pith. sign in

arxiv: 1907.10491 · v1 · pith:J37FPKG2new · submitted 2019-07-22 · 💻 cs.MA · eess.SP

Alternative Intersection Designs with Connected and Automated Vehicle

Zijia Zhong , Earl E. Lee This is my paper

Pith reviewed 2026-05-24 18:14 UTC · model grok-4.3

classification 💻 cs.MA eess.SP
keywords alternative intersection designsdiverging diamond interchangeconnected and automated vehiclestraffic simulationthroughputdriver confusiondiamond interchange
0
0 comments X

The pith

Redesigning a conventional diamond interchange as a diverging diamond raises throughput far more than adding connected and automated vehicles.

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

The paper tests whether alternative intersection layouts and connected-automated vehicle technology improve traffic flow when used separately or together. Microscopic simulations compare a conventional diamond interchange to a diverging diamond interchange and a restricted crossing U-turn design across different levels of vehicle automation. Results show that switching to the diverging diamond layout increases hourly throughput by about 950 vehicles, while full automation adds only 300 vehicles. Average vehicle delay follows the same pattern. Driver confusion at the new layouts produces a statistically detectable slowdown.

Core claim

Converting an existing conventional diamond interchange to a diverging diamond interchange improves the throughput of a CDI by 950 vehicles per hour, a near 20% improvement; whereas with full penetration of CAV, the throughput of a CDI is increased only by 300 vehicles per hour. A similar trend is observed in the average delay per vehicle as well. The negative impacts of driver's confusion are of statistical significance according to the ANOVA test.

What carries the argument

Microscopic traffic simulation that applies the diverging diamond interchange and restricted crossing U-turn layouts to a baseline conventional diamond interchange while varying connected and automated vehicle penetration rates.

If this is right

  • The diverging diamond layout produces larger capacity gains than connected and automated vehicle technology alone.
  • Driver confusion at alternative intersections reduces performance by a measurable amount.
  • Combining layout changes with vehicle automation yields additive benefits, but the geometry change accounts for most of the improvement.
  • The same ranking of interventions holds for delay as for throughput.

Where Pith is reading between the lines

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

  • Agencies could achieve quicker capacity relief by prioritizing geometric redesigns over waiting for high automation adoption.
  • The relative size of the two effects could help rank which existing interchanges should be converted first.
  • Real-world driver adaptation data collected after an actual conversion would test whether the modeled confusion penalty shrinks over time.

Load-bearing premise

The simulation model correctly captures how connected and automated vehicles actually drive and how human drivers react to unfamiliar intersection layouts.

What would settle it

A before-and-after field measurement of hourly vehicle throughput and average delay at an actual interchange before and after conversion to diverging diamond geometry, with and without automated vehicles present.

Figures

Figures reproduced from arXiv: 1907.10491 by Earl E. Lee, Zijia Zhong.

Figure 1
Figure 1. Figure 1: AID locations in contiguous U.S. (data source [ [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Configurations of selected DDI and RCUT convetional diamond interchange in early 2016 and open to trafic in late 2016 [22]. Four settings for DDI are simulated as shown in TABLE III. TABLE III: Simulation Cases for DDI Case CDI DDI AV MPR Base-CDI X 0% Base-DDI X 0% CAV-CDI X X 10-100% CAV-DDI X X 10-100% The arterial demand is assumed to be 3,000 vph for both westbound and eastbound direction. The traffic… view at source ↗
Figure 4
Figure 4. Figure 4: Average delay When it comes to RCUT, the flow-speed observations in three locations (diverging, upstream, and downstream) are shown in [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 3
Figure 3. Figure 3: Network throughput The average delay for each vehicle is plotted in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Flow-speed curve observed at the diverging area for [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Flow-speed curve observed at the diverging area for [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗
Figure 6
Figure 6. Figure 6: Impact of driver’s confusion The speed-flow diagram of the diverging area of the RCUT network is shown in [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
read the original abstract

Alternative intersection designs (AIDs) can improve the performance of an intersection by not only reducing the number of signal phases but also change the configuration of the conflicting points by re-routing traffic. However the AID studies have rarely been extended to Connected and Automated Vehicle (CAV) which is expected to revolutionize our transportation system. In this study, we investigate the potential benefits of CAV to two AIDs: the diverging diamond interchange (DDI) and the restricted crossing U-turn intersection. The potential enhancements of AID, CAV, and the combination of both are quantified via microscopic traffic simulation. We found that CAV is able to positively contribute to the performance of an intersection. However, converting an existing conventional diamond interchange (CDI) to a diverging one is a more effective way according to the simulation results. DDI improves the throughput of a CDI by 950 vehicles per hour, a near 20% improvement; whereas with full penetration of CAV, the throughput of a CDI is increased only by 300 vehicles per hour. A similar trend is observed in the average delay per vehicle as well. Furthermore, we assess the impact for the driver's confusion, a concern for deploying AIDs, on the traffic flow. According to the ANOVA test, the negative impacts of driver's confusion are of statistical significance.

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

2 major / 2 minor

Summary. The manuscript uses microscopic traffic simulation to compare alternative intersection designs (diverging diamond interchange DDI and restricted crossing U-turn RCUT) against a conventional diamond interchange (CDI), both with and without connected and automated vehicles (CAV). It claims that converting CDI to DDI raises throughput by 950 vehicles per hour (nearly 20%), outperforming the 300 vph gain from full CAV penetration on the original CDI; similar trends hold for average delay. ANOVA indicates that driver confusion has statistically significant negative effects on flow.

Significance. If the simulation model is shown to be realistic, the results would suggest that geometric redesign of interchanges can deliver substantially larger capacity gains than CAV adoption alone, with direct implications for prioritizing infrastructure changes versus technology deployment in transportation planning.

major comments (2)
  1. [Abstract] Abstract: The central quantitative claim (DDI improves CDI throughput by 950 vph while full CAV penetration improves it by only 300 vph) rests entirely on the microscopic simulation outputs, yet the manuscript supplies no description of the CAV behavioral parameters (headway, reaction time, lane-changing logic at 100% penetration), their sources, calibration procedure, or validation against field data; without these the 950-vs-300 differential cannot be assessed for robustness.
  2. [Results section (ANOVA)] Results section (ANOVA on driver confusion): Statistical significance is asserted but no effect sizes, F-statistics, degrees of freedom, or sensitivity checks to the confusion parameters are reported; this leaves the practical magnitude of the negative impact unquantified and weakens the claim that confusion is a deployability concern for AIDs.
minor comments (2)
  1. [Abstract] Abstract: 'assess the impact for the driver's confusion' contains a preposition error and should read 'of the driver's confusion'.
  2. [Throughout] Throughout: Throughput and delay results are given as point values with no error bars, standard deviations, or indication of the number of simulation replications, reducing interpretability of variability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on the simulation methodology and statistical reporting. We address each major comment below and will revise the manuscript to enhance transparency and completeness.

read point-by-point responses

  1. Referee: [Abstract] The central quantitative claim (DDI improves CDI throughput by 950 vph while full CAV penetration improves it by only 300 vph) rests entirely on the microscopic simulation outputs, yet the manuscript supplies no description of the CAV behavioral parameters (headway, reaction time, lane-changing logic at 100% penetration), their sources, calibration procedure, or validation against field data; without these the 950-vs-300 differential cannot be assessed for robustness.

    Authors: We agree that a dedicated description of the CAV parameters is required for assessing robustness. In the revised manuscript we will add a new subsection in the Methods section specifying the headway, reaction time, and lane-changing parameters used at 100% penetration (based on the Wiedemann model in VISSIM with reduced headways drawn from prior CAV literature). Calibration followed standard procedures from the VISSIM manual adjusted per established studies; we will cite the exact sources. Full field validation for 100% CAV is not possible given the current lack of real-world data at that penetration level, and we will explicitly note this limitation while emphasizing that the 950-vs-300 comparison is intended as a relative simulation-based indicator rather than an absolute prediction. revision: yes

  2. Referee: [Results section (ANOVA)] Statistical significance is asserted but no effect sizes, F-statistics, degrees of freedom, or sensitivity checks to the confusion parameters are reported; this leaves the practical magnitude of the negative impact unquantified and weakens the claim that confusion is a deployability concern for AIDs.

    Authors: We concur that the ANOVA reporting is incomplete. The revised Results section will report the full ANOVA table including F-statistics, degrees of freedom, p-values, and effect sizes (partial eta-squared). We will also add a sensitivity analysis varying the confusion parameter values and report the resulting changes in throughput to quantify practical significance. These additions will directly support the deployability discussion. revision: yes

Circularity Check

0 steps flagged

No circularity: throughput and delay results are direct simulation outputs

full rationale

The paper reports throughput gains (DDI +950 vph, CAV +300 vph) and delay metrics obtained from VISSIM microscopic simulations under varying geometries and CAV penetration levels. These quantities are generated by running the simulator on input traffic demands and vehicle behavior parameters; they are not obtained by solving equations whose right-hand sides contain the same quantities, nor by fitting parameters to a subset of the reported results and then re-presenting the fit as a prediction. The ANOVA on driver confusion is likewise a post-simulation statistical test on the generated delay and throughput samples. No self-citation chain, uniqueness theorem, or ansatz is invoked to justify the central comparison. The derivation chain is therefore self-contained against external simulation benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The central claims rest on standard traffic-flow modeling assumptions and scenario parameters chosen for the simulation; no new physical entities are introduced.

free parameters (2)
  • CAV penetration levels = 100%
    Full (100%) penetration used as an upper-bound scenario for comparison.
  • Driver confusion parameters
    Modeled to produce measurable flow reduction in the ANOVA test.
axioms (1)
  • domain assumption Microscopic simulation sufficiently captures comparative traffic dynamics for the chosen intersections and vehicle mixes.
    Invoked to justify all quantitative throughput and delay claims.

pith-pipeline@v0.9.0 · 5752 in / 1307 out tokens · 26501 ms · 2026-05-24T18:14:54.625869+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

  1. [1]

    Performance analysi s of parallel flow intersection and displaced left-turn intersection des igns,

    A. Dhatrak, P . Edara, and J. Bared, “Performance analysi s of parallel flow intersection and displaced left-turn intersection des igns,” Trans- portation Research Record: Journal of the Transportation R esearch Board, no. 2171, pp. 33–43, 2010

    work page 2010

  2. [2]

    A lternative Intersections and Interchange

    Institue for Transportation Research and Education, “A lternative Intersections and Interchange.” [Online]. Available: https://www.google. com/maps/d/u/0/viewer?mid=1MMZKi0RdQqjZjtBxmSTznUVnurY

    work page

  3. [3]

    V ehicl e-infrastructure integration (VII) initiative benefit-cost analysis versio n 2.3 (draft),

    V olpe National Transportation Systems Center, “V ehicl e-infrastructure integration (VII) initiative benefit-cost analysis versio n 2.3 (draft),” 2008

    work page 2008

  4. [4]

    Should the diverging diamond interchang e always be considered a diamond interchange form?

    G. Chlewicki, “Should the diverging diamond interchang e always be considered a diamond interchange form?” Transportation Research Record, vol. 2223, no. 1, pp. 88–95, 2011

    work page 2011

  5. [5]

    Development of a s ignal optimization model for diverging diamond interchange,

    X. Y ang, G.-L. Chang, and S. Rahwanji, “Development of a s ignal optimization model for diverging diamond interchange,” Journal of Transportation Engineering, vol. 140, no. 5, p. 4014010, 2014

    work page 2014

  6. [6]

    Alterna- tive intersections/interchanges: informational report ( AIIR),

    W. Hughes, R. Jagannathan, D. Sengupta, and J. E. Hummer, “Alterna- tive intersections/interchanges: informational report ( AIIR),” Tech. Rep., 2010

    work page 2010

  7. [7]

    Evaluation of Grand River Ave nue (M-5/M-102) Safety Improvement Project Before and After St udy,

    M. Scheuer and K. L. Kunde, “Evaluation of Grand River Ave nue (M-5/M-102) Safety Improvement Project Before and After St udy,” Michigan Department of Transportation, Lansing , 1996

    work page 1996

  8. [8]

    The Operational and Safety Aspects of the Michigan Design for Divided Highways,

    S. Castronovo, P . W. Dorothy, M. C. Scheuer, T. L. Maleck, and Others, “The Operational and Safety Aspects of the Michigan Design for Divided Highways,” V olume I, Michigan State University College of Engineering, 1995

    work page 1995

  9. [9]

    Evaluations of Manage d Lane Strategies for Arterial Deployment of Cooperative Adaptiv e Cruise Control,

    Z. Zhong, L. Joyoung, and L. Zhao, “Evaluations of Manage d Lane Strategies for Arterial Deployment of Cooperative Adaptiv e Cruise Control,” in 96th Transportation Research Board Annual Meeting , Washington DC, USA, 2017

    work page 2017

  10. [10]

    Automated car s: Queue discharge at signalized intersections with Assured-Clear -Distance- Ahead’driving strategies,

    S. Le Vine, X. Liu, F. Zheng, and J. Polak, “Automated car s: Queue discharge at signalized intersections with Assured-Clear -Distance- Ahead’driving strategies,” Transportation Research Part C: Emerging Technologies, vol. 62, pp. 35–54, 2016

    work page 2016

  11. [11]

    Isolated interse ction control for various levels of vehicle technology: Conventional, co nnected, and automated vehicles,

    K. Y ang, S. I. Guler, and M. Menendez, “Isolated interse ction control for various levels of vehicle technology: Conventional, co nnected, and automated vehicles,” Transportation Research Part C: Emerging Tech- nologies, vol. 72, pp. 109–129, 2016

    work page 2016

  12. [12]

    Operational perfo rmance comparison of four unconventional intersection designs us ing micro- simulation,

    J. Autey, T. Sayed, and M. El Esawey, “Operational perfo rmance comparison of four unconventional intersection designs us ing micro- simulation,” Journal of Advanced Transportation , vol. 47, no. 5, pp. 536–552, 2013

    work page 2013

  13. [13]

    Upstream signalized c rossover inter- section: optimization and performance issues,

    T. Sayed, P . Storer, and G. Wong, “Upstream signalized c rossover inter- section: optimization and performance issues,” Transportation Research Record: Journal of the Transportation Research Board , no. 1961, pp. 44–54, 2006

    work page 1961

  14. [14]

    Di- verging diamond interchange performance evaluation (I-44 and Route 13),

    V . Chilukuri, S. Siromaskul, M. Trueblood, T. Ryan, and Others, “Di- verging diamond interchange performance evaluation (I-44 and Route 13),” Missouri. Dept. of Transportation, Tech. Rep., 2011

    work page 2011

  15. [15]

    Travel time comparisons between seven unconventional arterial intersection designs,

    J. Reid and J. Hummer, “Travel time comparisons between seven unconventional arterial intersection designs,” Transportation Research Record: Journal of the Transportation Research Board , no. 1751, pp. 56–66, 2001

    work page 2001

  16. [16]

    FHW A tech brief: D rivers’ evaluation of the diverging diamond interchange,

    J. Bared, T. Granda, and A. Zineddin, “FHW A tech brief: D rivers’ evaluation of the diverging diamond interchange,” W ashington DC: Turner-Fairbank Highway Research Center . Tech Brief. FHWA-HRT-07- 048, 2007

    work page 2007

  17. [17]

    Transportation safety modeling and evalua tion: alternative geometric designs, enforcement, and airfield applications ,

    B. Claros, “Transportation safety modeling and evalua tion: alternative geometric designs, enforcement, and airfield applications ,” 2017

    work page 2017

  18. [18]

    Enhanced intel ligent driver model to access the impact of driving strategies on traffic ca pac- ity,

    A. Kesting, M. Treiber, and D. Helbing, “Enhanced intel ligent driver model to access the impact of driving strategies on traffic ca pac- ity,” Philosophical Transactions of the Royal Society A: Mathema ti- cal,Physical and Engineering Sciences , vol. 368, no. 1928, pp. 4585– 4605, 2010

    work page 1928

  19. [19]

    Simulation des Straenverkehrsflusses,

    R. Wiedemann, “Simulation des Straenverkehrsflusses, ” Ph.D. disserta- tion, Karlsruhe, 1974

    work page 1974

  20. [20]

    Modelling of rti-elements on multi-lane roads,

    ——, “Modelling of rti-elements on multi-lane roads,” i n Drive Confer- ence (1991: Brussels, Belgium) , vol. 2, 1991

    work page 1991

  21. [21]

    Congested Tra ffic States in Empirical Observations and Microscopic Simulations,

    M. Treiber, A. Hennecke, and D. Helbing, “Congested Tra ffic States in Empirical Observations and Microscopic Simulations,” pp. 1805–1824, 2000

    work page 2000

  22. [22]

    Sr 72/sr 1 diverging diamond interchange,

    “Sr 72/sr 1 diverging diamond interchange,” https://www.deldot.gov/ information/projects/SR72-SR1/index.shtml, accessed: 2019-05-01

    work page 2019

  23. [23]

    A policy on geometric desig n of highways and streets,

    M. W. Hancock and B. Wright, “A policy on geometric desig n of highways and streets,” 2013

    work page 2013

  24. [24]

    Advanced statistics: boo tstrapping confidence intervals for statistics with difficult distribu tions,

    J. S. Haukoos and R. J. Lewis, “Advanced statistics: boo tstrapping confidence intervals for statistics with difficult distribu tions,” Academic emergency medicine, vol. 12, no. 4, pp. 360–365, 2005

    work page 2005

  25. [25]

    Pairwise comparisons for one-way anova,

    “Pairwise comparisons for one-way anova,” https://support.minitab. com/en-us/minitab/18/help-and-how-to/modeling-stati stics/anova/ how-to/one-way-anova/interpret-the-results/ all-statistics-and-graphs/ pairwise-comparisons/, accessed: 2018-09-30

    work page 2018