The 2nd Conference on Traffic and Transportation Engineering, 2016, *, ** Published Online **** 2016 in SciRes. http://www.scirp.org/journal/wjet http://dx.doi.org/10.4236/wjet.2014.***** Evaluation and Improvement of the Roundabouts in Changchun Mo Zhou Department of Civil Engineering, National University of Singapore, Singapore 117576, Singapore Email: mollychou24@hotmail.com Received **** 2016 Abstract Among the grade crossing of city road in Changchun, roundabouts consist of a major part. They are often designed to connect the main roads due to their ability to eliminate or alter conflict points, making them safer than the traditional intersections. Also, they can beautify the city in a way no other crossings can. However, as a result of increasing traffic volume, the malfunction, or even paralysis of roundabouts is happening more and more frequent. It means that we need to evaluate the roundabouts on time, and make improvements to those whose performance are considered poor. By categorizing the roundabouts in Changchun, an example is selected from each kind. Through measuring the geometry condition, the V/C ratio in weaving sections, and delay, the poorly-performed roundabouts can be determined, and the enhancement proposals can be made according to the criteria. Subsequently, with the help of a micro-simulation software AIMSUN, a basic model is established using the OD data collected by field measurement, and error checking was then carried out. Afterwards, by putting forward a model parameter sensitivity analysis, the key parameters can be chosen. Then average queue length is applied to calibrate key parameters which influence the models, in order to ensure the minimum RMSE value between model output and field data. Finally the proposed improvements, therefore, can be tested based and assessed. Keywords Roundabouts; Evaluation; Micro-simulation; AIMSUN; Improvements 1. Introduction (Heading 1) Nowadays, several problems like the prolonged delay and increased queue length have gradually occurred in some roundabouts. Without timely evaluation and improvement on them, these small problems can get way worse. However, as for Changchun, who has at least three roundabouts in almost every arterial roads (as is marked with black five-pointed stars in Figure 1.1), malfunction of any single roundabout can be insufferable and disastrous. Therefore, it is highly necessary to assess the current running state of the roundabouts and evaluate proposed amendments by using micro-simulation. How to cite this paper: Mo Zhou (2016) Evaluation and Improvement of the Roundabouts in Changchun. *********, *, **-**. http://dx.doi.org/10.4236/wjet.2014.*****
Figure 1.1.Distribution Map of Roundabouts in Urban Areas of Changchun Research content and process can be briefly demonstrated in the technology map in Figure 1.2. Figure 1.2.Technology Map 2. Running Features of the Roundabouts in Changchun 2.1. Classification of Roundabouts According to the existence of traffic light and the number of connected lanes, the common roundabouts in Changchun can be classified into four types: 2
(1) Unsignalized four-leg roundabouts (eg. Weixing Square and Jingyang Square) (2) Signalized four-leg roundabouts (eg. Xi an Square) (3) Unsignalized five-leg roundabouts (eg. Zhengyang Square and Zhanqian Square) (4) Signalized five-leg roundabouts (eg. Nanhu Square and Xinmin Square) This thesis will choose an example from each types, and evaluate it, see if there s any hidden problem in it. 2.2. Basic Data Collection Detailed plan is shown as Table 2.1. Table 2.1. Data Collection Plan Location Time Content Weixing Square May 5 th Road Inventory and Peak Hour Flow 16:30-18:30 in Each Weaving Sections Weixing Square May 10 th Peak Hour Flow in Each Entrance 16:30-18:30 Sections and Turning Proportion Zhengyang Square th May 17 Road Inventory and Peak Hour Flow 16:30-18:30 in Each Weaving Sections Xinmin Square th May 18 Data needed for Individual Sample of 16:30-18:30 Stop Delay survey method Xi an Square th May 19 Data needed for Individual Sample of 16:30-18:30 Stop Delay survey method 3. Evaluation of the Roundabouts in Changchun 3.1. Unsignalized Four-leg Roundabouts (Weixing Square) As one of the most important roundabouts in Changchun, Weixing Square is selected as an example of this kind. First, the four weaving sections was marked as is shown in Figure 3.1. Figure 3.1.Weaving sections in Weixing Square The evaluation index of unsignalized multi-lane roundabout is the V/C ratio of weaving section. [1] After measuring the geometry data from digital map and collecting the peak hour flow, V/C ratio can be calculated with the help of the Index Calculator. Taking Weaving Section No.3 as an example, Figure 3.2 shows the user interface of the calculator. 3
Figure 3.2.User Interface of the Index Calculator V/C ratio is corresponding to level of service. [2] Therefore, the level of service of each weaving sections in Weixing Square can be determined. Table 3.1. The V/C Ratio and Level of Service of Each Weaving Sections in Weixing Square No. 1 2 3 4 V/C Ratio 0.72 0.95 0.77 0.97 Level of Service C F C F When the level of service of a weaving section reaches level F, severe congestion will occur, resulting in prolonged delay and increased queue length, which would definitely have a detrimental effect on the traffic flow, bringing hidden danger. As is shown in Table 3.1, service level of weaving section 2 and 4 have already reached F level. Consequently, the general running state of Weixing Square is unsatisfying, which means amendments should be made to this roundabout. 3.2. Unsignalized Five-leg Roundabout As for unsignalized roundabout, Zhengyang Square is chosen to be an example. Similarly, each weaving sections is marked in Figure 3.3. Figure 3.3.Weaving Sections in Zhengyang Square Also, by using the Index Calculator, V/C ratio and service level can be listed in Table 3.2. Table 3.2. The V/C Ratio and Level of Service of Each Weaving Sections in Zhengyang Square No. 1 2 3 4 5 V/C Ratio 0.72 0.95 0.77 0.97 0.89 Level of Service B E C B E 4
None of the service level reaches F, indicating that Zhengyang Square haven t shown any severe malfunction yet. The delay is relatively acceptable, and the conflicts are not so prominent. 3.3. Signalized Four-leg Roundabout Xi an Square, as one of the most important intersection in Changchun, is chosen as an example. Since the evaluation indexes of signalized roundabouts is the same as those of signalized cross intersection, [1] average delay is used to assess Xi an Square. When it comes to evaluation, delay is better virtually measured than estimated. So Individual Sample of Stop Delay survey method was used to collect data and calculate average [3] delay. Also according to HCM, service level can be corresponded. Table 3.3. The Average Delay and Level of Service of Each Entrance Sections in Xi an Square No. 1 2 3 4 Average Delay 26.3 19.0 39.8 38.4 Level of Service D C D D Based on Table 3.3, the traffic flow runs relatively smooth in Xi an Square. 3.4. Signalized Five-leg Roundabout Take one of the most famous roundabout in Changchun---Xinmin Square---as an example. In order to describe clearly, five legs are firstly marked with numbers in Figure 3.4. Figure 3.4. Five legs in Xinmin Square Similarly, average delay and service level of each sections can be calculated and listed as Table 3.4. Table 3.4. The Average Delay and Level of Service of Each Entrance Sections in Xinmin Square No. 1 2 3 4 5 Average Delay 34.2 26.7 23.7 37.8 36.4 Level of Service D D C D D As is shown above, traffic flow in each sections can be deemed steady. The average delay now is sufferable. Therefore, it is not urgent to make adjustments to this roundabout. In summary, Weixing Square, among the four mentioned roundabouts, particularly needs to be improved. 4. Micro-Simulation of the Roundabout 4.1. Basic Model Development Firstly, the base map is introduced into the new project of AIMSUN. 5
Figure 4.1. Base Model Based on the data that was collected before, traffic state settings can be filled in through calculation. In order to define OD matrix and allocate the generation and attraction, centroid configuration is also necessary. Figure 4.2 demonstrates the four centroid of this roundabout. Blue lines mean generation, while green lines represent attraction. In this picture, every single line corresponds with a specific percentage. Figure 4.2. Centroid Configuration After finishing the last step, OD matrix can be defined as Figure 4.3 and Figure 4.4 shows. Figure 4.3. OD Matrix of Car 6
Figure 4.4. OD Matrix of Bus Then related traffic demand and dynamic scenario can be created. By running the new experiment replica, AIMSUN is able to run the initial simulation. A screenshot (Figure 4.5) was taken during the simulation process. Figure 4.5. Screenshot of the Simulation Effect The last step was to carry out error checking, which includes three stages: software error checking, input data error checking and animation error checking. [4] After checking the software and input data, animation is supposed to be reviewed. Compared with the actual condition, the queue length in the simulation is slightly longer, which probably results from the give way to those who are entering the roundabout protocol. In addition, the running state in weaving section is slightly smoother than that in real-life scenario. This is because under the influence of temporal construction inside the roundabout, the central island is actually malformed, causing the number of lanes in some weaving sections less than 5. Hence, in general, the animation is very close to reality. 4.2. Model Parameter The Gipps model inside AIMSUN can be seen in an equation below. (1) Where: a n = max acceleration of vehicle n bn = max deceleration of vehicle n bn-1 = max deceleration of vehicle n-1 T.. = reaction time θ.. = reaction time to ensure safety = max expected speed of vehicle n in time t v n 7
v n (t). = speed of vehicle n in time t xn(t).= driven distance of vehicle n in time t Sn-1...= effective length of vehicle n-1 From the equation, the involved parameters are reaction time. speed limit, max acceleration, max deceleration, general deceleration, and so on. Take average queue length as the criterion to assess the sensitivity of the model, according to the pre-experiment, reaction time does not have a significant influence on average queue length, thus should not be considered as a parameter to be calibrated. Moreover, since the model involves roundabout, which contains give way protocol, max give-way time should be taken into consideration. After proposing appropriate value range of the parameters, the calibrating process can be carried out. In this case, the Root Mean Square Error method is used to get a minimum error value between the model measurements and the field measurements of V/C ratio. Through calculating under different range value, the minimum RMSE value can be found. The corresponding parameter value can therefore be determined. The calibrated parameters and the default ones are compared in Table 4.1. Table 4.1. Comparison of the Calibrated and the Default Parameters Parameter Default Calibrated average minimum maximum average minimum maximum Max Acceleration (m/s 2 ) 3 2.6 3.4 2.4 2.08 2.72 Max Decceleration (m/s 2 ) 6 5 7 5.4 4.5 6.3 Max Expected Speed(km/h) 110 80 150 99 72 135 Max Give-way Time (s) 10 5 15 8 4 12 Since there s no modification to the original model in the software, it is deemed unnecessary to carry out the validation process. 5. Simulation and Analysis of the Alternative Scenarios 5.1. Alternative Scenario 1: Addition of lane in weaving sections Add lanes to the north and south weaving section like Figure 5.1 shows. Figure 5.1. Addition of lane Then run the new experiment replica, record the V/C ratio of each weaving sections again. Compared to the field measurement, Table 5.1 can be obtained. 8
Table 5.1. Comparison of Alternative Scenario 1 and Real-life Scenario Weaving Section East North East South V/C Ratio of Realife Scenario 0.71 0.82 0.77 0.85 V/C Ratio of Alternative Scenario 1 0.67 0.73 0.75 0.74 Variation Range -5.08% -10.61% -2.22% -12.75% From table 5.1, it is obvious that V/C ratio in north and south weaving sections have dropped significantly, leading in the improvement of corresponding service level from F to C. Additionally, other indexes have also improved more or less as is shown in Table 5.2. Table 5.2. Variation Range of Other Indexes in Alternative Scenario 1 Index Queue Length Travel Time Delay Variation Range -1.39% -6.02% -23.17% 5.2. Alternative Scenario 2: Apply Signal Control Add lanes to the north and south weaving section like Figure 5.1 shows. As is shown in Figure 5.2, add fixed-time control to the base model. Figure 5.2. Apply Signal Control Create a new control plan and run this new experiment replica, and the comparison can be listed as Table 5.3. Table 5.3. Comparison of Alternative Scenario 2 and Real-life Scenario Weaving Section East North East South V/C Ratio of Realife Scenario 0.71 0.82 0.77 0.85 V/C Ratio of Alternative Scenario 2 0.42 0.65 0.50 0.68 Variation Range -40.68% -20.45% -36.40% -19.46% According to Table 5.3, the V/C ratio has dropped tremendously, while other indexes have deteriorated like Table 5.4 shows. Table 5.4. Variation Range of Other Indexes in Alternative Scenario 2 Index Queue Length Travel Time Delay Variation Range +55.96% +32.11% +89.50% All in all, signal control, who are generally considered as an effective method, is actually not suitable for this intersection, whereas addition of lane is much more appropriate. 9
Acknowledgements The guidance of two amazing supervisors is indispensable to the accomplishment of the thesis. One is my future supervisor Associate Professor Chin Hoong Chor from National University of Singapore, who broadened my comprehension to this topic, and introduced this great micro-simulation software to me. The other is my undergraduate thesis supervisor Yang Lili from Jilin University, who patiently answers my every questions and help me modify and refine this article over and over again. I couldn t finish this article without either of them. Therefore, I would like to express my sincere gratitude to them. References [1] The compilation group of Technical Guidelines for Traffic Impact Analysis of Construction Project (2010). Technical Guidelines for Traffic Impact Analysis of Construction Project. 1st Edition. [2] An Zizhen. (2008) Research on Capacity and Levels of Service for Roundabout. Master Thesis, Southeast University, China. [3] Li Qingyin, Sun Feng. (2011) Evaluation Indexes Selection and Level-of-service Classification for Roundabouts. Journal of Highway and Transportation Research and Development, 28, 131-135. [4] Septarina. (2012) Micro-Simulation of The Roundabout at Idrottsparken Using AIMSUN. Master Thesis, Linköping University, Sweden. 1 0