive safety filter lets a reinforcement-learned planner drift safely and beat an MPC benchmark.

desk verdict A useful RL+PSF drifting planner with real performance gains, but the 'Safe' claim outruns the evidence: the filter's frozen-state model does not enforce the stated constraints, and Fig. 5a shows a violation. read the letter →

arxiv 2506.22894 v1 pith:5Z7HMYXQ submitted 2025-06-28 cs.RO

Bei Zhou , Baha Zarrouki , Mattia Piccinini , Cheng Hu , Lei Xie , Johannes Betz This is my paper · ORCID

classification cs.RO

keywords safereinforcementlearningpredictivesafetyfilterautonomousdriftingmotionplanningdriftequilibriumroadfrictionpathtrackingDDPG

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 establish that a reinforcement-learning motion planner can plan autonomous drifting along a general path without expert demonstrations, provided a predictive safety filter screens its curvature commands. The planner learns two quantities online: a correction to the reference curvature and a correction to the road friction coefficient, which improve the drift-equilibrium calculation of a model predictive drift controller. If correct, the approach offers a learning-based alternative to MPC-based drift planning that is safer during training and faster online. The claim is supported by simulations where the safe RL planner reduces mean lateral deviation by about half and mean heading error by about 60 percent relative to the benchmark, while using 3.6 times less computation.

What carries the argument

The central object is the Predictive Safety Filter (PSF), an optimal control problem that minimizes the deviation between the RL-planned curvature and the filtered curvature subject to the path-tracking model $\dot e = V\sin\Delta\psi$ and $\Delta\dot\psi = \dot\beta + r - \kappa V\cos\Delta\psi/(1-\kappa e)$, with constraints $|e|<e_{\mathrm{th}}$ and $|\Delta\psi|<\Delta\psi_{\mathrm{th}}$ enforced over a short prediction horizon. The RL agent, trained with DDPG, simultaneously learns curvature adjustment $\kappa_{\mathrm{RL}} = \kappa_r + \varepsilon_\kappa$ and friction adjustment $\mu_{\mathrm{RL}} = \mu_n + \varepsilon_\mu$; these corrections refine the drift-equilibrium calculation that the low-level Model Predictive Drift Controller tracks. The PSF sits between the RL agent and the controller, rejecting or modifying curvature commands that would violate the safety constraints.

What would settle it

Run the trained Safe RL planner on a high-fidelity or physical vehicle through a transient drift where yaw rate and sideslip change rapidly, and record whether the actual lateral error exceeds $e_{\mathrm{th}} = 1.5$ m even though the PSF's filtered curvature satisfied $|e|<e_{\mathrm{th}}$ in its prediction model. One such violation during normal PSF operation would refute the claim that the filter prevents unsafe states.

Watch

Extended reading notes

Core claim

On the paper's own terms, the proposed Safe RL planner couples a DDPG agent that outputs reference-curvature and road-friction corrections with a Predictive Safety Filter that re-solves a short-horizon optimal control problem at every step to keep lateral and yaw errors within track bounds, and then feeds the filtered curvature to a model predictive drift controller. In Matlab-Carsim simulations over an 18-second clothoid path, this combination produced 49.6% lower mean lateral deviation and 61.1% lower mean heading error than an enhanced MPC planner benchmark, while requiring 3.6 times less average computation time. The same architecture also generalizes to three unseen tracks with varying curvature, keeping lower tracking RMSE than the fixed-parameter MPC benchmark on every test track.

Load-bearing premise

The safety filter checks the RL curvature using a simplified forecast that freezes the vehicle's speed, sideslip, and yaw rate, so its "safe" output may not be safe for the real vehicle during aggressive drifting.

Editorial extensions

If this is right

A learning-based drift planner can run online in real time on modest hardware, whereas the MPC benchmark's average 0.495 s solve time is far too slow for the 0.05 s control loop.
The predictive safety filter accelerates training convergence and raises average reward, so safety filtering acts as a training-time regularizer as well as a deployment safeguard.
Learning the road friction coefficient online reduces drift-state RMSE by 11.5% relative to a curvature-only RL planner, showing that model-error correction matters for drift performance.
The safe RL planner retains its tracking advantage over the MPC benchmark on unseen test tracks without retuning, supporting the paper's claim of generalization to variable-curvature environments.

Reading between the lines

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

The safety guarantee is only as strong as the PSF's frozen-parameter prediction model: if velocity, sideslip, and yaw rate change substantially within the prediction horizon during aggressive transients, the filtered curvature could violate the true constraints even though the simplified model reports a safe trajectory.
The same PSF-plus-RL architecture could transfer to other underactuated, safety-critical maneuvers such as emergency obstacle avoidance on low-friction roads, where the RL agent adapts to unknown parameters and the filter rejects unsafe reference commands.
A testable extension is to replace the frozen-parameter PSF model with a reachable-set or robust-tube forecast; the resulting losses in tracking performance and gains in safety could be measured directly against the present simulation results.
Because PSF corrections diminish as training progresses, the filter could be omitted at inference time to save computation, but doing so would sacrifice the explicit safety layer that the paper highlights as its contribution.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: the integration of RL-based online curvature and friction adaptation with a predictive safety filter is a real and sensible step for autonomous drifting, and the simulations support the performance claim against the MPC baseline. But the word 'Safe' is load-bearing, and the filter's guarantee is not actually shown: model (17) freezes V, beta, and r, and Fig. 5a shows an early termination at |e|>emax with the filter active. So the safety contribution is a heuristic that reduces errors, not a guarantee.

What is genuinely new: I don't recall another drifting planner that uses RL to adjust both the reference curvature and the friction coefficient in closed loop with a safety filter. The ablation separating curvature learning from friction learning is good, and the head-to-head against the enhanced MPC from [15] is fair. The 3.6x computation advantage is credible, and the training curves show the PSF does help convergence.

Soft spots: the central one is the mismatch between the 'Safe RL' language and what the PSF actually does. Footnote 1 admits V, beta, and r are held constant over the 10-step horizon. During hard drifting those states change fast, so the OCP's constraints |e|<1.5 m and |Δψ|<0.2 rad do not transfer to the real vehicle. Fig. 5a is not just a theoretical worry: the filter is active and the episode still terminates because |e| exceeds 5 m. That means the claim to 'prevent unsafe states' is overstated. The right fix is to soften the guarantee language or add an analysis of when constraint satisfaction on (17) implies safety on (1)-(3), e.g., Lipschitz bounds on the drift states over the horizon.

Additional minor issues: single training seed, no error bars on the box plots, and the generalization test with only three tracks is a start but not convincing. No code or data is available, which lowers confidence.

Bottom line: a useful engineering contribution that deserves a serious referee, but not with the current safety claims. I'd send it to review with major revisions expected: soften the claims, add formal or empirical evidence about the filter's effect on true constraints, and ideally release code/data. I wouldn't cite it as a safe RL method in my own work, but I might cite the empirical comparison.

Referee Report

2 major / 4 minor

Summary. The paper proposes a hierarchical Safe Reinforcement Learning (RL) motion planner for autonomous drifting. An RL agent (DDPG) learns a curvature offset and a friction coefficient correction to feed a Model Predictive Drift Controller (MPDC). A Predictive Safety Filter (PSF) solves a short-horizon optimal control problem to adjust the RL curvature reference so that a simplified path-tracking model satisfies |e|<eth and |Δψ|<Δψth. The authors validate in Matlab-Carsim simulations, reporting that the safe RL planner outperforms an MPC benchmark in tracking accuracy, drift-state RMSE, and computation time, and that the PSF accelerates training convergence. The paper also presents ablations isolating the effects of learning κ, learning μ, and adding the PSF, plus generalization tests on three unseen tracks.

Significance. If the safety and performance claims hold, this is a useful contribution to learning-based motion planning for highly dynamic maneuvers: it combines model-free RL with a modular model-based safety filter, avoiding expert demonstrations while keeping a principled controller. The paper is careful in its ablation design: the low-level MPDC is unchanged across all comparisons, and the reported improvements are consistent across lateral error, heading error, drift-state RMSE, computation time, and generalization tracks. The computational-efficiency comparison is a notable strength, as is the explicit training-time cost of the PSF. However, the central 'safe' claim is currently stated too strongly relative to the evidence, so the significance hinges on whether the safety guarantee can be substantiated for the full closed-loop system.

major comments (2)

[Section V, Eq. (17), footnote 1, Fig. 5a] The safety guarantee of the PSF is not established. The prediction model (17) keeps V, beta, and r constant over the horizon, and the constraints |e|<eth and |Δψ|<Δψth are enforced on this simplified kinematic model, not on the full drift model (1)-(3) or the Carsim vehicle. Since the low-level MPDC tracks a drift equilibrium rather than directly imposing path-tracking constraints, the realized closed-loop errors can differ substantially from the PSF predictions. This is not merely a theoretical gap: Fig. 5a shows an Episode-500 run where the PSF is active (κ0 deviates from κRL) yet the vehicle still triggers early termination at |e|>emax=5 m, well above eth=1.5 m. Thus the statements that the PSF 'prevents unsafe states' (contribution 2) and that the method 'ensures safe and efficient learning' (abstract) are unsupported. The authors should either provide a formal safety argument for the full closed-loop system (e.g., robust invariance or tube-based reasoning) or present a thorough empirical safety evaluation over all training episodes and test tracks, and revise the wording to 'reduces safety violations' rather than 'ensures safety'.
[Section V, Eq. (16)] The PSF OCP formulation does not discuss recursive feasibility or the behavior when the OCP is infeasible. The RL agent can propose a curvature for which no solution of (16) satisfies the constraints; the paper does not specify what action is executed in that case, how the RL update handles the lack of a safe action, or whether the MPDC can still receive a feasible reference. Since the PSF is the core safety mechanism, the absence of a feasibility analysis leaves a gap in the safety claim and in the practical implementation of the training loop.

minor comments (4)

[Section VI-H, Fig. 3c] The computational load in Fig. 3c is reported without clarifying whether it includes the MPDC solve time. Given that the RL planner operates at a 0.1 s sampling time and the reported Safe RL mean load is 0.137 s, the reader cannot determine whether the full system meets real-time constraints; please state explicitly what the box-plot values include and add the aggregate per-step latency.
[Section I-B, contribution 3] The claim of 'no prior expert knowledge' should be qualified: the equilibrium calculation in Section III-B fixes δ0 = -0.5 rad and uses a model-based drift dynamics, which constitutes model knowledge. Clarify that the claim refers to the absence of expert demonstrations or hand-tuned initial policies, not to the absence of any model-based prior.
[Section IV-B-3] In the reward description, the phrasing 'penalizes further the lateral error e further' is redundant; consider rewording to 'additionally penalizes the lateral error e once it exceeds the threshold'.
[Section VI-D] When reporting percentage improvements over the MPC benchmark, please include the absolute values of the errors in the same sentence (they appear only in Fig. 3), so the reader can assess the practical magnitude of the gains.

Circularity Check

0 steps flagged · score 2.0 of 10

No material circularity: the planner's advantage over the MPC benchmark is empirical and the benchmark is reproduced; minor self-citations are not load-bearing.

full rationale

The paper's central claims are empirical rather than derived from their inputs by construction. The RL agent's reward (13)-(15) minimizes lateral and heading errors, and the evaluation (Figs. 3-4) reports those same errors; this is objective alignment during training, not a hidden fit or a prediction made from fitted parameters. The learned friction correction mu_RL is a latent action tuned by DDPG to improve closed-loop reward, not a parameter fitted to the reported outcome. The PSF in (16) minimizes the distance to the RL curvature subject to constraints, so its output is not forced to equal any 'predicted' result. The MPC benchmark from [15] is a prior published controller that the authors re-implement and enhance; although [15] shares authors with this paper, the comparison is reproduced in simulation rather than imported as an unverified premise, so the self-citation is not load-bearing. The paper's main weakness is the PSF prediction model (17) freezing V, beta, and r over the horizon (footnote 1) and the early termination shown in Fig. 5a despite the filter being active; this undermines the strength of the safety guarantee, but it is a soundness and assumption gap, not a circular derivation. Therefore no circular step is identified, and the score reflects only minor non-load-bearing self-citation.

Assumptions & free parameters 9 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several hand-chosen thresholds and weights that are not automatically tuned. The safety guarantee depends on a simplified prediction model and on the assumption that the vehicle remains near drift equilibria. No new physical entities are introduced.

free parameters (9)

PSF lateral error threshold eth = 1.5 m
Chosen by hand in Table I; directly sets the safety constraint in eq. 16c and the reward penalty in eq. 14. Affects how aggressively the PSF modifies the RL action.
PSF heading error threshold Delta psi_th = 0.2 rad
Hand-selected in Table I; part of the safety constraint set X in eq. 16c.
Curvature adjustment bound kappa_th = 0.05 m^-1
Limits the RL action kappa_RL via eq. 8; chosen by hand to match feasible track curvatures.
Friction adjustment bound mu_th = 0.15
Limits the RL action mu_RL via eq. 9; hand-tuned in Table I to allow reasonable friction estimates without destabilizing the MPDC.
PSF rate limit Delta kappa_th = 0.01 m^-1
Constraint on curvature variation in eq. 16c; set to smooth the PSF output.
Constant steering angle delta0 = -0.5 rad
Assumed fixed steering angle for the drift equilibrium calculation in Section III-B; a modeling choice that anchors the equilibrium.
Reward weights lambda and c = lambda=10, c=100
Hand-tuned reward coefficients in eqs. 13-15 that balance lateral vs heading error and penalize MPDC infeasibility.
PSF cost weights Ws and Rs = Ws=10, Rs=1
Hand-selected in Table I for the PSF objective eq. 16a; these weights determine trade-off between tracking the RL action and smoothness.
MPDC weights Q and R = not specified
Referenced only as 'positive definite weighting matrices' in eq. 7; exact values are in reference [33], not in this paper. Their values likely affect the comparison.

assumptions (4)

domain assumption The PSF prediction model in eq. 17 keeps V, beta, and r constant over the prediction horizon.
Footnote 1 states this assumption explicitly. The safety filter's guarantees depend on the fidelity of this simplified model.
domain assumption The drift equilibrium calculation assumes a constant steering angle delta_eq = delta0 (Section III-B).
The MPDC references a fixed steering angle, which may not hold for transient drift maneuvers.
domain assumption The rear tire is at its friction limit, giving the friction circle relation Fyr = sqrt((mu Fzr)^2 - Fxr^2) in eq. 5.
This relation is valid only when the rear tire operates at the saturation limit, which is the intended drift regime but a strong modeling assumption.
domain assumption The single-track vehicle model (1)-(3) with Pacejka tires adequately captures the drifting dynamics.
This is the standard model used in the MPDC; its fidelity bounds the achievable control performance.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Safe Reinforcement Learning with a Predictive Safety Filter for Motion Planning and Control: A Drifting Vehicle Example." pith.science (2026). https://pith.science/paper/5Z7HMYXQ

@misc{pith2026250622894,
  author       = {Pith},
  title        = {Pith review of: Safe Reinforcement Learning with a Predictive Safety Filter for Motion Planning and Control: A Drifting Vehicle Example},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5Z7HMYXQ}},
  note         = {Machine review of arXiv:2506.22894}
}

read the original abstract

Autonomous drifting is a complex and crucial maneuver for safety-critical scenarios like slippery roads and emergency collision avoidance, requiring precise motion planning and control. Traditional motion planning methods often struggle with the high instability and unpredictability of drifting, particularly when operating at high speeds. Recent learning-based approaches have attempted to tackle this issue but often rely on expert knowledge or have limited exploration capabilities. Additionally, they do not effectively address safety concerns during learning and deployment. To overcome these limitations, we propose a novel Safe Reinforcement Learning (RL)-based motion planner for autonomous drifting. Our approach integrates an RL agent with model-based drift dynamics to determine desired drift motion states, while incorporating a Predictive Safety Filter (PSF) that adjusts the agent's actions online to prevent unsafe states. This ensures safe and efficient learning, and stable drift operation. We validate the effectiveness of our method through simulations on a Matlab-Carsim platform, demonstrating significant improvements in drift performance, reduced tracking errors, and computational efficiency compared to traditional methods. This strategy promises to extend the capabilities of autonomous vehicles in safety-critical maneuvers.

Figures

Figures reproduced from arXiv: 2506.22894 by the authors.

**Figure 1.** We design the PSF to act as a supervisory layer that [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** The average rewards, smoothened over 50 episodes, show that the [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Figures from the paper (3 more)

**Figure 3.** Figure 3: Box plots of various metrics for different planners during an [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]

**Figure 4.** Figure 4: Comparisons of path tracking performance among the proposed [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

**Figure 5.** Figure 5: Comparison of RL-planned curvature κRL and PSF-adjusted safe curvature κ0 in specific episodes of the Safe RL planner: As training progresses, the safety filter’s adjustments to the RL-planned curvature diminish, reflecting the agent’s improved ability to generate safe…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 29 canonical work pages

[15]

A novel model predictive controller for the drifting vehicle to track a circular trajectory,

C. Hu, L. Xie, Z. Zhang, and H. Xiong, “A novel model predictive controller for the drifting vehicle to track a circular trajectory,” Vehicle System Dynamics, vol. 0, no. 0, pp. 1–30, 2024

work page 2024
[1]

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

J. Betz, H. Zheng, A. Liniger, U. Rosolia, P. Karle, M. Behl, V . Krovi, and R. Mangharam, “Autonomous vehicles on the edge: A survey on autonomous vehicle racing,” IEEE Open Journal of Intelligent Transportation Systems, vol. 3, pp. 458–488, 2022

2022
[2]

How optimal is the minimum-time manoeuvre of an artificial race driver?

M. Piccinini, S. Taddei, E. Pagot, E. Bertolazzi, and F. Biral, “How optimal is the minimum-time manoeuvre of an artificial race driver?” Vehicle System Dynamics, vol. 0, no. 0, pp. 1–28, 2024

2024
[3]

Simultaneous stabilization and tracking of basic automobile drifting trajectories,

J. Y . Goh and J. C. Gerdes, “Simultaneous stabilization and tracking of basic automobile drifting trajectories,” in 2016 IEEE Intelligent Vehicles Symposium (IV), Jun. 2016, pp. 597–602

work page 2016
[4]

Dynamic drifting control for general path tracking of autonomous vehicles,

G. Chen, X. Zhao, Z. Gao, and M. Hua, “Dynamic drifting control for general path tracking of autonomous vehicles,” IEEE Transactions on Intelligent Vehicles, vol. 8, no. 3, pp. 2527–2537, Mar. 2023. 0 2 4 6 8 10 12 14 16 18 Time (s) 0.02 0.03 0.04 0.05 0.06Reference Curvature (1/m) RL-planned Curvature RL PSF Safe Curvature 0 Early Termination: | | > maxe...

work page 2023
[5]

Modeling and control for dynamic drifting trajectories,

T. P. Weber and J. C. Gerdes, “Modeling and control for dynamic drifting trajectories,” IEEE Transactions on Intelligent Vehicles, vol. 9, no. 2, pp. 3731–3741, Feb. 2024

work page 2024
[6]

Autonomous drifting using torque vectoring: Innovating active safety,

B. Lenzo, T. Goel, and J. Christian Gerdes, “Autonomous drifting using torque vectoring: Innovating active safety,” IEEE Transactions on Intelligent Transportation Systems , vol. 25, no. 11, pp. 17 931–17 939, Nov. 2024

work page 2024
[7]

Teaching a vehicle to autonomously drift: A data-based approach using neural networks,

M. Acosta and S. Kanarachos, “Teaching a vehicle to autonomously drift: A data-based approach using neural networks,” Knowledge-Based Systems, vol. 153, pp. 12–28, Aug. 2018

work page 2018

Show all 33 references

[8]

High-speed autonomous drifting with deep reinforcement learning,

P. Cai, X. Mei, L. Tai, Y . Sun, and M. Liu, “High-speed autonomous drifting with deep reinforcement learning,” IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 1247–1254, Apr. 2020

2020
[9]

Deep drifting: Autonomous drifting of arbitrary trajectories using deep reinforcement learning,

F. Domberg, C. C. Wembers, H. Patel, and G. Schildbach, “Deep drifting: Autonomous drifting of arbitrary trajectories using deep reinforcement learning,” in 2022 International Conference on Robotics and Automation (ICRA) , May 2022, pp. 7753–7759

2022
[10]

Autonomous drifting using simulation-aided reinforcement learning,

M. Cutler and J. P. How, “Autonomous drifting using simulation-aided reinforcement learning,” in 2016 IEEE International Conference on Robotics and Automation (ICRA) , May 2016, pp. 5442–5448

2016
[11]

One model to drift them all: Physics-informed conditional diffusion model for driving at the limits,

F. Djeumou, T. J. Lew, N. DING, M. Thompson, M. Suminaka, M. Greiff, and J. Subosits, “One model to drift them all: Physics-informed conditional diffusion model for driving at the limits,” in 8th Annual Conference on Robot Learning , 2024

2024
[12]

Learning-based mpc controller for drift control of autonomous vehicles,

X. Zhou, C. Hu, R. Duo, H. Xiong, Y . Qi, Z. Zhang, H. Su, and L. Xie, “Learning-based mpc controller for drift control of autonomous vehicles,” in 2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC). Macau, China: IEEE, Oct. 2022, pp. 322–328

2022
[13]

Autonomous driving at the handling limit using residual reinforcement learning,

X. Hou, J. Zhang, C. He, Y . Ji, J. Zhang, and J. Han, “Autonomous driving at the handling limit using residual reinforcement learning,” Advanced Engineering Informatics , vol. 54, p. 101754, Oct. 2022

2022
[14]

A harmonized approach: Beyond-the-limit control for autonomous vehicles balancing performance and safety in unpredictable environments,

S. Zhao, J. Zhang, X. He, C. He, X. Hou, H. Huang, and J. Han, “A harmonized approach: Beyond-the-limit control for autonomous vehicles balancing performance and safety in unpredictable environments,” IEEE Transactions on Intelligent Transportation Systems, pp. 1–14, 2024

2024
[16]

High-speed cornering control and real-vehicle deployment for autonomous electric vehicles,

S. Zhao, J. Zhang, N. Masoud, Y . Jiang, H. Huang, and T. Liu, “High-speed cornering control and real-vehicle deployment for autonomous electric vehicles,” Nov. 2024

2024
[17]

A review of safe reinforcement learning: Methods, theories and applications,

S. Gu, L. Yang, Y . Du, G. Chen, F. Walter, J. Wang, and A. Knoll, “A review of safe reinforcement learning: Methods, theories and applications,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

2024
[18]

Safe reinforcement learning using data-driven predictive control,

M. Selim, A. Alanwar, M. W. El-Kharashi, H. M. Abbas, and K. H. Johansson, “Safe reinforcement learning using data-driven predictive control,” in 2022 5th International Conference on Communications, Signal Processing, and their Applications (ICCSPA) . IEEE, 2022, pp. 1–6

2022
[19]

Saute RL: Almost surely safe reinforcement learning using state augmentation,

A. Sootla, A. I. Cowen-Rivers, T. Jafferjee, Z. Wang, D. H. Mguni, J. Wang, and H. Ammar, “Saute RL: Almost surely safe reinforcement learning using state augmentation,” in Proceedings of the 39th International Conference on Machine Learning , vol. 162. PMLR, 17–23 Jul 2022, p...

2022
[20]

Safe reinforcement learning via probabilistic logic shields,

W.-C. Yang, G. Marra, G. Rens, and L. De Raedt, “Safe reinforcement learning via probabilistic logic shields,” arXiv preprint arXiv:2303.03226, 2023

2023 arXiv
[21]

A safe reinforcement learning driven weights-varying model predictive control for autonomous vehicle motion control,

B. Zarrouki, M. Spanakakis, and J. Betz, “A safe reinforcement learning driven weights-varying model predictive control for autonomous vehicle motion control,” in 2024 IEEE Intelligent Vehicles Symposium (IV), 2024, pp. 1401–1408

2024
[22]

The safety filter: A unified view of safety-critical control in autonomous systems,

K.-C. Hsu, H. Hu, and J. F. Fisac, “The safety filter: A unified view of safety-critical control in autonomous systems,” Annual Review of Control, Robotics, and Autonomous Systems , vol. 7
[23]

Safety filters for black-box dynamical systems by learning discriminating hyperplanes,

W. Lavanakul, J. Choi, K. Sreenath, and C. Tomlin, “Safety filters for black-box dynamical systems by learning discriminating hyperplanes,” in Proceedings of the 6th Annual Learning for Dynamics amp; Control Conference, ser. Proceedings of Machine Learning Research, vol. 242. ...

2024
[24]

Safety filtering while training: Improving the performance and sample efficiency of reinforcement learning agents,

F. P. Bejarano, L. Brunke, and A. P. Schoellig, “Safety filtering while training: Improving the performance and sample efficiency of reinforcement learning agents,” IEEE Robotics and Automation Letters, vol. 10, no. 1, pp. 788–795, 2025

2025
[25]

A predictive safety filter for learning-based control of constrained nonlinear dynamical systems,

K. P. Wabersich and M. N. Zeilinger, “A predictive safety filter for learning-based control of constrained nonlinear dynamical systems,” Automatica, vol. 129, p. 109597, 2021

2021
[26]

Predictive safety filter using system level synthesis,

A. Leeman, J. K ¨ohler, S. Bennani, and M. Zeilinger, “Predictive safety filter using system level synthesis,” in Proceedings of The 5th Annual Learning for Dynamics and Control Conference , vol. 211. PMLR, 15–16 Jun 2023, pp. 1180–1192

2023
[27]

Tyre modelling for use in vehicle dynamics studies,

E. Bakker, L. Nyborg, and H. B. Pacejka, “Tyre modelling for use in vehicle dynamics studies,” SAE Transactions, pp. 190–204, 1987

1987
[28]

Toward automated vehicle control beyond the stability limits: Drifting along a general path,

J. Y . Goh, T. Goel, and J. Christian Gerdes, “Toward automated vehicle control beyond the stability limits: Drifting along a general path,” Journal of Dynamic Systems, Measurement, and Control , vol. 142, no. 021004, Nov. 2019

2019
[29]

A controller framework for autonomous drifting: Design, stability, and experimental validation,

R. Y . Hindiyeh and J. Christian Gerdes, “A controller framework for autonomous drifting: Design, stability, and experimental validation,” Journal of Dynamic Systems, Measurement, and Control, vol. 136, no. 051015, Jul. 2014

2014
[30]

Learning-based hierarchical model predictive control for drift vehicles,

B. Zhou, C. Hu, Y . Shi, X. Hu, L. Xie, and H. Su, “Learning-based hierarchical model predictive control for drift vehicles,” in 2024 American Control Conference (ACC) , Jul. 2024, pp. 3524–3530

2024
[31]

Continuous control with deep reinforcement learning,

T. P. Lillicrap, J. J. Hunt, A. Pritzel, N. Heess, T. Erez, Y . Tassa, D. Silver, and D. Wierstra, “Continuous control with deep reinforcement learning,” Jul. 2019

2019
[32]

A predictive safety filter for learning-based racing control,

B. Tearle, K. P. Wabersich, A. Carron, and M. N. Zeilinger, “A predictive safety filter for learning-based racing control,” IEEE Robotics and Automation Letters , vol. 6, no. 4, pp. 7635–7642, Oct. 2021

2021
[33]

Adaptive learning-based model predictive control strategy for drift vehicles,

B. Zhou, C. Hu, J. Zeng, Z. Li, J. Betz, L. Xie, and H. Su, “Adaptive learning-based model predictive control strategy for drift vehicles,” Robotics and Autonomous Systems , p. 104941, 2025

2025

Pith tools

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

[1] [15]

A novel model predictive controller for the drifting vehicle to track a circular trajectory,

C. Hu, L. Xie, Z. Zhang, and H. Xiong, “A novel model predictive controller for the drifting vehicle to track a circular trajectory,” Vehicle System Dynamics, vol. 0, no. 0, pp. 1–30, 2024

work page 2024

[2] [1]

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

J. Betz, H. Zheng, A. Liniger, U. Rosolia, P. Karle, M. Behl, V . Krovi, and R. Mangharam, “Autonomous vehicles on the edge: A survey on autonomous vehicle racing,” IEEE Open Journal of Intelligent Transportation Systems, vol. 3, pp. 458–488, 2022

2022

[3] [2]

How optimal is the minimum-time manoeuvre of an artificial race driver?

M. Piccinini, S. Taddei, E. Pagot, E. Bertolazzi, and F. Biral, “How optimal is the minimum-time manoeuvre of an artificial race driver?” Vehicle System Dynamics, vol. 0, no. 0, pp. 1–28, 2024

2024

[4] [3]

Simultaneous stabilization and tracking of basic automobile drifting trajectories,

J. Y . Goh and J. C. Gerdes, “Simultaneous stabilization and tracking of basic automobile drifting trajectories,” in 2016 IEEE Intelligent Vehicles Symposium (IV), Jun. 2016, pp. 597–602

work page 2016

[5] [4]

Dynamic drifting control for general path tracking of autonomous vehicles,

G. Chen, X. Zhao, Z. Gao, and M. Hua, “Dynamic drifting control for general path tracking of autonomous vehicles,” IEEE Transactions on Intelligent Vehicles, vol. 8, no. 3, pp. 2527–2537, Mar. 2023. 0 2 4 6 8 10 12 14 16 18 Time (s) 0.02 0.03 0.04 0.05 0.06Reference Curvature (1/m) RL-planned Curvature RL PSF Safe Curvature 0 Early Termination: | | > maxe...

work page 2023

[6] [5]

Modeling and control for dynamic drifting trajectories,

T. P. Weber and J. C. Gerdes, “Modeling and control for dynamic drifting trajectories,” IEEE Transactions on Intelligent Vehicles, vol. 9, no. 2, pp. 3731–3741, Feb. 2024

work page 2024

[7] [6]

Autonomous drifting using torque vectoring: Innovating active safety,

B. Lenzo, T. Goel, and J. Christian Gerdes, “Autonomous drifting using torque vectoring: Innovating active safety,” IEEE Transactions on Intelligent Transportation Systems , vol. 25, no. 11, pp. 17 931–17 939, Nov. 2024

work page 2024

[8] [7]

Teaching a vehicle to autonomously drift: A data-based approach using neural networks,

M. Acosta and S. Kanarachos, “Teaching a vehicle to autonomously drift: A data-based approach using neural networks,” Knowledge-Based Systems, vol. 153, pp. 12–28, Aug. 2018

work page 2018