15th Scientific and Technical Conference Transport Systems Theory and Practice Department of Transport Systems and Traffic Engineering, Faculty of Transport, Silesian University of Technology, Katowice, Poland, Sep 2018 Gap Acceptance Cycles for Modelling Roundabout Capacity and Performance Rahmi Akçelik
Paper Content and Disclaimer Paper Content Introduction Modelling Roundabout Capacity by Gap Acceptance Cycles Back of Queue and Cycle Average Queue A Simple Roundabout Example Concluding Remarks The author is the developer of the SIDRA INTERSECTION model used in the study presented in this paper. 2 of 24
Gap-acceptance theory beyond capacity modelling Gap-acceptance theory has been used widely for estimation of CAPACITY at roundabouts and sign-controlled (stop and give-way) intersections. Models for estimating delay, queue length and stop rate for roundabouts and other unsignalised intersections were developed by the author using traffic signal analogy. This helps with the estimation of fuel consumption and emissions for unsignalised intersections. The models discussed in this paper have been implemented in the SIDRA INTERSECTION software. 2000-208 Akcelik & Associates Pty Ltd support.sidrasolutions.com 3 of 24
Gap-acceptance theory beyond capacity modelling This paper will describe the basic method that uses gap acceptance cycles for modelling performance measures with a focus on the modelling of queue length at roundabouts. The method is applicable to two-way sign control (give-way and stop) as well. Back of Queue vs Cycle Average Queue is discussed in detail. A simple single-lane roundabout example is given to explain important aspects of modelling the queue length. 4 of 24
Gap acceptance cycle This figure depicts modelling of gap acceptance cycles and its application to the modelling of capacity and performance at unsignalised intersections. Gap Acceptance Cycle Time Blocked Time Unblocked Time A gap acceptance cycle consists of blocked period (vehicles waiting due to lack of an acceptable gap) and unblocked period (vehicles departing when an acceptable gap occurs). This is similar to a signal cycle that consists of a red period and a green period. 5 of 24
Gap acceptance capacity Q = s x u s = 3600 / t f where Q = capacity (veh/h) u = unblocked time ratio (the proportion of time when the vehicles can depart from the queue) s = saturation flow rate (veh/h) t f = follow-up headway (seconds) (saturation headway) 6 of 24
Comparison of gap acceptance capacity models The increasing difference at high opposing flow rates is due to the bunching assumptions in headway distributions. Akçelik, R. (2007). A Review of Gap-Acceptance Capacity Models. 29th Conference of Australian Institutes of Transport Research (CAITR), University of South Australia, Adelaide, Australia. 7 of 24
Back of Queue and Cycle-Average Queue Back of queue is used commonly for modelling signalised intersection performance. General literature, various guidelines and traffic theory text books present only the cycle-average queue based on traditional gap acceptance and queuing theory models for unsignalised intersections. This discrepancy continues to exist in the signalised and unsignalised intersection chapters of US Highway Capacity Manual Edition 6. 8 of 24
Back of Queue for design of appropriate queuing space Back of queue is a more useful performance measure since it is relevant to the design of appropriate queuing space: short lane design to avoid queue spillback into adjacent lanes modelling blockage of upstream intersection lanes (queue spillback). Blocked Time Unblocked Time 9 of 24
Back of Queue and Delay Back of queue and delay are not necessarily consistent in terms of magnitude. Low delay associated with a long back of queue At signalised intersections, this is a result of large green time ratio and a high arrival flow rate. At roundabouts, this occurs with large unblocked time ratio (due to low circulating flow rate) and high entry flow rate. Blocked Time Unblocked Time 10 of 24
Back of Queue and Delay Back of queue and delay are not necessarily consistent in terms of magnitude. Large delay associated with a short back of queue At signalised intersections, this is a result of small green time ratio and a low arrival flow rate. At roundabouts, this occurs with low unblocked time ratio (due to large circulating flow rate) and low entry flow rate. 11 of 24
A simple roundabout example Example to demonstrate the relationship between back of queue and cycle-average queue and present the related aspects of modelling using gap acceptance cycles for varying entry and circulating flow rates. Results were obtained using the SIDRA INTERSECTION standard software setup for driving on the right-hand side of the road. 12 of 24
SIDRA capacity model treating roundabout as an interactive system The SIDRA Standard roundabout capacity model is based on analysis of a roundabout as a closed system with interactions among roundabout entries Capacity constraint Bunched headway distribution model for the circulating flow Lane balance of circulating flow rates Unbalanced flow conditions (OD pattern and queuing on approach roads) NOT as a series of T intersections 13 of 24
Driver behavior at roundabouts In the SIDRA Standard model, critical gap and follow-up headway values are reduced with increased circulating flow based on the Australian research. Dominant lane follow-up headway (s) 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 More capacity Di = 30, ne = 1 Di = 50, ne = 2 Di = 80, ne = 3 Shorter departure headways Dominant lane critical gap (s) 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 More capacity Di = 30, ne = 1 Di = 50, ne = 2 Di = 80, ne = 3 Observed (Dominant lanes) Shorter acceptable gaps 0.0 0 300 600 900 1200 1500 1800 2100 2400 2700 0.0 0 300 600 900 1200 1500 1800 2100 2400 2700 Circulating flow (pcu/h) Circulating flow (pcu/h) 14 of 24
Entry capacity as a function of the circulating flow rate for arrival flow rates of 300, 600 and 1000 veh/h Capacity differences for the three arrival flow rates for low circulating flow rates are due to the effect of the ratio of entry flow rate to the circulating flow rate (higher values of this ratio give higher capacities in the model). This is an important feature in modelling unbalanced flow conditions. Entry capacity (veh/h) 1600 1400 1200 1000 800 600 qa = 1000 qa = 600 qa = 300 400 0 200 400 600 800 1000 1200 Circulating Flow Rate (pcu/h) 15 of 24
Critical gap and follow-up headway Critical gap and follow-up headway values reduced with increased circulating flow rates for the single-lane roundabout example Follow-up Headway and Critical Gap (s) 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 Follow-up Headway (tf) 1.0 Critical Gap (tc) 0.5 0.0 200 300 400 500 600 700 800 900 1000 Circulating Flow Rate (pcu/h) 16 of 24
Gap acceptance cycle parameters Blocked and unblocked times and the gap acceptance cycle time as a function of the circulating flow rate for the case of arrival flow rate of 300 veh/h Gap Acceptance Cycle Parameters (s) 30 Gap Acceptance Cycle Time 25 Unblocked Time Blocked Time 20 15 10 5 0 200 300 400 500 600 700 800 900 1000 Circulating Flow Rate (pcu/h) 17 of 24
Comparison of Back of Queue and Cycle-Average Queue The difference between the values of average back of queue and cycle-average queue increase with increasing arrival flow rate. 25 25 Average Queue (veh) 20 15 10 5 Nb (qa = 300) Nc (qa = 300) Average Queue (veh) 20 15 10 5 Nb (qa = 1000) Nc (qa = 1000) 0 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 Degree of Saturation 0 0.60 0.70 0.80 0.90 1.00 1.10 Degree of Saturation Average Back of Queue and Cycle-Average Queue as a function of the degree of saturation. 18 of 24
Comparison of Back of Queue and Cycle-Average Queue The correlation of the average back of queue and cycle-average queue for arrival flow rates of 300, 600 and 1000 veh/h. The difference between the values of average back of queue and cycleaverage queue increases with increasing arrival flow rate. Average Back of Queue (veh) 25 20 15 10 5 qa = 1000 qa = 600 qa = 300 0 0 5 10 15 20 25 Cycle-Average Queue 19 of 24
Stopline Delay Stopline Delay values as a function of circulating flow rate. This is used to calculate the cycle-average queue: N c where = q x d q = arrival flow rate (veh/s) d = Stopline delay (seconds) Stopline Delay (s) 50 40 30 20 10 0 qa = 1000 qa = 600 qa = 300 0 200 400 600 800 1000 1200 Circulating Flow Rate (pcu/h) 20 of 24
Average Back of Queue Average Back of Queue values as a function of circulating flow rate Average Back of Queue (veh) 25 20 15 10 5 qa = 1000 qa = 600 qa = 300 0 0 200 400 600 800 1000 1200 Circulating Flow Rate (pcu/h) 21 of 24
Concluding Remarks The results given here are from an analytical model developed by the author and are for a simple single-lane roundabout case used for the purpose of this paper. Further research is recommended by means of microsimulation analysis and real-life surveys. The research should consider complications that arise in real-life situations including the effect of short lanes, variations in various geometric and driver behaviour parameters, slip lanes, effect of upstream signals and pedestrians. 22 of 24
References Papers and reports based on research by the author and colleagues can be downloaded from http://www.sidrasolutions.com/resources/articles 23 of 24
END OF PRESENTATION Thank you! Rahmi Akçelik www.sidrasolutions.com sidrasolutions.com youtube.com/sidrasolutions