Development of an Optimal Ramp Metering Control Strategy for I-12 LSU Research Team Sherif S. Ishak, Osama Osman, Saleh Mousa, Sogand Karblaieali, Peter Bakhit
2 Outline Introduction Objectives Study area Methodology Data collection Model development and calibration Results and analysis Conclusions
3 Introduction First use in Louisiana on I-12 beginning June 2010 (Essen lane) and ending November 2010 16 ramp meters along I-12 between Essen Ln. and Walker/LA Operation: Fixed time (2 sec green/red), WB (6:00-9:00 am), EB (3:00-7:00 pm) Previous evaluation study did not show any improvements; construction and poor data quality This study presents a comprehensive evaluation of various adaptive ramp metering strategies
5 Main Objective Identify the optimal ramp metering strategy (combination of strategies) for I-12 on-ramps with the objective of alleviating congestion and improving the operational characteristics of the corridor.
6 Study Area Ramp meter locations
Methodology
8 Methodology Data Collection Geometric Data Traffic Data Build Simulation Model Current Ramp Metering Strategy Coding Model Calibration Test the Different Ramp Metering Strategies Comparative Analysis & Evaluation Selection of the Optimal strategy
9 Selected Ramp Metering Strategies 1. Fixed Time Strategy (current implementation) 2. ALINEA Local Ramp Metering Strategy 3. HERO (Coordinated Ramp Metering Strategy)
12 Simulation Scenarios Fixed (Base Case) ALINEA Mixed (ALINEA + HERO)
Proposed Locations for different RM strategies 13
14 Measures of Performance (MOPs) Travel time Speed Vehicle Hours Traveled (VHT)
15 Performed Analysis Corridor level analysis: ANOVA analysis -significant difference in the performance Post-hoc (Tukey) analysis best performing strategy Section level analysis Post-hoc (Tukey) analysis sections benefited the most from each strategy
Model Development and Calibration
17 Data Collection 1- Traffic Counts ATMS software page 2- Travel Times BlueTOAD data - Calibration
Simulation Model 18
19 Model Calibration Simulation model operates at conditions close to the actual conditions Fixed Time Queue Override Performance measures: 1. Traffic flow 2. Speed profiles
Traffic Volumes (Veh/hr) 20 Model Calibration-Traffic Flow For the Eastbound direction, the difference in volumes does not exceed 2.0% 7000 6000 5000 Actual results Approved Trial 4000 3000 2000 1000 0 Ramp Locations
Traffic Volume (Veh/hr) 21 Model Calibration-Traffic Flow (cont d) For the Westbound direction, the difference in volumes does not exceed 4.0% 7000 6000 5000 Actual results Approved Trial 4000 3000 2000 1000 0 Ramp Locations
Speed (mph) 22 Model Calibration-Speed For the Eastbound direction, the speed difference does not exceed 14.0% 70 60 50 40 30 Approved Trial Actual Speed 20 10 0 Essen Jefferson Airline Sherwood Millerville Oneal Range Juban Walker Ramp Locations
Speed (mph) 23 Model Calibration-Speed (cont d) For the Westbound direction, the speed difference does not exceed 15.0% 80 70 60 50 40 30 20 10 Approved Trial Actual Speed 0 Essen Jefferson Airline Sherwood Millerville O'Neal Range Juban Walker Ramp Locations
Results and Analysis
28 Simulation Results Output from 20 simulation runs. Three performance measures Speed Travel time Vehicles hours travelled (VHT) Analysis performed on two different levels: Corridor level analysis Sectional level analysis
29 Speed Corridor level analysis ANOVA test was performed to compare between the different strategies (at 5% level of significance) Eastbound direction, no statistically significant change in the operating speed was realized Westbound direction, analysis results show statistically significant change
30 Speed Corridor level analysis Westbound direction, to determine the strategy with the lowest travel times, Post Hoc tests (Tukey) were performed p-values for ANOVA pairwise comparison (level of significance, α=5%) Strategy Fixed- Time ALINEA Mixed Fixed-Time 0.209 0.001 ALINEA 0.209 0.079 Mixed 0.001 0.079
Speed (mph) Speed (mph) 31 Speed Section level analysis 70 70 60 60 50 50 40 40 30 20 10 0 Fixed-Time ALINEA Mixed 30 20 10 0 Fixed-Time ALINEA Mixed Freeway Segments Along I-12 Eastbound Direction Freeway Segments Along I-12 Westbound Direction
32 Speed Section level analysis : p-values for ANOVA pairwise comparison (at 5% significant level) Strategy Range-O neal Fixed- Time ALINEA Mixed Fixed-Time <0.001 0.009 ALINEA <0.001 0.101 Mixed 0.009 0.101 Strategy O neal-millerville Fixed- Time ALINEA Mixed Fixed-Time 0.725 <0.001 ALINEA 0.725 0.005 Mixed <0.001 0.005
Conclusions
Speed (mph) 42 Eastbound Direction Free flowing with the fixed-time strategy; as result, no further improvements could be achieved 70 60 50 40 30 20 10 Fixed-Time 0 Freeway Segments Along I-12 Analysis results showed that the current demand on the eastbound direction can be controlled by the Fixed Time Control strategy
43 Westbound Direction ALINEA and mixed strategies showed significant improvements compared to the fixed-time control Three main locations benefited the most from both strategies, (Range-O neal, O neal-millerville and Millerville-Sherwood) Performance of the two strategies was compared, the mixed control was superior to ALINEA local control Mixed Strategy is recommended for implementation
44 Implementation Plan H H A A A A H F H F A F F F F F A: ALINEA strategy H: HERO strategy F : Fixed strategy
45 ALINEA Implementation Downstream Traffic Conditions Queue Detection Vehicle Presence
46 HERO Implementation Downstream Traffic Conditions Queue & Arrival Rates Vehicle Presence Departure Rates
Thank You LSU Research Team