pith. sign in

arxiv: 2509.07464 · v2 · submitted 2025-09-09 · 💻 cs.RO · cs.SY· eess.SY

Safe and Nonconservative Contingency Planning for Autonomous Vehicles via Online Learning-Based Reachable Set Barriers

Rui Yang , Lei Zheng , Shuzhi Sam Ge , Jun Ma This is my paper

Pith reviewed 2026-05-18 18:17 UTC · model grok-4.3

classification 💻 cs.RO cs.SYeess.SY
keywords autonomous vehiclescontingency planningforward reachable setsonline learningsafety barrierstrajectory optimizationhuman-driven vehiclesuncertainty quantification
0
0 comments X

The pith

Autonomous vehicles maintain invariant safety by using online-learned forward reachable sets of human vehicles as barrier constraints in contingency trajectory planning.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a real-time contingency planning framework for autonomous vehicles that learns human-driven vehicle control-intent sets in an event-triggered manner. This learning dynamically quantifies multimodal uncertainties and incrementally updates forward reachable sets without assuming accurate future trajectory predictions. These reachable sets are converted into barrier constraints that are embedded directly into a trajectory optimization problem solved via consensus ADMM. The result is planning that stays safe and feasible while adapting to reduce unnecessary conservatism, as shown in highway and urban simulations plus real-world tests.

Core claim

Event-triggered online learning of HV control-intent sets produces refined forward reachable sets that, when imposed as barrier constraints inside contingency trajectory optimization, guarantee invariant safety for the autonomous vehicle even under multimodal behavioral uncertainties and perception errors, all while preserving driving efficiency through real-time adaptation.

What carries the argument

Forward reachable sets (FRSs) generated from online-learned HV control-intent sets, enforced as invariant safety barrier constraints within consensus ADMM-based contingency trajectory optimization.

If this is right

  • Safety holds without requiring precise prediction of other vehicles' future paths.
  • Planning avoids excessive conservatism by continuously refining uncertainty bounds.
  • Feasibility of the optimization is maintained while safety constraints remain active.
  • Efficiency and comfort metrics improve in both highway and urban driving under uncertainty.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same barrier mechanism could be applied to other uncertain agents such as pedestrians or cyclists by learning their intent sets online.
  • Over longer horizons the method might reduce dependence on high-fidelity perception by letting learned reachable sets absorb sensor noise.
  • Integration with reinforcement learning for faster intent-set updates could further lower computational load during frequent events.

Load-bearing premise

The event-triggered online learning must produce forward reachable sets that truly contain every possible future behavior of the human vehicle so that the barriers enforce genuine invariant safety.

What would settle it

A controlled experiment in which a human-driven vehicle executes a maneuver outside the learned reachable set and collides with the autonomous vehicle despite the active barrier constraints would disprove the safety guarantee.

Figures

Figures reproduced from arXiv: 2509.07464 by Jun Ma, Lei Zheng, Rui Yang, Shuzhi Sam Ge.

Figure 1
Figure 1. Figure 1: Overview of the proposed safe contingency trajectory planning frame [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the FRS-based safety barrier. At predicted time step [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Snapshots of the EV navigating dense traffic in the NGSIM dataset. The EV (red) jointly optimizes a nominal trajectory (red) and a contingency [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Trajectory and speed profiles of the EV under different planning [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Evolution of longitudinal and lateral acceleration profiles of EV [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Snapshots of the EV safely navigating through an unsignalized intersection. (a)-(d) demonstrate the EV’s (red) response to a southbound HV’s abrupt [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Trajectory comparison in the intersection scenario demonstrates [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The minimum EV-HV distances (dmin) show that our method maintains appropriate safety margins. The Deterministic Barrier Planner and ST-RHC fail to avoid collision, while the Worst-case Barrier Planner and the Uncertainty-Aware planner exhibit unnecessarily conservative distances. of contingency planning capability results in abrupt lateral maneuvers and degraded driving efficiency. The discontinuous nature… view at source ↗
Figure 9
Figure 9. Figure 9: Impact of the consensus step parameter Ns on intersection navigation behaviors. Smaller values enable aggressive maneuvers while larger values enforce conservative strategies with preemptive deceleration. 3) Discussion. The effectiveness of the proposed method is affected by two critical parameters: the consensus steps parameter Ns and the weight ps. These parameters collectively determine how the vehicle … view at source ↗
Figure 12
Figure 12. Figure 12: Overtaking trajectory comparison under different uncertainty levels: [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
Figure 11
Figure 11. Figure 11: Overlaying multiple frames of experiments showing (a) Scenario I: [PITH_FULL_IMAGE:figures/full_fig_p014_11.png] view at source ↗
Figure 13
Figure 13. Figure 13: Evolution of control-intent set volume during online learning. The ˆ [PITH_FULL_IMAGE:figures/full_fig_p014_13.png] view at source ↗
read the original abstract

Autonomous vehicles must navigate dynamically uncertain environments while balancing safety and efficiency. This challenge is exacerbated by unpredictable human-driven vehicle (HV) behaviors and perception inaccuracies, necessitating planners that adapt to evolving uncertainties while maintaining safe trajectories. Overly conservative planning degrades driving efficiency, while deterministic methods risk failure in unexpected scenarios. To address these issues, we propose a real-time contingency trajectory optimization framework. Our method employs event-triggered online learning of HV control-intent sets to dynamically quantify multimodal HV uncertainties and incrementally refine their forward reachable sets (FRSs). Crucially, we enforce invariant safety through FRS-based barrier constraints that ensure safety without reliance on accurate trajectory prediction. These constraints are seamlessly embedded in contingency trajectory optimization and solved efficiently through consensus alternating direction method of multipliers (ADMM). The system continuously adapts to HV behavioral uncertainties, preserving feasibility and safety without excessive conservatism. High-fidelity simulations on highway and urban scenarios, along with a series of real-world experiments, demonstrate significant improvements in driving efficiency and passenger comfort while maintaining safety under uncertainty. The project page is available at https://pathetiue.github.io/frscp.github.io/.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper proposes a real-time contingency trajectory optimization framework for autonomous vehicles operating in environments with uncertain human-driven vehicle (HV) behaviors. It employs event-triggered online learning to dynamically refine multimodal HV control-intent sets and their associated forward reachable sets (FRSs). These FRSs are used to construct barrier constraints that are embedded in a contingency planner solved via consensus ADMM, with the goal of enforcing invariant safety without relying on accurate trajectory predictions while reducing conservatism. The approach is evaluated in high-fidelity highway and urban simulations plus real-world experiments, reporting gains in efficiency and comfort.

Significance. If the safety claims hold with formal guarantees, the work would offer a practical advance in balancing safety and efficiency for AVs in mixed-traffic settings by adapting online to observed uncertainties rather than using fixed conservative bounds. The combination of event-triggered learning, reachable-set barriers, and ADMM-based optimization is technically interesting for real-time deployment, though its impact depends on establishing that the learned sets and FRSs deliver the claimed invariant properties.

major comments (1)
  1. [Abstract and method description] The central safety claim—that satisfying the FRS-based barrier constraints renders the ego trajectory invariant-safe for every HV behavior inside the learned sets—requires that the online learner produces conservative over-approximations and that the FRS propagation is a true outer approximation. The manuscript describes the event-triggered update and FRS construction but provides neither a proof of conservatism for the learned control-intent sets nor an error bound on the learning rule or FRS computation; this renders the invariant safety conditional on an unproven modeling assumption rather than enforced by construction.
minor comments (1)
  1. [Abstract] The abstract states that the system 'preserves feasibility and safety without excessive conservatism,' but the precise definition of 'excessive' and the quantitative metric used to demonstrate this in the experiments should be clarified for reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback, which highlights an important aspect of our safety claims. We address the major comment below and outline revisions to strengthen the formal guarantees.

read point-by-point responses

  1. Referee: [Abstract and method description] The central safety claim—that satisfying the FRS-based barrier constraints renders the ego trajectory invariant-safe for every HV behavior inside the learned sets—requires that the online learner produces conservative over-approximations and that the FRS propagation is a true outer approximation. The manuscript describes the event-triggered update and FRS construction but provides neither a proof of conservatism for the learned control-intent sets nor an error bound on the learning rule or FRS computation; this renders the invariant safety conditional on an unproven modeling assumption rather than enforced by construction.

    Authors: We agree that the invariant-safety claim depends on the learned sets and FRSs being conservative over-approximations, and that the manuscript currently describes the event-triggered update and FRS construction without supplying a formal proof or explicit error bounds. The design of the learning rule is intended to maintain over-approximation by construction (via inclusion of all observed modes and incremental expansion), yet this property is not proven. In the revised manuscript we will add a dedicated subsection containing (i) a theorem stating that the event-triggered update rule produces a conservative superset of the true control-intent distribution under the stated Lipschitz and boundedness assumptions on HV dynamics, and (ii) a discretization-error bound for the FRS propagation that quantifies the Hausdorff distance between the computed and true reachable sets. These additions will make the safety guarantee conditional on the proven properties rather than on an implicit modeling assumption. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation uses standard FRS and barrier methods

full rationale

The paper claims invariant safety via FRS-based barrier constraints embedded in contingency optimization solved by ADMM. This rests on set-theoretic properties of forward reachable sets constructed from learned control-intent sets and standard optimization, without any quoted reduction where a prediction or safety guarantee equals its own fitted input or self-citation by construction. The event-triggered learning refines sets from data, but the barrier enforcement step does not tautologically redefine safety in terms of the learning outputs. No self-definitional, fitted-prediction, or uniqueness-imported patterns appear in the provided derivation chain. The framework is self-contained against external set-theoretic benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete; the central claim rests on standard domain assumptions from control theory and optimization whose details are not extractable here.

axioms (1)
  • domain assumption Forward reachable sets computed from learned control-intent sets can be used to enforce invariant safety constraints
    Invoked when the paper states that FRS-based barriers ensure safety without accurate trajectory prediction.

pith-pipeline@v0.9.0 · 5738 in / 1272 out tokens · 47395 ms · 2026-05-18T18:17:36.009020+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

63 extracted references · 63 canonical work pages

  1. [1]

    Milestones in autonomous driving and intelligent vehicles: Survey of surveys,

    L. Chen, Y . Li, C. Huang, B. Li, Y . Xing, D. Tian, L. Li, Z. Hu, X. Na, Z. Liet al., “Milestones in autonomous driving and intelligent vehicles: Survey of surveys,”IEEE Transactions on Intelligent Vehicles, vol. 8, no. 2, pp. 1046–1056, 2022

    work page 2022

  2. [2]

    Barrier-enhanced parallel homotopic trajectory optimization for safety-critical autonomous driv- ing,

    L. Zheng, R. Yang, M. Yu Wang, and J. Ma, “Barrier-enhanced parallel homotopic trajectory optimization for safety-critical autonomous driv- ing,”IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 2, pp. 2169–2186, 2025

    work page 2025

  3. [3]

    Active exploration in iterative Gaussian process regression for uncertainty modeling in autonomous racing,

    T. Benciolini, C. Tang, M. Leibold, C. Weaver, M. Tomizuka, and W. Zhan, “Active exploration in iterative Gaussian process regression for uncertainty modeling in autonomous racing,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1301–1316, 2025

    work page 2025

  4. [4]

    RACP: Risk-aware contingency planning with multi-modal predictions,

    K. A. Mustafa, D. J. Ornia, J. Kober, and J. Alonso-Mora, “RACP: Risk-aware contingency planning with multi-modal predictions,”IEEE Transactions on Intelligent Vehicles, pp. 1–16, 2024

    work page 2024

  5. [5]

    Automated lane merging via game theory and branch model predictive control,

    L. Zhang, S. Han, and S. Grammatico, “Automated lane merging via game theory and branch model predictive control,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1258–1269, 2025

    work page 2025

  6. [6]

    Interaction-aware traffic prediction and scenario-based model predictive control for autonomous vehicles on highways,

    X. Zhang, S. Zeinali, and G. Schildbach, “Interaction-aware traffic prediction and scenario-based model predictive control for autonomous vehicles on highways,”IEEE Transactions on Control Systems Technol- ogy, vol. 33, no. 4, pp. 1235–1245, 2025

    work page 2025

  7. [7]

    Spatiotemporal receding horizon control with proactive interaction towards autonomous driving in dense traffic,

    L. Zheng, R. Yang, Z. Peng, M. Y . Wang, and J. Ma, “Spatiotemporal receding horizon control with proactive interaction towards autonomous driving in dense traffic,”IEEE Transactions on Intelligent Vehicles, vol. 9, no. 11, pp. 6853–6868, 2024

    work page 2024

  8. [8]

    Set-based prediction of traffic participants on arbitrary road networks,

    M. Althoff and S. Magdici, “Set-based prediction of traffic participants on arbitrary road networks,”IEEE Transactions on Intelligent Vehicles, vol. 1, no. 2, pp. 187–202, 2016

    work page 2016

  9. [9]

    Proactive emergency collision avoidance for automated driving in highway scenarios,

    L. Gharavi, A. Dabiri, J. Verkuijlen, B. De Schutter, and S. Baldi, “Proactive emergency collision avoidance for automated driving in highway scenarios,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1164–1177, 2025

    work page 2025

  10. [10]

    Queues and artificial potential trenches for multirobot formations,

    S. Ge and C.-H. Fua, “Queues and artificial potential trenches for multirobot formations,”IEEE Transactions on Robotics, vol. 21, no. 4, pp. 646–656, 2005

    work page 2005

  11. [11]

    Chance-constrained collision avoidance for mavs in dynamic environments,

    H. Zhu and J. Alonso-Mora, “Chance-constrained collision avoidance for mavs in dynamic environments,”IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 776–783, 2019

    work page 2019

  12. [12]

    Predictive control for autonomous driving with uncertain, multimodal predictions,

    S. H. Nair, H. Lee, E. Joa, Y . Wang, H. E. Tseng, and F. Borrelli, “Predictive control for autonomous driving with uncertain, multimodal predictions,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1178–1192, 2025

    work page 2025

  13. [13]

    Scenario- based motion planning with bounded probability of collision,

    O. de Groot, L. Ferranti, D. M. Gavrila, and J. Alonso-Mora, “Scenario- based motion planning with bounded probability of collision,”The International Journal of Robotics Research, 2025

    work page 2025

  14. [14]

    Stochastic model predictive control with a safety guarantee for automated driving,

    T. Br ¨udigam, M. Olbrich, D. Wollherr, and M. Leibold, “Stochastic model predictive control with a safety guarantee for automated driving,” IEEE Transactions on Intelligent Vehicles, vol. 8, no. 1, pp. 22–36, 2021

    work page 2021

  15. [15]

    Recursively feasible chance-constrained model predictive control under gaussian mixture model uncertainty,

    K. Ren, C. Chen, H. Sung, H. Ahn, I. M. Mitchell, and M. Kamgarpour, “Recursively feasible chance-constrained model predictive control under gaussian mixture model uncertainty,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1193–1206, 2025

    work page 2025

  16. [16]

    Non-gaussian chance- constrained trajectory planning for autonomous vehicles under agent uncertainty,

    A. Wang, A. Jasour, and B. C. Williams, “Non-gaussian chance- constrained trajectory planning for autonomous vehicles under agent uncertainty,”IEEE Robotics and Automation Letters, vol. 5, no. 4, pp. 6041–6048, 2020

    work page 2020

  17. [17]

    Scenario-based trajectory optimization in uncertain dynamic environ- ments,

    O. de Groot, B. Brito, L. Ferranti, D. Gavrila, and J. Alonso-Mora, “Scenario-based trajectory optimization in uncertain dynamic environ- ments,”IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5389– 5396, 2021

    work page 2021

  18. [18]

    Robust convex optimization,

    A. Ben-Tal and A. Nemirovski, “Robust convex optimization,”Mathe- matics of Operations Research, vol. 23, no. 4, pp. 769–805, 1998

    work page 1998

  19. [19]

    Safe motion planner for autonomous driving based on LPV MPC and reachability analysis,

    ´A. Carrizosa-Rend ´on, V . Puig, and F. Nejjari, “Safe motion planner for autonomous driving based on LPV MPC and reachability analysis,” Control Engineering Practice, vol. 147, p. 105932, 2024

    work page 2024

  20. [20]

    Real-time robust receding horizon planning using Hamilton–Jacobi reachability analysis,

    H. Seo, D. Lee, C. Y . Son, I. Jang, C. J. Tomlin, and H. J. Kim, “Real-time robust receding horizon planning using Hamilton–Jacobi reachability analysis,”IEEE Transactions on Robotics, vol. 39, no. 1, pp. 90–109, 2022

    work page 2022

  21. [21]

    Using reachable sets for trajec- tory planning of automated vehicles,

    S. Manzinger, C. Pek, and M. Althoff, “Using reachable sets for trajec- tory planning of automated vehicles,”IEEE Transactions on Intelligent Vehicles, vol. 6, no. 2, pp. 232–248, 2020

    work page 2020

  22. [22]

    Set-based prediction of traffic participants considering occlusions and traffic rules,

    M. Koschi and M. Althoff, “Set-based prediction of traffic participants considering occlusions and traffic rules,”IEEE Transactions on Intelli- gent Vehicles, vol. 6, no. 2, pp. 249–265, 2020

    work page 2020

  23. [23]

    Experimental valida- tion of safe MPC for autonomous driving in uncertain environments,

    I. Batkovic, A. Gupta, M. Zanon, and P. Falcone, “Experimental valida- tion of safe MPC for autonomous driving in uncertain environments,” IEEE Transactions on Control Systems Technology, vol. 31, no. 5, pp. 2027–2042, 2023

    work page 2027

  24. [24]

    MPC with integrated evasive maneuvers for failure-safe automated driving,

    T. Skibik, A. P. Vinod, A. Weiss, and S. Di Cairano, “MPC with integrated evasive maneuvers for failure-safe automated driving,” in2023 American Control Conference (ACC), 2023, pp. 1122–1128

    work page 2023

  25. [25]

    Computationally efficient fail-safe trajectory planning for self-driving vehicles using convex optimization,

    C. Pek and M. Althoff, “Computationally efficient fail-safe trajectory planning for self-driving vehicles using convex optimization,” inInter- national Conference on Intelligent Transportation Systems (ITSC), 2018, pp. 1447–1454

    work page 2018

  26. [26]

    Fail-safe motion planning for online verification of autonomous vehicles using convex optimization,

    ——, “Fail-safe motion planning for online verification of autonomous vehicles using convex optimization,”IEEE Transactions on Robotics, vol. 37, no. 3, pp. 798–814, 2020

    work page 2020

  27. [27]

    Safe and non-conservative trajectory planning for autonomous driving handling unanticipated behaviors of traffic participants,

    T. Benciolini, M. Fink, N. G ¨uzelkaya, D. Wollherr, and M. Leibold, “Safe and non-conservative trajectory planning for autonomous driving handling unanticipated behaviors of traffic participants,”arXiv preprint arXiv:2406.13396, 2024

    work page arXiv 2024

  28. [28]

    Trust-aware safe control for autonomous nav- igation: Estimation of system-to-human trust for trust-adaptive control barrier functions,

    S. Ejaz and M. Inoue, “Trust-aware safe control for autonomous nav- igation: Estimation of system-to-human trust for trust-adaptive control barrier functions,”IEEE Transactions on Control Systems Technology, vol. 33, no. 4, pp. 1151–1163, 2025

    work page 2025

  29. [29]

    Model predictive control with preview: Recursive feasibility and stability,

    X. Fang and W.-H. Chen, “Model predictive control with preview: Recursive feasibility and stability,”IEEE Control Systems Letters, vol. 6, pp. 2647–2652, 2022

    work page 2022

  30. [30]

    Invariant sets for constrained nonlinear discrete-time systems with application to feasibility in model predictive control,

    E. Kerrigan and J. Maciejowski, “Invariant sets for constrained nonlinear discrete-time systems with application to feasibility in model predictive control,” inIEEE Conference on Decision and Control (CDC), vol. 5, 2000, pp. 4951–4956

    work page 2000

  31. [31]

    Safety-critical model predictive control with discrete-time control barrier function,

    J. Zeng, B. Zhang, and K. Sreenath, “Safety-critical model predictive control with discrete-time control barrier function,” inAmerican Control Conference (ACC), 2021, pp. 3882–3889

    work page 2021

  32. [32]

    Robust control barrier–value functions for safety-critical control,

    J. J. Choi, D. Lee, K. Sreenath, C. J. Tomlin, and S. L. Herbert, “Robust control barrier–value functions for safety-critical control,” in IEEE Conference on Decision and Control (CDC), 2021, pp. 6814– 6821

    work page 2021

  33. [33]

    Fast, smooth, and safe: implicit control barrier functions through reach- avoid differential dynamic programming,

    A. R. Kumar, K.-C. Hsu, P. J. Ramadge, and J. F. Fisac, “Fast, smooth, and safe: implicit control barrier functions through reach- avoid differential dynamic programming,”IEEE Control Systems Letters, vol. 7, pp. 2994–2999, 2023

    work page 2023

  34. [34]

    Safety-critical coordination of legged robots via layered controllers and forward reachable set based control barrier functions,

    J. Kim, J. Lee, and A. D. Ames, “Safety-critical coordination of legged robots via layered controllers and forward reachable set based control barrier functions,” inIEEE International Conference on Robotics and Automation (ICRA), 2024, pp. 3478–3484

    work page 2024

  35. [35]

    Risk-aware optimal control for automated overtaking with safety guarantees,

    Y . Gao, F. J. Jiang, L. Xie, and K. H. Johansson, “Risk-aware optimal control for automated overtaking with safety guarantees,”IEEE Trans- actions on Control Systems Technology, vol. 30, no. 4, pp. 1460–1472, 2021

    work page 2021

  36. [36]

    Reachability-based confidence-aware probabilistic collision detection in highway driving,

    X. Wang, Z. Li, J. Alonso-Mora, and M. Wang, “Reachability-based confidence-aware probabilistic collision detection in highway driving,” Engineering, vol. 33, pp. 90–107, 2024

    work page 2024

  37. [37]

    A Hamilton-Jacobi reachability-based framework for predicting and analyzing human motion for safe planning,

    S. Bansal, A. Bajcsy, E. Ratner, A. D. Dragan, and C. J. Tomlin, “A Hamilton-Jacobi reachability-based framework for predicting and analyzing human motion for safe planning,” inIEEE International Conference on Robotics and Automation (ICRA), 2020, pp. 7149–7155

    work page 2020

  38. [38]

    Robust, informative human-in-the-loop predictions via empirical reachable sets,

    K. Driggs-Campbell, R. Dong, and R. Bajcsy, “Robust, informative human-in-the-loop predictions via empirical reachable sets,”IEEE Transactions on Intelligent Vehicles, vol. 3, no. 3, pp. 300–309, 2018

    work page 2018

  39. [39]

    Convex approximation of probabilistic reachable sets from small samples using self-supervised neural networks,

    J. Xiang and J. Chen, “Convex approximation of probabilistic reachable sets from small samples using self-supervised neural networks,”arXiv preprint arXiv:2411.14356, 2024

    work page arXiv 2024

  40. [40]

    Estimating high probability reachable sets using Gaussian processes,

    M. E. Cao, M. Bloch, and S. Coogan, “Estimating high probability reachable sets using Gaussian processes,” inIEEE Conference on Decision and Control (CDC), 2021, pp. 3881–3886

    work page 2021

  41. [41]

    Data-driven reachable set computation using adaptive gaussian process classification and monte carlo methods,

    A. Devonport and M. Arcak, “Data-driven reachable set computation using adaptive gaussian process classification and monte carlo methods,” inAmerican control conference (ACC), 2020, pp. 2629–2634

    work page 2020

  42. [42]

    Active learning for estimating reachable sets for systems with unknown dynamics,

    A. Chakrabarty, C. Danielson, S. Di Cairano, and A. Raghunathan, “Active learning for estimating reachable sets for systems with unknown dynamics,”IEEE Transactions on Cybernetics, vol. 52, no. 4, pp. 2531– 2542, 2020

    work page 2020

  43. [43]

    Robust predictive motion planning by learning obstacle uncertainty,

    J. Zhou, Y . Gao, O. Johansson, B. Olofsson, and E. Frisk, “Robust predictive motion planning by learning obstacle uncertainty,”IEEE 16 Transactions on Control Systems Technology, vol. 33, no. 3, pp. 1006– 1020, 2025

    work page 2025

  44. [44]

    Robust motion planning for autonomous vehicles based on environment and uncertainty-aware reachability prediction,

    J. Zhou, Y . Gao, B. Olofsson, and E. Frisk, “Robust motion planning for autonomous vehicles based on environment and uncertainty-aware reachability prediction,”Control Engineering Practice, vol. 160, p. 106319, 2025

    work page 2025

  45. [45]

    Contingency planning over probabilistic obstacle predictions for autonomous road vehicles,

    J. Hardy and M. Campbell, “Contingency planning over probabilistic obstacle predictions for autonomous road vehicles,”IEEE Transactions on Robotics, vol. 29, no. 4, pp. 913–929, 2013

    work page 2013

  46. [46]

    Interactive multi-modal motion planning with branch model predictive control,

    Y . Chen, U. Rosolia, W. Ubellacker, N. Csomay-Shanklin, and A. D. Ames, “Interactive multi-modal motion planning with branch model predictive control,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 5365–5372, 2022

    work page 2022

  47. [47]

    A non-conservatively defensive strategy for urban autonomous driving,

    W. Zhan, C. Liu, C.-Y . Chan, and M. Tomizuka, “A non-conservatively defensive strategy for urban autonomous driving,” inIEEE International Conference on Intelligent Transportation Systems (ITSC), 2016, pp. 459– 464

    work page 2016

  48. [48]

    Invariant safe contingency model predictive control for intersection coordination of mixed traffic,

    X. Chen and J. M ˚artensson, “Invariant safe contingency model predictive control for intersection coordination of mixed traffic,” inIEEE Interna- tional Conference on Intelligent Transportation Systems (ITSC), 2023, pp. 3369–3376

    work page 2023

  49. [49]

    Alternat- ing direction method of multipliers for constrained iterative LQR in autonomous driving,

    J. Ma, Z. Cheng, X. Zhang, M. Tomizuka, and T. H. Lee, “Alternat- ing direction method of multipliers for constrained iterative LQR in autonomous driving,”IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 12, pp. 23 031–23 042, 2022

    work page 2022

  50. [50]

    Multi-modal model predictive control through batch non-holonomic trajectory optimization: Application to highway driving,

    V . K. Adajania, A. Sharma, A. Gupta, H. Masnavi, K. M. Krishna, and A. K. Singh, “Multi-modal model predictive control through batch non-holonomic trajectory optimization: Application to highway driving,” IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 4220–4227, 2022

    work page 2022

  51. [51]

    Occlusion-aware contingency safety-critical planning for autonomous vehicles,

    L. Zheng, R. Yang, M. Zheng, Z. Peng, M. Y . Wang, and J. Ma, “Occlusion-aware contingency safety-critical planning for autonomous vehicles,”arXiv preprint arXiv:2502.06359, 2025

    work page arXiv 2025

  52. [52]

    Interactive joint planning for autonomous vehicles,

    Y . Chen, S. Veer, P. Karkus, and M. Pavone, “Interactive joint planning for autonomous vehicles,”IEEE Robotics and Automation Letters, vol. 9, no. 2, pp. 987–994, 2023

    work page 2023

  53. [53]

    S. P. Boyd and L. Vandenberghe,Convex Optimization. Cambridge, UK: Cambridge University Press, 2004

    work page 2004

  54. [54]

    On Khachiyan’s algorithm for the computation of minimum-volume enclosing ellipsoids,

    M. J. Todd and E. A. Yıldırım, “On Khachiyan’s algorithm for the computation of minimum-volume enclosing ellipsoids,”Discrete Applied Mathematics, vol. 155, no. 13, pp. 1731–1744, 2007

    work page 2007

  55. [55]

    Safe model- based reinforcement learning with stability guarantees,

    F. Berkenkamp, M. Turchetta, A. Schoellig, and A. Krause, “Safe model- based reinforcement learning with stability guarantees,”Advances in Neural Information Processing Systems, vol. 30, 2017

    work page 2017

  56. [56]

    Discrete control barrier functions for safety-critical control of discrete systems with application to bipedal robot navigation

    A. Agrawal and K. Sreenath, “Discrete control barrier functions for safety-critical control of discrete systems with application to bipedal robot navigation.” inRobotics: Science and Systems, 2017

    work page 2017

  57. [57]

    Farin,Curves and Surfaces for CAGD: A Practical Guide, 5th ed

    G. Farin,Curves and Surfaces for CAGD: A Practical Guide, 5th ed. San Francisco, CA, USA: Morgan Kaufmann Publishers Inc., 2001

    work page 2001

  58. [58]

    A survey on trajectory-prediction methods for autonomous driving,

    Y . Huang, J. Du, Z. Yang, Z. Zhou, L. Zhang, and H. Chen, “A survey on trajectory-prediction methods for autonomous driving,”IEEE Transactions on Intelligent Vehicles, vol. 7, no. 3, pp. 652–674, 2022

    work page 2022

  59. [59]

    Training neural networks without gradients: A scalable ADMM ap- proach,

    G. Taylor, R. Burmeister, Z. Xu, B. Singh, A. Patel, and T. Goldstein, “Training neural networks without gradients: A scalable ADMM ap- proach,” inInternational Conference on Machine Learning, 2016, pp. 2722–2731

    work page 2016

  60. [60]

    Optimal parameter selection for the alternating direction method of multipliers (ADMM): Quadratic problems,

    E. Ghadimi, A. Teixeira, I. Shames, and M. Johansson, “Optimal parameter selection for the alternating direction method of multipliers (ADMM): Quadratic problems,”IEEE Transactions on Automatic Con- trol, vol. 60, no. 3, pp. 644–658, 2015

    work page 2015

  61. [61]

    Distributed optimization and statistical learning via the alternating direction method of multipliers,

    S. Boyd, N. Parikh, E. Chu, B. Peleato, J. Ecksteinet al., “Distributed optimization and statistical learning via the alternating direction method of multipliers,”Foundations and Trends in Machine Learning, vol. 3, no. 1, pp. 1–122, 2011

    work page 2011

  62. [62]

    ACADO toolkit – An open source framework for automatic control and dynamic optimization,

    B. Houska, H. Ferreau, and M. Diehl, “ACADO toolkit – An open source framework for automatic control and dynamic optimization,”Optimal Control Applications and Methods, vol. 32, no. 3, pp. 298–312, 2011

    work page 2011

  63. [63]

    Safe and real- time consistent planning for autonomous vehicles in partially observed environments via parallel consensus optimization,

    L. Zheng, R. Yang, M. Zheng, M. Y . Wang, and J. Ma, “Safe and real- time consistent planning for autonomous vehicles in partially observed environments via parallel consensus optimization,”IEEE Transactions on Intelligent Transportation Systems, 2025

    work page 2025