Pith. sign in

REVIEW 4 major objections 6 minor 37 references

Integrated Eco-Driving and Powertrain Optimization for Hybrid Vehicles in Complex Urban Traffic

T0 review · 4 major / 6 minor · reviewed 2026-07-31 · grok-4.5

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

Pith's one-line read One mixed-integer plan jointly optimizes hybrid-vehicle motion, lanes, intersections, and powertrain, cutting equivalent energy by about one third versus kinematics-only planning.

desk verdict Careful integrated MINLP that couples multi-lane urban rules with hybrid EMS; energy edge over KINO is real in-sim but rests partly on foresight and accounting choices. read the letter →

arxiv 2607.25114 v1 pith:DIW3DQKG submitted 2026-07-27 eess.SY cs.SYmath.OC

Milad Hasanzadeh , Amin Kargarian , Mehdi Farasat This is my paper · ORCID
classification eess.SYcs.SYmath.OC
keywords eco-drivinghybridvehiclesfinite-horizonplanningmixed-integeroptimizationurbandrivinglanechangingsignalizedintersectionspowertrainenergymanagement
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

Urban eco-driving forces a vehicle to choose speed, acceleration, lanes, safe gaps, intersection behavior, and energy use at the same time. Most prior work optimizes only some of those pieces and handles the rest separately. This paper builds a single finite-horizon mixed-integer planner for hybrid vehicles that decides longitudinal motion, lane occupancy and changes, car-following and lane-change safety, signalized and unsignalized intersection rules, and hybrid powertrain operation together. In 810 simulated urban cases the full planner stays fully feasible and about as mobile and comfortable as a kinematics-only optimizer, while using roughly one-third less equivalent energy—about 0.88 kWh on average, or 0.30 kWh/km—by allocating engine, motor, battery, and regeneration jointly with the driving plan. A sympathetic reader cares because the result is a reference trajectory that is already energy-aware, rule-compliant, and safe before any lower-level controller tracks it.

What carries the argument

The unified finite-horizon mixed-integer formulation (problem (36)): binary lane occupancy and lane-change variables, hard collision-avoidance plus soft nominal spacing, signalized entry/safe-to-clear/downstream-clearance constraints, unsignalized first-entry and priority-dependent stop/yield duration, discrete hybrid powertrain dynamics, and a four-level lexicographic objective (nominal safety, permission compliance, comfort, then equivalent energy and progress).

What would settle it

Rerun the same 810 cases (or a closed-loop receding-horizon version) with noisy or intentionally wrong front/rear trajectories, signal timing, or clearance flags; if success rate collapses or the energy advantage over kinematics-only disappears under realistic prediction error, the central claim fails.

Watch

Extended reading notes

Core claim

The authors show that embedding lane logic, hard and soft safety gaps, signal safe-to-clear and downstream-clearance rules, unsignalized stop-and-yield logic, and hybrid powertrain dynamics in one lexicographic mixed-integer finite-horizon problem yields urban plans that are 100% feasible across ten multi-lane scenarios and substantially more energy-efficient than rule-following, fixed-lane, overtaking-enabled, and kinematics-only baselines, without sacrificing mobility or comfort relative to the kinematics-only optimizer.

Load-bearing premise

Over the whole planning horizon the planner treats surrounding vehicles, lane availability, signals, and priority counts as known fixed inputs; if those forecasts are wrong, the claimed safety, legality, and energy gains need not hold.

Editorial extensions

If this is right

  • A single planner can supply speed, lane, and power-demand references that already satisfy urban rules and hybrid energy limits before tracking control.
  • Separating kinematics from powertrain leaves roughly one-third of equivalent energy on the table under the paper’s benchmark.
  • Fixed-lane or incomplete rule models fail on closures and unsignalized stops; the integrated rule set is required for full feasibility in the tested corridor mix.
  • Lexicographic ranking keeps energy savings from buying violations of nominal gaps, lane permissions, or comfort limits.
  • The main practical cost of integration is higher solve time (about 15 s per run versus under 1 s for kinematics-only).

Reading between the lines

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

  • Real deployment will need a receding-horizon wrapper and uncertainty-aware traffic prediction; the open-loop perfect-foresight plan is a reference generator, not a finished controller.
  • The same constraint block could be reused for battery-electric or fuel-cell powertrains by swapping only the energy submodel while keeping lane and intersection logic.
  • Computation time suggests warm-starting, horizon shortening, or decomposition may be needed before on-board use at one-second updates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper presents a finite-horizon mixed-integer nonlinear planning formulation for urban hybrid-vehicle eco-driving that jointly optimizes longitudinal motion, lane occupancy and lane changes, car-following and lane-change safety (hard minimum gaps plus soft nominal margins), signalized-intersection entry with safe-to-clear and downstream-clearance conditions, unsignalized stop-and-yield behavior, and a discrete hybrid powertrain model (engine on/off and power, motor motoring/generating, battery SOC, quadratic fuel map). A lexicographic objective (35g) ranks nominal safety, lane-permission compliance, comfort, and an eco-driving level combining fuel energy, battery discharge, smoothness, lane-change penalties, and forward progress. The method (Full) is compared against RF, HCS-i NLC, OEAC-i, and a kinematics-only variant (KINO) over 10 scenarios × 81 initial conditions (810 runs per method). Full and KINO achieve 100% success; Full averages ≈0.88 kWh equivalent energy (≈0.30 kWh/km) versus ≈1.30 kWh adjusted for KINO, a claimed ~one-third reduction at matched distance (~2.97 km) and comfort, at the cost of ~15 s average solve time.

Significance. If the results hold, the contribution is a genuinely unified, optimization-compatible treatment of urban driving logic (lane permissions, mandatory/emergency lane changes, safe-to-clear signal entry, downstream blocking, priority-based stop-and-yield) together with hybrid powertrain operation in one finite-horizon problem — a combination that, as the introduction argues, prior work treats only in subsets. The constraint set (28)–(34) is carefully and transparently specified (first-entry logic, big-M activation, lane-transition consistency), the simulation campaign is broad (4050 attempted runs with failures penalized rather than excluded), and the authors release complete code, data, and exported results in a public repository [37], which materially supports reproducibility. The KINO ablation is in principle the right instrument for isolating the value of joint motion-plus-powertrain optimization. However, the headline quantitative claim (~one-third energy reduction) currently rests on two accounting choices that the manuscript does not adequately pin down, and on a powertrain model in which the claimed regenerative-braking path appears unreachable; these must be resolved before th

major comments (4)
  1. [§IV-A and Eq. (39)] The central quantitative claim (Full ≈0.88 kWh vs KINO ≈1.30 kWh adjusted) depends on how KINO's energy is obtained, and this is not specified. §IV-A states only that for KINO 'hybrid-energy quantities are reconstructed from the returned motion trajectory using the same post-processing model.' It is nowhere stated whether this post-processor optimizes the engine/motor split for KINO's wheel-power profile (e.g., a DP or QP over (34i)–(34t) with the trajectory fixed) or applies a fixed rule-based EMS. KINO's reported values (108.45 g fuel, 0.004 kWh battery) are consistent with a near-engine-only rule. If so, the comparison measures 'optimized vs. unoptimized power split' rather than 'joint vs. sequential planning,' which is the stated contribution. Given that sequential schemes with an optimal EMS layer typically recover a large fraction of the joint benefit, the authors must (a) fully sp
  2. [Eq. (39), Table II (λb = 1), and §IV-B] The equivalent-energy metric credits gross battery discharge max{Pb,0} at λb = 1, i.e., 1 kWh of battery discharge is priced at 1 kWh of fuel LHV, while terminal SOC is held only within 0.005 of SOC_init (≈0.025 kWh net on a 5 kWh pack). In charge-sustaining operation, battery throughput that originated from engine generation costs roughly 1/(ηeηm) ≈ 2.5–3 kWh fuel-LHV per kWh; only the regenerative share is near-free. The paper never reports how Full's 0.129 kWh average discharge splits between regen and engine generation. The Full–KINO gap is 45.6 g fuel (≈0.545 kWh) minus 0.125 kWh additional battery use; a physically consistent charge-sustaining accounting could plausibly shrink the ~32% headline to single digits. The authors should report the regen/engine decomposition of battery throughput and provide a sensitivity of the Full–KINO gap to λb ∈ [1, 3] (or a net-SOC-corrected fuel-eq
  3. [§III-E, Eqs. (34h)–(34j) and (34n)] As written, the model contains no path for braking kinetic energy to reach the battery. Constraint (34n) imposes F_tr ≥ 0, so P_wh = F_tr·v ≥ 0 in (34h); the braking force F_br in (34d)/(34o) is purely dissipative and never enters the power balance (34i). Consequently P_m,− > 0 in (34i) can only be sourced from engine power exceeding wheel demand, i.e., engine-driven charging — regenerative braking from deceleration cannot occur, despite 'regenerative braking' being advertised in the abstract, §I, and §III-E. This is load-bearing in two ways: (i) a claimed model feature appears unrepresented; (ii) it implies that essentially all of Full's battery discharge is engine-generated, which sharpens Major Comment 2. The authors should either extend the model (e.g., signed wheel power with a regen path through P_m,−) or remove/correct the regenerative-braking claims and re-audit the energy accoun
  4. [§II and §III-G (problem (36))] Over the full N = 240 s horizon, surrounding-vehicle trajectories (13)–(14), lane availability/permissions, safe-to-clear indicators ΓSIG (26), downstream clearance Bclear (27), and priority counts qNS (19) are treated as known deterministic inputs. The safety, rule-compliance, and energy claims are therefore conditional on perfect 240-second foresight of exogenous traffic — a strong premise that is only implicitly acknowledged (the conclusion lists 'uncertainty-aware traffic prediction' as future work). Since the safety constraints (29)–(32) are activated by these predicted quantities, prediction error can invalidate the certified gaps. The manuscript should state this assumption prominently (abstract/§III-G), clarify which quantities are realistically obtainable from SPaT/V2X versus requiring prediction, and provide at least a limited robustness check (e.g., perturbed surrounding-vehic
minor comments (6)
  1. [Eq. (20)] The rule τ_NS = 1 + q_NS (with q_NS ∈ {0,1,2,3}, Δt = 1 s, so a maximum 4 s stop) is asserted rather than justified. A short rationale or citation for the one-second-per-priority-vehicle yielding model would help; as stated it reads as an ad-hoc environment axiom.
  2. [§III-F, Eq. (35f) vs. Eq. (34t)] The relationship between SOC_tar in (34t) and SOC_init in the quadratic charge-sustaining term of (35f) is unclear, as is the role of the 'charge-sustaining tolerance 0.005' in Table II (is (34t) SOC_N ≥ SOC_init − 0.005?). Please define SOC_tar explicitly and state which weights in (35f) are active in the reported experiments (e.g., is the optional w_dis term used, given it duplicates λb Eb,dis?).
  3. [§IV-A] HCS-i NLC and OEAC-i are 'adapted to the common benchmark and are not exact reproductions of the original methods.' Please say more about what was changed; otherwise the feasibility failures in Table I (e.g., 0/81 on B010) risk reflecting the adaptation rather than the published methods, and the comparison text should be tempered accordingly.
  4. [§IV-A / §IV-B] Numerical details needed for reproducibility: the big-M values used in (29)–(32), the lexicographic objective weights (wE, wCS, wa, wΔa, wL, wR, wss), and the distribution (not just the ~15 s mean) of Full solve times against the 300 s limit, including how many runs hit the time limit or the 10^-3 MIP gap.
  5. [Eq. (34b)/(34d)] The position update uses v_k while the velocity update uses forces at step k (forward Euler). With Δt = 1 s this is a coarse discretization for acceleration transients near stops; a sentence acknowledging the discretization error (or a smaller-Δt check) would be appropriate.
  6. [Presentation] Typographical/grammar items: mismatched quotes around 'Other Event' and 'late'/'mild' in §IV-A; 'thanks to integration of the vehicles dynamics' (§IV-C) should be 'vehicle dynamics'; Figures 5–7 are low-contrast in the preprint and the stacked fuel/battery bars in Fig. 7(b) would benefit from numeric labels; the GitHub access date in [37] is incomplete ('accessed: July , 2026').

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: constructive MIP eco-driving planner; energy results are post-solve simulation metrics, not inputs restated as predictions.

full rationale

This paper builds a finite-horizon mixed-integer planner (36) from explicit lane, safety, intersection, and hybrid-powertrain constraints (28)–(34) and a lexicographic objective (35), then evaluates it in closed simulation against RF, HCS-i NLC, OEAC-i, and KINO over 810 cases. Equivalent energy (39) is computed from solved fuel mass and battery discharge after optimization; it is not algebraically forced by a fitted constant renamed as a prediction. Design choices (gap parameters, weights, λb, SOC target) are stated inputs, not circular derivations. Self-citations ([21], [24], [27], [30], [32]) concern unrelated control/OPF/quantum topics and do not underwrite the eco-driving formulation or the Full-vs-KINO energy gap. Skeptical concerns about KINO post-processing fidelity or λb=1 accounting affect correctness of the comparison, not circularity of the derivation chain. No self-definitional loop, fitted-input-as-prediction, or load-bearing self-citation uniqueness claim is present.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

Claims rest on standard vehicle/hybrid dynamics, deterministic perfect foresight of traffic and signals, big-M mixed-integer encoding of rules, a quadratic fuel map, and hand-chosen lexicographic/objective weights and safety margins. No new physical entities. Energy superiority is a numerical consequence of solving that model, not a fitted law.

free parameters (6)
  • Lexicographic eco weights (wE, λb, wCS, wdis, wa, wΔa, wL, wR, wss)
    Fourth-level objective (35f) depends on these relative weights; manuscript does not report numeric values used in the 810 runs, yet they shape the reported energy–comfort–lane-change tradeoff.
  • Nominal/hard spacing and time-headway parameters (dcf_min, d0, Th, dF/R,*_min, T*_h, T*_rel)
    Safety constraints (29)–(30) are parameterized by these design constants; they define what ‘safe’ and ‘nominal’ mean and are not derived in-paper.
  • Comfort bounds amin, amax, jmax and soft slacks
    Soft acceleration/jerk limits (34f)–(34g) and Jdyn depend on chosen comfort envelopes.
  • Fuel map coefficients γ0, γ1, γ2
    Instantaneous fuel rate (34l) is a quadratic-in-power model with coefficients treated as known; values not tabulated.
  • Safe-to-clear / stopping parameters TSIG_r, bsafe and big-M constants
    Signal stopping feasibility (32h) and all big-M activations depend on chosen reaction/braking and M values that affect feasibility encoding.
  • Battery equivalence factor λb (=1 in Table II) and SOC_tar / charge-sustaining tolerance = λb=1; SOC tolerance 0.005
    Equivalent energy (39) and terminal SOC constraint directly use these; λb=1 is a modeling choice equating battery and fuel joules.
assumptions (7)
  • domain assumption Longitudinal motion is adequately captured by discrete double-integrator dynamics with quasi-steady resistance (aero, rolling, grade) and additive traction/brake forces (34b)–(34e).
    Standard point-mass longitudinal model; lateral dynamics during lane change are omitted except via occupancy binaries.
  • domain assumption Hybrid powertrain power balance, constant efficiencies ηdr, ηm, and quadratic engine fuel map suffice for planning-level energy (34h)–(34m), citing Sciarretta & Guzzella-style relations [35].
    Ignores transient thermal, gear, and detailed BSFC map effects.
  • domain assumption All road, traffic, signal, and priority data over the horizon are known deterministic parameters (Section II).
    Enables open-loop finite-horizon optimality; stated as environment inputs, not estimated online.
  • domain assumption Hard minimum gaps plus penalized nominal gaps, with body-length correction Δbody, are an adequate safety model; soft slacks never relax hard collision constraints.
    Core of (29)–(30); standard in MIP driving planners but not a formal safety proof under uncertainty.
  • ad hoc to paper Unsignalized required stop duration equals 1 + number of priority crossing vehicles (20), and signal entry is allowed only if precomputed ΓSIG=1 and Bclear allows.
    Specific encoding of stop-and-yield and safe-to-clear chosen for MIP compatibility; other traffic codes could differ.
  • ad hoc to paper Lexicographic order Jsafe ≻ Jperm ≻ Jdyn ≻ Jeco is the correct priority for urban eco-driving (35g).
    Design choice that prevents energy from buying safety violations; not empirically validated against human or legal preference weights.
  • domain assumption Mixed-integer nonlinear program with Gurobi nonconvex quadratic mode yields plans that represent attainable vehicle behavior when tracked.
    No lower-level tracking or actuator limits beyond force/power bounds are closed in the loop.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Integrated Eco-Driving and Powertrain Optimization for Hybrid Vehicles in Complex Urban Traffic." pith.science (2026). https://pith.science/paper/DIW3DQKG

@misc{pith2026260725114,
  author       = {Pith},
  title        = {Pith review of: Integrated Eco-Driving and Powertrain Optimization for Hybrid Vehicles in Complex Urban Traffic},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DIW3DQKG}},
  note         = {Machine review of arXiv:2607.25114}
}
read the original abstract

Urban eco-driving requires the simultaneous planning of speed, acceleration, lane decisions, surrounding-vehicle safety, intersection rules, and vehicle energy use. Existing studies commonly optimize only selected aspects of urban driving, such as longitudinal motion, lane changing, intersection crossing, or powertrain energy management, while treating the remaining decisions separately. Unlike these studies, this paper develops an integrated finite-horizon eco-driving planning framework for urban hybrid vehicles that jointly optimizes traffic behavior and powertrain operation within a unified mixed-integer formulation. The novelty of the proposed method lies in simultaneously modeling longitudinal motion, lane occupancy and changes, car-following and lane-change safety, signalized and unsignalized intersection behavior, and hybrid powertrain operation. The formulation captures lane availability, lane-dependent speed limits, mandatory and emergency lane changes, safe-to-clear signal conditions, downstream clearance, stop-and-yield rules, engine and motor power, battery state of charge, regenerative braking, and fuel consumption. Simulation studies compare the proposed method with rule-following, signal-aware, overtaking-enabled, and kinematic-only optimization baselines. The results demonstrate improved energy performance while maintaining feasible, safe, comfortable, and traffic-rule-compliant urban driving plans.

Figures

Figures reproduced from arXiv: 2607.25114 by the authors.

Figure 1
Figure 1. Multi-lane road data The raw traffic-light permission is represented by G SIG i,k ∈ {0, 1}, i ∈ ISIG, k ∈ K. (25) The value GSIG i,k = 1 means that the signal is permissive at step k, while GSIG i,k = 0 means that the signal is non-permissive. The raw signal state alone is not sufficient for safe intersection entry because the vehicle must also be able to clear the full signalized control zone before the permissive … view at source ↗
Figure 3
Figure 3. Unsignalized-intersection data [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Signalized-intersection data A. Proposed Lane Occupancy and Lane-Changing Constraints The proposed formulation uses binary variables to represent lane occupancy and lane-change actions. Let yk,ℓ = 1 if the host vehicle occupies lane ℓ at step k. Let z L k,ℓ = 1 denote a left lane change from lane ℓ to lane ℓ + 1, and let z R k,ℓ = 1 denote a right lane change from lane ℓ to lane ℓ − 1 during the interval from k to k… view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: Feasibility, adjusted energy, and mobility results [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Comfort, lane changes, and computation time [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Adjusted equivalent-energy intensity and energy composition [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 1 linked inside Pith

  1. [37]

    OSRBUD: Codes and results,

    LSU RAISE Laboratory, “OSRBUD: Codes and results,” https://github. com/LSU-RAISE-LAB/OSRBUD, 2026, accessed: July , 2026

    2026

  2. [1]

    A survey on the coordination of connected and automated vehicles at intersections and merging at highway on-ramps,

    J. Rios-Torres and A. A. Malikopoulos, “A survey on the coordination of connected and automated vehicles at intersections and merging at highway on-ramps,”IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 5, pp. 1066–1077, 2017

    2017

  3. [2]

    A survey of motion planning and control techniques for self-driving urban vehicles,

    B. Paden, M. ˇC´ap, S. Z. Yong, D. Yershov, and E. Frazzoli, “A survey of motion planning and control techniques for self-driving urban vehicles,” IEEE Transactions on Intelligent V ehicles, vol. 1, no. 1, pp. 33–55, 2016

    2016

  4. [3]

    Estimating vehicle fuel consumption and emissions based on instantaneous speed and acceleration levels,

    K. Ahn, H. Rakha, A. Trani, and M. Van Aerde, “Estimating vehicle fuel consumption and emissions based on instantaneous speed and acceleration levels,”Journal of Transportation Engineering, vol. 128, no. 2, pp. 182–190, 2002

    2002

  5. [4]

    Energy and emissions impacts of a freeway-based dynamic eco-driving system,

    M. Barth and K. Boriboonsomsin, “Energy and emissions impacts of a freeway-based dynamic eco-driving system,”Transportation Research Part D: Transport and Environment, vol. 14, no. 6, pp. 400–410, 2009

    2009

  6. [5]

    Control of connected and automated vehicles: State of the art and future challenges,

    J. Guanetti, Y . Kim, and F. Borrelli, “Control of connected and automated vehicles: State of the art and future challenges,”Annual Reviews in Control, vol. 45, pp. 18–40, 2018

    2018

  7. [6]

    Optimal eco-driving control of connected and autonomous vehicles through signalized intersections,

    C. Sun, J. Guanetti, F. Borrelli, and S. J. Moura, “Optimal eco-driving control of connected and autonomous vehicles through signalized intersections,”IEEE Internet of Things Journal, vol. 7, no. 5, pp. 3759–3773, 2020

    2020

  8. [7]

    Safe and eco-driving control for connected and automated electric vehicles using an analytical state-constrained optimal solution,

    J. Han, A. Sciarretta, L. L. Ojeda, G. De Nunzio, and L. Thibault, “Safe and eco-driving control for connected and automated electric vehicles using an analytical state-constrained optimal solution,”IEEE Transactions on Intelligent V ehicles, vol. 3, no. 2, pp. 163–172, 2018

    2018

Show all 37 references
  1. [8]

    Lane change and merge maneuvers for connected and automated vehicles: A survey,

    D. Bevly, X. Cao, M. Gordon, G. Ozbilgin, D. Kari, B. Nelson, J. Woodruff, M. Barth, C. Murray, A. Kurtet al., “Lane change and merge maneuvers for connected and automated vehicles: A survey,”IEEE Transactions on Intelligent V ehicles, vol. 1, no. 1, pp. 105–120, 2016

    2016

  2. [9]

    Overtaking-enabled eco-approach control at signalized intersections for connected and automated vehicles,

    H. Dong, W. Zhuang, G. Wu, Z. Li, G. Yin, and Z. Song, “Overtaking-enabled eco-approach control at signalized intersections for connected and automated vehicles,”IEEE Transactions on Intelligent Transportation Systems, vol. 25, no. 5, pp. 4527–4539, 2024

    2024

  3. [10]

    Predictive cruise control: Utilizing upcoming traffic signal information for improving fuel economy and reducing trip time,

    B. Asadi and A. Vahidi, “Predictive cruise control: Utilizing upcoming traffic signal information for improving fuel economy and reducing trip time,”IEEE Transactions on Control Systems Technology, vol. 19, no. 3, pp. 707–714, 2011

    2011

  4. [11]

    Review on eco-driving control for connected and automated vehicles,

    J. Li, A. Fotouhi, Y . Liu, Y . Zhang, and Z. Chen, “Review on eco-driving control for connected and automated vehicles,”Renewable and Sustainable Energy Reviews, vol. 189, p. 114025, 2024

    2024

  5. [12]

    Eco-driving control for CA Vs at signalized intersections: Adapting to traffic uncertainties,

    K. Lu, Y . Chen, Y . Tong, J. Zhang, Y . Luo, and J. Wang, “Eco-driving control for CA Vs at signalized intersections: Adapting to traffic uncertainties,”Transportation Research Part D: Transport and Environment, vol. 132, p. 104270, 2024

    2024

  6. [13]

    A decentralized energy-optimal control framework for connected automated vehicles at signal-free intersections,

    A. A. Malikopoulos, C. G. Cassandras, and Y . J. Zhang, “A decentralized energy-optimal control framework for connected automated vehicles at signal-free intersections,”Automatica, vol. 93, pp. 244–256, 2018

    2018

  7. [14]

    Cooperative eco-driving at signalized intersections in a partially connected and automated vehicle environment,

    Z. Wang, G. Wu, and M. J. Barth, “Cooperative eco-driving at signalized intersections in a partially connected and automated vehicle environment,” IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 5, pp. 2029–2038, 2020

  8. [15]

    Hierarchical energy-efficient control for CA Vs at multiple signalized intersections considering queue effects,

    S. Dong, H. Chen, B. Gao, L. Guo, and Q. Liu, “Hierarchical energy-efficient control for CA Vs at multiple signalized intersections considering queue effects,”IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 8, pp. 11 643–11 653, 2022

    2022

  9. [16]

    A generic approach to eco-driving of connected automated vehicles in mixed urban traffic and heterogeneous power conditions,

    Y . Hu, P. Yang, M. Zhao, D. Li, L. Zhang, S. Hu, W. Hua, W. Ji, Y . Wang, and J. Guo, “A generic approach to eco-driving of connected automated vehicles in mixed urban traffic and heterogeneous power conditions,” IEEE Transactions on Intelligent Transportation Systems, vol. 2...

    2023

  10. [17]

    Multilane automated driving with optimal control and mixed-integer programming,

    R. A. Dollar and A. Vahidi, “Multilane automated driving with optimal control and mixed-integer programming,”IEEE Transactions on Control Systems Technology, vol. 29, no. 6, pp. 2561–2574, 2021

    2021

  11. [18]

    Efficient and eco lane-changing trajectory planning for connected and automated vehicles: A deep reinforcement learning-based method,

    S. Jing, Y . Feng, F. Hui, J. Liu, X. Zhao, and A. J. Khattak, “Efficient and eco lane-changing trajectory planning for connected and automated vehicles: A deep reinforcement learning-based method,”IEEE Transactions on Intelligent Transportation Systems, 2025, early Access

    2025

  12. [19]

    Lane-changing-enabled eco-approach control for autonomous vehicles at a signalized intersection in a mixed-traffic environment,

    J. Liu, T. Qu, S. Dong, and B. Gao, “Lane-changing-enabled eco-approach control for autonomous vehicles at a signalized intersection in a mixed-traffic environment,”IET Intelligent Transport Systems, vol. 19, no. 1, p. e70036, 2025

    2025

  13. [20]

    Lane-changing-enabled eco-driving control for a plug-in hybrid electric vehicle under consecutive signalized-intersection conditions,

    Y . Jia, Y . Liu, Y . Zhang, Z. Chen, Z. Lei, and Y . Zhang, “Lane-changing-enabled eco-driving control for a plug-in hybrid electric vehicle under consecutive signalized-intersection conditions,”Green Energy and Intelligent Transportation, p. 100311, 2025

    2025

  14. [21]

    Distributed fixed-time rotating encirclement control of linear multi-agent systems with moving targets,

    M. Hasanzadeh, M. Baradarannia, and F. Hashemzadeh, “Distributed fixed-time rotating encirclement control of linear multi-agent systems with moving targets,”Journal of the Franklin Institute, vol. 361, no. 11, p. 106970, 2024

    2024

  15. [22]

    A survey of autonomous driving: Common practices and emerging technologies,

    E. Yurtsever, J. Lambert, A. Carballo, and K. Takeda, “A survey of autonomous driving: Common practices and emerging technologies,” IEEE Access, vol. 8, pp. 58 443–58 469, 2020

    2020

  16. [23]

    GlidePath: Eco-friendly automated approach and departure at signalized intersections,

    O. D. Altan, G. Wu, M. J. Barth, K. Boriboonsomsin, and J. A. Stark, “GlidePath: Eco-friendly automated approach and departure at signalized intersections,”IEEE Transactions on Intelligent V ehicles, vol. 2, no. 4, pp. 266–277, 2017

    2017

  17. [24]

    Encirclement control for PDE-based leader-follower multi-agent systems with targets in a sphere: Part II—rotation encirclement,

    M. Hasanzadeh and S.-X. Tang, “Encirclement control for PDE-based leader-follower multi-agent systems with targets in a sphere: Part II—rotation encirclement,” inProceedings of the 63rd IEEE Conference on Decision and Control (CDC). IEEE, 2024, pp. 7160–7165

    2024

  18. [25]

    A review of end-to-end autonomous driving in urban environments,

    D. Coelho and M. Oliveira, “A review of end-to-end autonomous driving in urban environments,”IEEE Access, vol. 10, pp. 75 296–75 311, 2022

    2022

  19. [26]

    Energy-optimal speed control for electric vehicles on signalized arterials,

    X. Wu, X. He, G. Yu, A. Harmandayan, and Y . Wang, “Energy-optimal speed control for electric vehicles on signalized arterials,”IEEE Transactions on Intelligent Transportation Systems, vol. 16, no. 5, pp. 2786–2796, 2015

    2015

  20. [27]

    ADMM enhancement techniques for distributed optimal power flow,

    M. Hasanzadeh and A. Kargarian, “ADMM enhancement techniques for distributed optimal power flow,”IEEE Transactions on Power Systems, vol. 41, no. 2, pp. 1321–1333, 2026

    2026

  21. [28]

    Prediction-based eco-approach and departure at signalized intersections with speed forecasting on preceding vehicles,

    F. Ye, P. Hao, X. Qi, G. Wu, K. Boriboonsomsin, and M. J. Barth, “Prediction-based eco-approach and departure at signalized intersections with speed forecasting on preceding vehicles,”IEEE Transactions on Intelligent Transportation Systems, vol. 20, no. 4, pp. 1378–1389, 2019

    2019

  22. [29]

    Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles,

    SAE On-Road Automated Driving Committee, “Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles,” SAE International, Tech. Rep. SAE J3016, 2021

    2021

  23. [30]

    Distributed implementation of variational quantum eigensolver to solve QUBO problems,

    M. Hasanzadeh and A. Kargarian, “Distributed implementation of variational quantum eigensolver to solve QUBO problems,”arXiv preprint arXiv:2508.17471, 2025

    2025 arXiv

  24. [31]

    A time- and energy-efficient routing algorithm for electric vehicles based on historical driving data,

    A. M. Bozorgi, M. Farasat, and A. Mahmoud, “A time- and energy-efficient routing algorithm for electric vehicles based on historical driving data,”IEEE Transactions on Intelligent V ehicles, vol. 2, no. 4, pp. 308–320, 2017

    2017

  25. [32]

    Dynamic average consensus as distributed PDE-based control for multi-agent systems,

    M. Hasanzadeh and S.-X. Tang, “Dynamic average consensus as distributed PDE-based control for multi-agent systems,” inProceedings of the 2024 European Control Conference (ECC). IEEE, 2024, pp. 828–833

    2024

  26. [33]

    Global EV outlook 2025,

    International Energy Agency, “Global EV outlook 2025,” International Energy Agency, Paris, France, Tech. Rep., May 2025

    2025

  27. [34]

    2025 EPA automotive trends report,

    U.S. Environmental Protection Agency, “2025 EPA automotive trends report,” U.S. Environmental Protection Agency, Tech. Rep. EPA-420-R-26-001, Feb. 2026

    2025

  28. [35]

    Control of hybrid electric vehicles,

    A. Sciarretta and L. Guzzella, “Control of hybrid electric vehicles,”IEEE Control Systems Magazine, vol. 27, no. 2, pp. 60–70, 2007

    2007

  29. [36]

    Gurobi optimizer reference manual,

    Gurobi Optimization, LLC, “Gurobi optimizer reference manual,” [Online]. Available: https://www.gurobi.com, 2026, accessed: Jul. 21, 2026

    2026

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.