Pith. sign in

REVIEW 5 major objections 8 minor 8 cited by

Adaptive Field Effect Planner for Safe Interactive Autonomous Driving on Curved Roads

T0 review · 5 major / 8 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read This paper claims that an adaptive risk field, Frenet coordinates, and an improved particle swarm optimizer can together produce safe, comfortable lane changes on curved roads, and supports the claim with simulations of a two-lane…

desk verdict A readable integration of known planner components, but the Frenet dynamic transformation is wrong and the promised real-world validation is absent. read the letter →

arxiv 2504.14747 v1 pith:FHDT4JXR submitted 2025-04-20 eess.SY cs.SY

Qinghao Li , Zhen Tian , Xiaodan Wang , Jinming Yang , Zhihao Lin This is my paper · ORCID
classification eess.SYcs.SY
keywords autonomousdrivinginteractivecurvyroadriskfieldquinticpolynomialcurveparticleswarmoptimizationFrenetcoordinatestrajectoryplanning
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 single planner can handle interactive lane changes on curved roads by combining three components usually used separately: an artificial-potential-field risk field for safety, Frenet coordinates for road-aligned motion, and an improved particle swarm optimizer (IPSO) for trajectory selection. The proposed planner decides when to change lanes by comparing repulsive, attractive, and lane-change risk fields, generates smooth candidate paths with quintic polynomials in the Frenet frame, and picks the best path under comfort and vehicle-dynamics constraints. Simulations on a two-lane road with inner radius 64 m and outer radius 70 m show the ego vehicle completing lane changes while maintaining safety margins and keeping lateral acceleration below 0.4g. The authors also report that IPSO converges in about 0.004 seconds, faster than standard PSO, a genetic algorithm, and an active-set method. If the framework works as described, it offers a parameter-based alternative to learning-based planners for curved-road interactive driving.

What carries the argument

The central object is the combined risk-field-Frenet-IPSO planner. The risk field is a sum of three potential functions: lane-keeping attraction Ua, front-vehicle repulsion Ub, and lane-change field Uc with a feasibility function Φ(Ξ). The Frenet frame, using arc length s and lateral offset d, converts curved-road planning into a road-aligned problem, and the quintic polynomial y(t) = a0 + a1 t + a2 $t^{2}$ + a3 $t^{3}$ + a4 $t^{4}$ + a5 $t^{5}$ supplies smooth candidate lateral trajectories. IPSO then searches the polynomial parameter space using an adaptive inertia weight w(t) and time-varying cognitive and social coefficients c1(t) and c2(t), with a penalty function handling constraints. The mechanism that carries the argument is the coupling: the risk fields decide whether and when to change lanes, and the Frenet-based quintic generator plus IPSO decides how, balancing safety and comfort within dynamic limits.

What would settle it

Run a round-trip test on the published transformation: take a known circular-arc trajectory of radius 64 m with a known lateral offset, convert the global state to Frenet coordinates with Eq. (6), convert back with Eqs. (7)-(8), and compare with the original global state. If the round-trip error grows with curvature, or if the sign of the lateral offset does not match a point whose left or right side is known, the transformation is not correct and the reported trajectories are not truly road-aligned.

Watch

Extended reading notes

Core claim

The central claim is that interactive lane changing on curved roads can be solved by an integrated framework in which a dynamic risk field, adapted from artificial potential fields, quantifies the threat from front, rear, and adjacent vehicles and triggers lane changes when three conditions hold: the front repulsive field exceeds its threshold, the lane-change risk field is below its threshold, and the attraction field favors the adjacent lane. In the Frenet frame of the reference path, a quintic polynomial generates smooth lateral-position profiles from boundary conditions, and IPSO with adaptive inertia and learning factors minimizes a weighted cost of terminal position, lateral acceleration, yaw rate, side-slip, and steering angle subject to dynamic constraints. According to the paper, this yields safe, comfortable, dynamically feasible lane changes on a 64-70 m radius curved road, with faster convergence than PSO, GA, and Active Set methods.

Load-bearing premise

The load-bearing premise is that the Frenet coordinate transformation in Section II-D, specifically Eq. (6), correctly maps global positions, velocities, and accelerations to road-aligned coordinates on curved reference paths; if that transformation is wrong or numerically unstable, every trajectory produced by the downstream optimizer is invalid.

Editorial extensions

If this is right

  • The planner can complete interactive lane changes on a curved two-lane road with inner radius 64 m and outer radius 70 m while keeping lateral acceleration within 0.4g and yaw rate within friction limits.
  • Lane-change timing adapts to surrounding traffic: when the rear vehicle is faster than the adjacent vehicle, the repulsive field from the closing gap grows and triggers an earlier lane change.
  • The IPSO optimizer with adaptive inertia and time-varying learning factors converges to a feasible trajectory in about 0.004 seconds, faster than standard PSO (0.005 s), GA (0.038 s), and Active Set (0.052 s) on the tested case.
  • Because the risk fields are recomputed from the current vehicle states, the framework can adjust lane-change intention online as the front, rear, and adjacent vehicles change speed or position.

Reading between the lines

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

  • Implicit in the paper: because the risk-field layer is agnostic to the road's curvature except through the Frenet frame, the same trigger logic should transfer to non-circular roads, such as highway curves or mountain roads, once a reference path is given; the paper only demonstrates circular geometry.
  • A testable extension: replace the fixed threshold values in the lane-change triggering condition with adaptive thresholds learned from traffic density or time-to-collision, and measure whether earlier or later triggers reduce the cost function; the paper does not specify how the thresholds were set.
  • The convergence comparison is measured on a single scenario with one small difference between IPSO and PSO (0.004 vs 0.005 s); an independent check on multiple scenarios with random initializations would show whether the adaptive parameters, rather than chance, produce the speed-up.
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.

Referee Report

5 major / 8 minor

Summary. This paper proposes an integrated trajectory planning framework for autonomous lane changes on curved two-lane roads, combining an adaptive risk field based on artificial potential fields, a Frenet-frame quintic polynomial trajectory generator, and an improved particle swarm optimizer (IPSO). The paper claims that the framework maintains safety margins and comfort while converging faster than PSO, GA, and Active Set methods, and states in the abstract that it is validated through extensive simulations and real-world scenarios. Section V, however, reports only two hand-picked Matlab test cases plus one qualitative wider-road case and a computation-time bar chart, with no real-world experiments.

Significance. If the framework worked as described, it would be a useful engineering contribution to interactive lane-change planning on curved roads, particularly the idea of a dynamic risk field that adjusts lane-change intention based on surrounding vehicle behavior. The paper's clear decomposition of the pipeline into risk field, Frenet formulation, quintic generation, and IPSO is a strength. However, the reported results are not reproducible without numerous omitted parameter values and undefined terms, and the central dynamic transformation appears to be mathematically incorrect, so the claimed validation is not currently supported. The paper provides no code or machine-checkable proofs to offset these gaps.

major comments (5)
  1. [Section II-D, Eq. (6)] The dynamic Frenet coordinate transformation in Eq. (6) is mathematically incorrect. The standard Frenet kinematics with r = p(s) + d n(s) gives sddot = (a_t + 2 κ sdot ddot)/(1 - κ d) and dddot = a_n - κ sdot^2 (1 - κ d), where a_t = xddot cosψ + yddot sinψ and a_n = yddot cosψ - xddot sinψ. Equation (6) instead adds κ sdot^2 to the sddot numerator and subtracts only κ sdot^2 from the dddot term, omitting the 2 κ sdot ddot coupling and the (1 - κ d) factor. For uniform circular motion at constant speed with d = 0, a_t = 0 but the printed formulas yield nonzero sddot and dddot. Since the feasibility constraints in Eq. (19) and the cost function in Eq. (24) are evaluated in this frame, the claim of dynamically feasible trajectories is not supported. Additionally, the lateral-offset sign in Eq. (5) uses sign(xdot yddot - xddot ydot), which is not the geometric side of the reference path; it vanishes for a stationary vehicle and flips when the velocity direction reverses at the same position.
  2. [Section V and Abstract] The abstract states that the framework is 'validated through extensive simulations and real-world scenarios', but Section V contains only two hand-picked Matlab test cases, one qualitative wider-road example with no quantitative measures, and a single computation-time bar chart. No real-world experiments, sensor data, statistical analysis, or comparison of trajectory safety and comfort against baseline planners are reported. The load-bearing validation claim in the abstract is therefore unsupported by the evidence in the manuscript.
  3. [Sections III-IV, Eq. (27) and Eq. (14)] The model depends on many parameters that are never specified: α, β, γ0, σ0, λ, ξ, vsafe, u_b,threshold, u_c,threshold, the cost weights w1..w5, and the IPSO bounds and learning-factor endpoints. Moreover, the lane-change feasibility function Φ(Ξ) introduced in Eq. (14) is never defined. Because the lane-change trigger conditions in Eq. (27) rely on thresholds that are not given, the reported lane-change times and safety margins are consequences of undisclosed tunings rather than independent predictions, making the validation circular and the paper unreproducible.
  4. [Section V, Fig. 6] The claim that IPSO converges faster than PSO, GA, and Active Set is not substantiated. Fig. 6 reports a single bar per algorithm with no variance or confidence intervals, the label 'Case 3' does not correspond to either of the two test cases listed in the text, and no convergence criterion or tolerance is stated. The reported difference between IPSO (0.004 s) and PSO (0.005 s) is likely within run-to-run noise, and no statistical test is provided. There is also no end-to-end comparison of the planner's safety or comfort performance against baseline planners, so the central planning performance claim is not evaluated.
  5. [Section II-E, Eq. (19)] The dynamic feasibility constraints in Eq. (19) include limits on lateral acceleration, yaw rate, side-slip angle, and steering angle, but the manuscript provides no vehicle model or mapping between the quintic polynomial trajectory and the quantities δ, β, and ψdot. Without such a mapping, the penalty-based constraint handling in Section IV-D cannot demonstrate that the generated trajectories satisfy Eq. (19), and the 'dynamically feasible' characterization in the abstract is not justified.
minor comments (8)
  1. [Section V, first paragraph] There is a typo 'iniital' for 'initial', and 'radium' should be 'radius'; the symbol 'S2' used in the analysis of Case 1 is never defined.
  2. [Section V, text around Figs. 1-2] The text says 'Figs. 1(a) through (d) show the smooth curves of velocity, acceleration, steering angle, and yaw rate' and 'Fig. 2(a) illustrates the trajectories', but the figure captions show the opposite assignment: Fig. 1 contains trajectories and field values, while Fig. 2 contains the smooth curves. The figure references appear to be swapped.
  3. [Section III-A, Eq. (9)] The weighting matrix W in the norm term ∥Xego - Xref∥^2_W is not defined, and the symbols α and β are reused with different meanings in Eq. (12), which is confusing.
  4. [Section IV-C, Eq. (24)] The term x_te,max in J1 is never defined, and the objective of maximizing terminal position for a lane-change maneuver is ambiguous; it should be stated whether this encodes progress, target-lane completion, or something else.
  5. [Section II-E, Eq. (19)] The steering-angle limit δmax = 2° is extremely restrictive for a lane change on a road with a 64-m inner radius, and no source or justification is provided for this constraint.
  6. [Section V, Fig. 5 and text] The time points t=10,20,30,40,50 in Fig. 5 appear to be simulation steps, but the text reports 'lane-changing time is 18' without stating whether this is seconds or time steps; the simulation timestep is never defined.
  7. [Section V, Fig. 6] The caption refers to 'Case 3', which is not among the two test cases defined earlier in the section, and the comparison lacks a dedicated GA citation (Ref. [23] is a trajectory-planning paper, not a genetic-algorithm reference).
  8. [Section IV-E, Eq. (27)] Lower-case u_a, u_b, and u_c are used in the trigger conditions without formal definition; the text should state explicitly that these denote the field values U_A, U_B, and U_C evaluated at the ego state.

Circularity Check

1 steps flagged · score 4.0 of 10

Lane-change timing is threshold-crossing by construction; no load-bearing self-citation found.

  1. fitted input called prediction [Eq. (27) and Section V, Case 1/2 discussion]
    "A lane change is initiated when: ub > ub,threshold, uc < uc,threshold, ua(current) > ua(adjacent) (Eq. 27). ... The lane-changing time is 18, which is earlier than in the previous cases. This is due to the higher speed of the RV compared to the IV ... The earlier lane-changing time demonstrates the adaptability of the proposed framework."

    Eq. (27) defines the moment of lane-change initiation as the first time the three threshold inequalities hold. Therefore the reported lane-change time is, by construction, the crossing time of the paper's own hand-set thresholds; it is not an independent prediction of the framework. Presenting that timing as 'demonstrating adaptability' is restating the trigger definition rather than validating it. The thresholds and field coefficients are never reported, so the behavior is indistinguishable from a pre-set trigger rule.

full rationale

The paper's central derivation chain (risk-field decision, Frenet transformation, quintic generator, IPSO selection) is a closed-loop planner rather than a first-principles prediction; most of it is self-consistent rather than circular. No load-bearing self-citation was found: works such as [21] are background citations for Frenet coordinates, and no uniqueness theorem is imported from the authors' prior work. The Frenet transformation in Eq. (6) is nonstandard and potentially invalid dynamically, but that is a correctness risk, not a circularity. The one clear circularity is in the validation of lane-change adaptability: Eq. (27) defines a lane change as the crossing of three thresholds, so the reported lane-change time is the threshold-crossing time by construction. Reporting that time as evidence of adaptability restates the trigger definition; the thresholds and coefficients are unreported, so the result cannot be independently checked. This is partial self-generated validation, but the IPSO/PSO/GA/Active-Set convergence comparison and the trajectory-generation mechanics have independent content, so the score is moderate rather than severe.

Assumptions & free parameters 8 free parameters · 5 assumptions · 1 invented entities

The central claim depends on many unpublished coefficients and thresholds, and the safety logic is asserted through a superposition of hand-designed fields. There is no external benchmark, formal proof, or reproduced artifact to ground the framework.

free parameters (8)
  • alpha in Eq. (9), attraction field strength
    Controls lane-keeping pull; no value or tuning procedure given.
  • gamma0 in Eq. (12) and weights alpha, beta
    Baseline repulsion strength and velocity-dependent weights; values not reported.
  • sigma0 in Eq. (13), spatial decay
    Sets how far the front-vehicle repulsion reaches; value not reported.
  • lambda in Eq. (14), lane-change field strength
    Scales the lane-change risk field; value not reported.
  • xi in Eq. (14), angular scaling factor
    Weights angular differences in the lane-change field; value not reported.
  • Lane-change thresholds u_b,threshold and u_c,threshold in Eq. (27)
    Determine when a lane change is triggered; values not reported.
  • Cost weights w1 to w5 in Eq. (23)
    Balance terminal position, lateral acceleration, yaw rate, side-slip, and steering; values not reported.
  • IPSO bounds wmin, wmax, c1,start, c1,end, c2,start, c2,end
    Adaptive inertia and learning factor endpoints in Eq. (20); values not reported.
assumptions (5)
  • domain assumption Reference path is smooth and the nearest-point projection in Eq. (5) is unique.
    The Frenet transformation depends on a unique (xp,yp); high curvature or ambiguous projection would break it.
  • ad hoc to paper Superposition U_total = U_A + U_B + U_C is a valid safety measure.
    No safety proof; the fields are designed to produce the desired behavior.
  • ad hoc to paper Lane-change trigger conditions in Eq. (27) are sufficient and not overly aggressive.
    Thresholds are asserted; no formal safety analysis or calibration against observed driving.
  • standard math The quintic boundary-value matrix M in Eq. (18) is invertible.
    Standard interpolation result for t_e != t_s; not discussed but implicit.
  • domain assumption IPSO finds a near-optimal trajectory within the fixed iteration budget.
    Non-convex cost in Eq. (23); no convergence guarantee, only one empirical timing.
invented entities (1)
  • Lane change feasibility function Phi(Xi)
    purpose: Gates the lane-change risk field in Eq. (14) to decide whether a lane change is safe.
    Never defined or evaluated; carries the safety decision but has no falsifiable content.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Field Effect Planner for Safe Interactive Autonomous Driving on Curved Roads." pith.science (2026). https://pith.science/paper/FHDT4JXR

@misc{pith2026250414747,
  author       = {Pith},
  title        = {Pith review of: Adaptive Field Effect Planner for Safe Interactive Autonomous Driving on Curved Roads},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FHDT4JXR}},
  note         = {Machine review of arXiv:2504.14747}
}
read the original abstract

Autonomous driving has garnered significant attention for its potential to improve safety, traffic efficiency, and user convenience. However, the dynamic and complex nature of interactive driving poses significant challenges, including the need to navigate non-linear road geometries, handle dynamic obstacles, and meet stringent safety and comfort requirements. Traditional approaches, such as artificial potential fields (APF), often fall short in addressing these complexities independently, necessitating the development of integrated and adaptive frameworks. This paper presents a novel approach to autonomous vehicle navigation that integrates artificial potential fields, Frenet coordinates, and improved particle swarm optimization (IPSO). A dynamic risk field, adapted from traditional APF, is proposed to ensure interactive safety by quantifying risks and dynamically adjusting lane-changing intentions based on surrounding vehicle behavior. Frenet coordinates are utilized to simplify trajectory planning on non-straight roads, while an enhanced quintic polynomial trajectory generator ensures smooth and comfortable path transitions. Additionally, an IPSO algorithm optimizes trajectory selection in real time, balancing safety and user comfort within a feasible input range. The proposed framework is validated through extensive simulations and real-world scenarios, demonstrating its ability to navigate complex traffic environments, maintain safety margins, and generate smooth, dynamically feasible trajectories.

Figures

Figures reproduced from arXiv: 2504.14747 by the authors.

Figure 3
Figure 3. Results of the decision-making for Case 2: (a) the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 2
Figure 2. Results of the lane-changing for Case 1: (a) the curve [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 5
Figure 5. Results of the decision making and lane-changing in a [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figures from the paper (1 more)
Figure 6
Figure 6. Figure 6: Comparison of computational time to convergence for [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Enhanced Mean Field Game for Interactive Decision-Making with Varied Stylish Multi-Vehicles

    cs.RO 2025-08 reject novelty 5.0 of 10

    A mean-field-game lane-change planner with six hand-set driving styles claims collision-free simulation results, but its promised NGSIM calibration and baseline comparisons do not appear in the body.

  2. Mean Field Game-Based Interactive Trajectory Planning Using Physics-Inspired Unified Potential Fields

    cs.RO 2025-09 reject novelty 4.0 of 10

    An autonomous driving planner that merges mean-field game interactions and potential fields into one unified field, claiming Nash equilibrium convergence but providing an incomplete proof and limited simulation evidence.

  3. Attention and Risk-Aware Decision Framework for Safe Autonomous Driving

    cs.RO 2025-09 reject novelty 4.0 of 10

    An improved PPO for highway driving uses a hybrid risk field, attention modules, a balanced reward, and a rule-based safety filter to reduce collisions and speed up training in simulation.

  4. Adaptive Evolution Factor Risk Ellipse Framework for Reliable and Safe Autonomous Driving

    cs.RO 2025-09 reject novelty 4.0 of 10

    An adaptive risk-field-plus-MPC controller with a sigmoid evolution factor and TTC/TWH-based risk ellipses is claimed to achieve collision-free overtaking and lane changes in simulation.

  5. A Risk-aware Spatial-temporal Trajectory Planning Framework for Autonomous Vehicles Using QP-MPC and Dynamic Hazard Fields

    cs.RO 2025-08 reject novelty 4.0 of 10

    Integrates a dynamic hazard field into QP-MPC trajectory planning and reports smoother, safer lane changes, overtaking, and intersection crossings in simulation.

  6. Safe and Efficient Lane-Changing for Autonomous Vehicles: An Improved Double Quintic Polynomial Approach with Time-to-Collision Evaluation

    cs.RO 2025-08 reject novelty 4.0 of 10

    An optimization-based double quintic lane-change planner that penalizes low time-to-collision is proposed and tested in simulation.

  7. Leveraging Vision-Language Large Models for Interpretable Video Action Recognition with Semantic Tokenization

    cs.CV 2025-09 reject novelty 3.0 of 10

    LVLM-VAR transforms video into 'semantic action tokens' and uses a LoRA-tuned vision-language model to classify actions and generate explanations, reporting 94.1% on NTU RGB+D X-Sub.

  8. Scenario-based Decision-making Using Game Theory for Interactive Autonomous Driving: A Survey

    cs.RO 2025-09 reject novelty 1.0 of 10

    A scenario-based survey of game-theoretic autonomous driving decision-making that claims comprehensiveness but is undermined by a non-systematic methodology and numerous internal errors.

Reference graph

Works this paper leans on

24 extracted references · 18 canonical work pages · cited by 8 Pith papers

  1. [1]

    A review on autonomous vehicles: Progress, methods and challenges,

    D. Parekh, N. Poddar et al., “A review on autonomous vehicles: Progress, methods and challenges,” Electronics, vol. 11, no. 14, p. 2162, 2022

    work page 2022

  2. [2]

    Decision-making method of autonomous vehicles in urban environments considering traffic laws,

    X. Lu et al., “Decision-making method of autonomous vehicles in urban environments considering traffic laws,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 11, pp. 21 641–21 652, 2022

    work page 2022

  3. [3]

    A conflicts-free, speed-lossless kan-based reinforcement learning decision system for interactive driving in roundabouts,

    Z. Lin et al. , “A conflicts-free, speed-lossless kan-based reinforcement learning decision system for interactive driving in roundabouts,” arXiv preprint arXiv:2408.08242, 2024

    arXiv 2024

  4. [4]

    Self-driving cars: A survey,

    C. Badue et al. , “Self-driving cars: A survey,” Expert Systems with Applications, vol. 165, p. 113816, 2021

    work page 2021

  5. [5]

    Enhanced visual slam for collision-free driving with lightweight autonomous cars,

    Z. Lin et al. , “Enhanced visual slam for collision-free driving with lightweight autonomous cars,” Sensors, vol. 24, no. 19, p. 6258, 2024

    2024

  6. [6]

    Autonomous driving manoeuvres in urban road traffic environment: A study on roundabouts,

    J. P ´erez et al. , “Autonomous driving manoeuvres in urban road traffic environment: A study on roundabouts,” vol. 18, 08 2011

    work page 2011

  7. [7]

    Fdnet: Fourier transform guided dual-channel underwater image enhancement diffusion network,

    Z. Zhu et al., “Fdnet: Fourier transform guided dual-channel underwater image enhancement diffusion network,” Science China Technological Sciences, vol. 68, no. 1, p. 1100403, 2025

    2025

  8. [8]

    Slam2: Simultaneous localization and multimode mapping for indoor dynamic environments,

    Z. Lin et al., “Slam2: Simultaneous localization and multimode mapping for indoor dynamic environments,” Pattern Recognition , vol. 158, p. 111054, 2025

    2025

Show all 24 references
  1. [9]

    Podb: A learning-based polari- metric object detection benchmark for road scenes in adverse weather conditions,

    Z. Zhu, X. Li, J. Zhai, and H. Hu, “Podb: A learning-based polari- metric object detection benchmark for road scenes in adverse weather conditions,” Information Fusion, vol. 108, p. 102385, 2024

    2024

  2. [10]

    Efficient and balanced exploration-driven decision making for autonomous racing using local information,

    Z. Tian et al. , “Efficient and balanced exploration-driven decision making for autonomous racing using local information,” IEEE Trans. Intell. Veh., 2024

    2024

  3. [11]

    An efficient high-risk lane-changing scenario edge cases generation method for autonomous vehicle safety testing,

    S. Jing et al., “An efficient high-risk lane-changing scenario edge cases generation method for autonomous vehicle safety testing,” IEEE Trans. Intell. Veh., pp. 1–13, 2024

    2024

  4. [12]

    Evaluating scenario-based decision-making for interactive autonomous driving using rational criteria: A survey,

    Z. Tian et al., “Evaluating scenario-based decision-making for interactive autonomous driving using rational criteria: A survey,” arXiv preprint arXiv:2501.01886, 2025

    2025 arXiv

  5. [13]

    A multi-vehicle game-theoretic framework for decision making and planning of autonomous vehicles in mixed traffic,

    Y . Yan et al. , “A multi-vehicle game-theoretic framework for decision making and planning of autonomous vehicles in mixed traffic,” IEEE Trans. Intell. Veh., vol. 8, no. 11, pp. 4572–4587, 2023

    2023

  6. [14]

    Driving style recognition for intelligent vehicle control and advanced driver assistance: A survey,

    C. M. Martinez et al., “Driving style recognition for intelligent vehicle control and advanced driver assistance: A survey,” IEEE Trans. Intell. Transp. Syst., vol. 19, no. 3, pp. 666–676, 2017

    2017

  7. [15]

    A comprehensive review of recent advances in automated guided vehicle technologies: Dynamic obstacle avoidance in complex environment toward autonomous capa- bility,

    M. Aizat, N. Qistina, and W. Rahiman, “A comprehensive review of recent advances in automated guided vehicle technologies: Dynamic obstacle avoidance in complex environment toward autonomous capa- bility,” IEEE Transactions on Instrumentation and Measurement , 2023

    2023

  8. [16]

    Emergency obstacle avoidance trajectory planning method of intelligent vehicles based on improved hybrid a,

    G. Chen et al. , “Emergency obstacle avoidance trajectory planning method of intelligent vehicles based on improved hybrid a,” SAE International Journal of Vehicle Dynamics, Stability, and NVH , vol. 8, no. 10-08-01-0001, pp. 3–19, 2023

    2023

  9. [17]

    Assessing and explaining collision risk in dynamic environments for autonomous driving safety,

    R. Nahata, D. Omeiza, R. Howard, and L. Kunze, “Assessing and explaining collision risk in dynamic environments for autonomous driving safety,” in 2021 IEEE international intelligent transportation systems conference (ITSC) . IEEE, 2021, pp. 223–230

    2021

  10. [18]

    Standards for passenger comfort in automated vehicles: Acceleration and jerk,

    K. N. de Winkel, T. Irmak, R. Happee, and B. Shyrokau, “Standards for passenger comfort in automated vehicles: Acceleration and jerk,” Applied Ergonomics, vol. 106, p. 103881, 2023

    2023

  11. [19]

    An improved artificial potential field method for path planning and formation control of the multi-uav systems,

    Z. Pan, C. Zhang, Y . Xia, H. Xiong, and X. Shao, “An improved artificial potential field method for path planning and formation control of the multi-uav systems,” IEEE Transactions on Circuits and Systems II: Express Briefs , vol. 69, no. 3, pp. 1129–1133, 2021

    2021

  12. [20]

    An integrated framework of decision making and motion planning for autonomous vehicles considering social behaviors,

    P. Hang et al., “An integrated framework of decision making and motion planning for autonomous vehicles considering social behaviors,” IEEE Trans. Veh. Technol., vol. 69, no. 12, pp. 14 458–14 469, 2020

    2020

  13. [21]

    Research on path planning for autonomous vehicle based on frenet system,

    Y . Wang and Z. Lin, “Research on path planning for autonomous vehicle based on frenet system,” Journal of engineering research, vol. 11, no. 2, p. 100080, 2023

    2023

  14. [22]

    Lf-net: A learning-based frenet planning approach for urban autonomous driving,

    Z. Yu and oithers, “Lf-net: A learning-based frenet planning approach for urban autonomous driving,” IEEE Trans. Intell. Veh., 2023

    2023

  15. [23]

    Optimal trajectory planning and robust tracking using vehicle model inversion,

    S. Victor, J.-B. Receveur et al., “Optimal trajectory planning and robust tracking using vehicle model inversion,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 5, pp. 4556–4569, 2022

    2022

  16. [24]

    An active-set algorithm for norm constrained quadratic problems,

    N. Rontsis, P. J. Goulart, and Y . Nakatsukasa, “An active-set algorithm for norm constrained quadratic problems,” Mathematical Programming, vol. 193, no. 1, pp. 447–483, 2022

    2022

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.