1. 1 Modeling Active Traffic Management for the I-80 Integrated Corridor Mobility (ICM) Project Terry Klim, P.E. Kevin Fehon, P.E. DKS Associates D

2. 2 Simulation of Active Traffic Management Elements n Adaptive Ramp Metering n Variable Speed Limits (VSL) n Incident Lane Management/Lane Control Signs

3. 3 The I-80 Corridor n 21-mile segment from Carquinez Bridge to Bay Bridge n Major parallel arterial (San Pablo Avenue)

4. Project Need n One of most congested corridors in Bay Area n Extensive congestion AM, PM and weekends n High accident rate along the freeway n Projected 30-45% demand increase by 2035 n Limited room for physical improvements 4

5. AM Peak Congestion Map 5

6. PM Peak Congestion Map 6

7. I-80 ICM Project 7

8. ATM Project Elements n Area-wide adaptive ramp metering u Meters installed at all ramps in both directions n Variable speed limits u Signs installed throughout corridor in both directions n Overhead lane control signals u Signs installed in westbound direction, within high- incident segment only 8

9. Simulation Approach n Corridor modeled using Paramics n Multi-hour AM (6 to 10 AM) and PM (2 to 7 PM) peak periods n Simulated both recurring and incident conditions n Involved application of: u New Ramp Metering and Lane Management plug-ins u Recently-added VSL module 9

10. Model Network n Over 25 miles of Freeway n 92 ramps n Over 50 miles of arterial roadway n 55 signalized intersections 10

11. Model Network (con’t) 11

12. Analysis Scenarios 12 2015 ConditionRecurring Congestion Incident Condition No-Build Ramp Metering Ramp Metering + Variable Speed Limit Ramp Metering + Variable Speed Limit + Lane Use Management

13. Adaptive Ramp Metering n Plan is for area-wide adaptive metering n Exact algorithm still being defined 13

14. Ramp Metering - Testing n Applied existing algorithm used by Caltrans District 4 (TOS 2.0) u “Locally adaptive” u Requires rate-vs-occupancy look-up tables u Includes “queue detection/override” capability n Metering active only WB in AM, both directions in PM 14

15. Ramp Metering – Plug-in n Developed by CLR Analytics n Requires user-supplied input file in XML format u Rate look-up tables u Start-up and shutdown control parameters u Queue override parameters 15

16. Ramp Meter – Modeling Insights n Additional network coding u Detectors u Splitting of multi-lane ramps n Development of metering rate-vs-occupancy look-up tables n Plug-in limitations: u Lack of automated “all green” u Number of meter lanes 16

17. Ramp Metering – Results n Ramp Metering provides improvement to the freeway without degrading arterial operation u I-80 hours of delay decrease by 20% in the westbound direction u Limited benefit eastbound due to pattern of congestion – most heavily congested at start of corridor u System wide hours of delay decreases by 8% u Negligible change in arterial hours of delay 17

18. Variable Speed Limits n General philosophy: u Modify upstream speed limits based on downstream conditions n Strategy objective evolved during study u Speed harmonization - slow traffic approaching bottlenecks to smooth speed profile and improve flow u End-of-queue warning during incidents 18

19. VSL - Testing n Examined as stand-alone “speed harmonization” strategy for AM westbound direction only u Tested over 30 VSL scenarios n Included in incident analysis for both directions and peak periods 19

20. VSL Module n Added to Version 6 at our request for this project n Allows for adjustment of speed limit based on lookup tables and monitored conditions n Requires coding of: u VSL routes - segments where VSL can be applied u VSL rules – if/then rules identifying under what conditions speed limits will be modified u Mainline detectors

21. VSL Module - Routes VSL Route 1 VSL Route 2 VSL Route 3 Detector Group 1 Detector Group 2 Detector Group 3

22. VSL Module - Rules Detector polling interval: allows the user to set how often data is gathered from loop detectors Transition time: the period of time to go from the current speed limit to the new target speed Settling time: the time that must elapse between two concurrent changes to speed limits

23. VSL – Modeling Insights n Module was easy and straightforward to use n Challenge: u Defining appropriate rules and parameters n Module limitations: u Inability to set segment-to-segment rules u Enforceable vs advisory speed limits – driver response u HOV Lane speed limit 23

24. VSL – Results n Produced an increase in average freeway travel time and delay u One scenario yielded a decrease n Reduced delay within congested, downstream segment n Key benefit is expected reduction in rear-end accidents 24

25. Incident Lane Management n General philosophy: u “Close” lane(s) upstream of incident u Use overhead gantry signs to provide warning of downstream incident/lane closure n Objective is to provide for earlier and smoother transition/lane changing in advance of incident

26. Incident Lane Management Sample Lane Control Sign and VSL combination 26

27. Lane Management - Testing n Focused on AM westbound n Involves: u Initiating incident directly within Paramics u “Activating” lane management after a short delay to simulate time to detect incident n Tested several incident scenarios 27

28. Lane Management – Plug-In n Developed by CLR Analytics n Requires XML input file u Lane closure zone/links u Decision zone/links in which drivers change lanes u CMS messages u Start/end time

29. Lane Management – Plug-In Model 29 Right two lanes closed ahead Right lane closed ahead Decision Link Closure Link Decision Link

30. Lane Management – Modeling Insights n Plug-in was easy and straightforward to use n Challenge: u Validation of results n Limitation: u “Zones” are defined by links rather than distance 30

31. Lane Management Results n Improved flow past incident location in all cases n Reduced delay on freeway, as well as arterials n Level of benefit varied significantly based on incident details 31

32. Institutional Challenges/Issues n Lack of ATM experience u Uncertainty about how to apply/use the ATM elements n Stakeholder skepticism n Desire for “exact” results 32

33. Conclusions n Paramics provided excellent vehicle to simulate active freeway management strategies n New VSL module and plug-ins worked as expected – but have limitations n ICM Project is moving forward with all ATM elements, although objectives have evolved 33

34. 34

35. 35

36. 36

37. 37

38. 38

39. 39

40. 40