modeling the effects of bus stops on bicycle traffic flow by cellular...

Research ArticleModeling the Effects of Bus Stops on Bicycle Traffic Flow byCellular Automata

Zhaoguo Huang 1234 Xiucheng Guo 134 Chunbo Zhang 134 and Hongying Zhang 2

1 Jiangsu Key Laboratory of Urban ITS Southeast University China2School of Civil Engineering Lanzhou University of Technology China3Jiangsu Province Collaborative Innovation Centre of Modern Urban Traffic Technologies Southeast University China4School of Transportation Southeast University China

Correspondence should be addressed to Xiucheng Guo seuguo163com

Received 28 February 2018 Revised 5 July 2018 Accepted 6 August 2018 Published 2 September 2018

Academic Editor Antonio Comi

Copyright copy 2018 ZhaoguoHuang et alThis is an open access article distributed under theCreative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Since currently huge traffic demands for public bus and bicycles exist in themajority of cities of China it is highly likely that the busstop has undeniable impacts on the bicycle flow that is close to the bus stop In this paper we proposed the test in several bicyclepaths beside bus stops in Nanjing and aimed at exploring the specific effects of bus stop on the bicycle flow nearby We assumedthere were two such effects space restriction and pedestrian conflicts Further we built up a cellular automation (CA) model tostudy the feature of these effects and find out not only the outcome of these two effects but also the inner correlation between them

1 Introduction

The bicycle travel for utilitarian purpose plays a significantrole in the large cities of China Abundant city roads areequipped with facility for thousands of bicycle user for thepurposes of commuting or relaxing However the design ofbicycle facility on the road [1] which is not the focus indesigning a road has several problems that cause bicycletraffic flow hardly to run smoothly such as the red-lightrunning [2 3] and bicycle conflicts with other road users [4]One of the problems on the bicycle path is the effect of busstop as showed in Figure 1 on bicycle traffic flow

In this study we assumed the bus stop has twomain kindsof effects on bicycle traffic flow (1) pedestrians who crossthe bicycle path in order to travel between the bus stopsand the main pedestrian street bus users have to cross thebicycle path and thus cause delay and potential dangers onthe bicycle (2) some bus stops specifically harbor-shapedbus stops demanding land space from bicycle path reducethe width of bicycle path and influence the bicycle flowperformance

In past several decades traffic designers planners andscholars have paid much attention on the effects of bus

stop on traffic flow One prominent achievement is a set ofguidelines for applying in designing and locating bus stopThe guidelines concluded the factors needed to considerwhen designing the l bus stop and also provided informationabout trade-offs among various bus stop treatments [5]Koshy and Arasan studied the impacts of bus stops onthe speed of vehicles by applying the microscopic trafficsimulation model [6] Moura et al analysed optimal bus stoplocation using a sequentially applied two-stage model Theresults obtained from the proposed model showed importantdifferences in the commercial speed of the buses dependingon the final location of the stop [7] Zhao et al analysed thecapacity drop caused by the combination effects of signalisedintersections and bus stops [8 9] Fitzpatrick and Nowlininvestigated the effects of bus stop design on suburban arterialoperations [10] Nevertheless only a few researches relatedto the impacts of bus stop on bicycle traffic flow Zhang etal analysed the influences of various types of bus stops ontraffic operation of bicycles vehicles and buses Four types ofstops were considered according to the geometric feature andlane arrangement The results showed that type 3 and type 4stop had the least impact on bicycle speed [11] Ranasinghe etal investigated the saturation headway variation at signalised

HindawiJournal of Advanced TransportationVolume 2018 Article ID 5876104 8 pageshttpsdoiorg10115520185876104

2 Journal of Advanced Transportation

Figure 1 Effects of bus stop on bicycle traffic flow

intersection approaches with a downstream bus stop andbicycle lane The results showed that buses significantly affectchange mean saturation headway where bicycles do not [12]Borjessonca et al optimised the number of bus stops andprices for car bus and cycling in the busiest inner citycorridor in StockholmThe authors found that in the welfareoptimum the bus service only requires a small subsidy due tocongestion in the bus lane crowding in the buses and extraboarding and alighting time per passenger [13]

Evaluation methods of Bicycle Level of Service (BLOS)have continually seen improvements and refinements BLOSis set up based on the perceived hindrance to user by Botmaas a function of pedestrian and bicyclist volumes path widthand speed [14] According to 2010 Highway Capacity Manual(HCM) the disturbing events of bicycles have been dividedinto passing events and meeting event and occur whenthe bicycle encounters a bicycle in the same direction andopposing direction respectively HCM also define that theweighted sum of the number passing and meeting events iscriterion used to determine level of service for bicycle paths[15] Additionally some research proposed several specialcases of BLOS evaluation model as the extended applicationof the generalized BLOS evaluation model Li et al provideda special case of BLOS evaluation model conditioning onphysically separated bicycle roadway [16 17] Minh et alconducted a field video and found that the speed differencesbetween passing and passed motorcycle were big [18]

Another area of intense research in the recent past isthe application of cellular automation model (CA model)in simulation of traffic flow This has been spurred by theincreasing presence and utility of computer used for researchThis tool is widely used to determine LOS measures forvehicle land and is expected to analyse the level of servicefor bike path construction CA model has the advantagethat it could operate computationally simply while providingrich behavior data for comparison to the empirical resultsand collection of the features of the bike flow The basicprocess for CA model is the simulation rulesrsquo establishmentand parametric calibration Lin has summarized the previousstudies and set up a set of detailed simulation definitions andupdate rules for CA model and validated the model withsimulation experiments [19] Gregory and Alex establishedCA model to analyse the service level of bicycle roadwayand calibrated the model with data [20] Zhao used the CAmethod to model the characteristics of bicycle passing eventsin mixed bicycle traffic on separated bicycle paths The CAmodel was calibrated with the use of the field data to simulate

the passing events in the mixed bicycle traffic The resultsshowed that the CA model could simulate the features ofbicycle passing events well [21] Luo et al proposed a CAmodel to simulate the car and bicycle heterogeneous trafficon urban road To capture the complex interactions betweenthese two types of vehicles a novel occupancy rule is adoptedin the proposed model Results indicated that the constantand fixed occupancy rule adopted in the previous studiesmight lead to overestimation of car flux in the heterogeneoustraffic flows with different bicycle densities [22] Xue et almodeled the bicycle flow using an improved Burgers CAmodel and the results showed that the improved model couldeffectively simulate the bicycle flow [23]

Literature review results showed the bus stops have asignificant negative impact on traffic operation and this isparticularly true for bicycle flows When a bus stop installedon a bike lane the bicycle operation may be hindered becauseof the heavy pedestrian activities at the bus stop resulting inthe disturbance of bicycle flow As such there is a need toinclude the effects of bus stops when modeling the bicycletraffic flow With the development of simulation techniquesCA model provided us with an efficient and flexible wayto capture the bicycle flow characteristics at the bus stopbike lane Within the CA model each bicycle and pedestriancan be set as one cell The parameters for each cell suchas the acceleration of bicycle can be also set in a flexibleway Consequently the interactions between pedestrians andbicycles can be simulated accurately

The objective of this study is to investigate the effects ofbus stop on the bicycle traffic flow A cellular automation(CA) model is developed for this goal Accordingly theremainder of this paper is organized as follows The nextsection specifies the data collection The CA model formodeling bicycle traffic flow is given in Section 3 Section 4introduces the CA models results and discussion Finallyconclusions are presented in Section 5

2 Data Collection and Description

Data of bicycle traffic flow on exclusive bicycle path iscollected for the later simulation work Field studies werecarried out on two sections of exclusive bicycle paths inthe center of Nanjing city West Beijing Road and LongpanRoad We selected the sections near Gulou Bus Stop onWest Beijing Road which is a conventional bus stop andthe section around a harbor-shaped bus stop Gangzicun BusStop on Longpan Road to compare and analyse the effectof alternative bus stop on bicycle traffic flow Both selectedsections are generally 35 m wide smooth bicycle path whilethe section near the harbor-shaped bus stop on LongpanRoad is 25 m wide

Cyclists are typically students recreational users andcommuters The selected bicycle paths have exclusive useHowever on certain sections of paths beside the bus stopspedestrians were allowed through to the bus stop To collectfield data traffic scenes on the bicycle path section wererecorded with video cameras by placing cameras on thepedestrian streets along the selected bicycle paths as shownin Figure 2 One hour of video was recorded on Gulou Bus

Journal of Advanced Transportation 3

Figure 2 Field data collection

Stop and one hour of video was recorded on Gangzicun BusStopThe recording hours are enough for data collection dueto the heavy bicycle traffic volume

The recorded video tapes were reviewed in the laboratoryfor data reduction The media software VideoStudio wasused to estimate the bicyclepedestrians flows speeds andconflicts From the video we found that there exist two differ-ent treatments when cyclists meet conflicts with pedestrianSome cyclists would reduce the bicycle speed and wait pedes-trians crossing the path in prior while others may maintainthe speed through the conflict area before the pedestrians

3 CA Model

The model implementation tool in this paper is MATLABR2015a In this paper we simulated mixed bicycle flowand pedestrian flow with a multilane CA model whichincludes three parts CA model definition updating rulesand simulation rules

31 CA Model Definition

311 Cell Size The CA model simulates bicycle roadway inunit of the cell which represents a discrete specific sectionof the roadway Cell size largely depends on the trafficdensity on the roadway and bicyclersquos physical size During thesimulation each cell has only two states occupied or not Wecan refer to other researchrsquos definition on cell size to decidethis index Gregory Gould set 7 feet as the length of a bicyclecell while both Zhang and Liu adopted 2meters as the lengthof a bicycle cell

Considering the impact of acceleration limits is 18ms2(this will be discussed later) and bicyclesrsquo physical size we set2mas the length of a bicycle cell and choose 05m as the widthof a bicycle cell

312 Dimension We have two dimensions set in our sim-ulation The x dimension represents the direction in whichbicycle flow travels and the y dimension shows the lateraldistance which has boundaries as shown in Figure 3 Theboundaries should be corresponded with the number ofbicycle lanes

In time t the current position of bicycle n is xn(t)and y (t) and d1n(t) d

2

n(t) and d4n(t) respectively representthe forward space headways between current cell of the

2m

05m

y

x

d1

n(t)

d2

n(t)

d3

n(t)

a1

n(t)

a2

n(t)

Figure 3 Illustration of the schematic diagram of cells on bicycleroadways

x

y

Figure 4 Illustration of modeling scene for conventional bus stop

bicycle n and the nonoccupied cell in the left front straightfront and right front side

313 Modeling Bicycle Size We set two modeling scenesone for a bicycle path near the harbor-shaped bus stop andanother for a bicycle path near the conventional bus stop

The model for the bicycle path with the conventional busstop is shown in Figure 4

The modeled bicycle path is a matrix consisting of 7 times 80cells where the width of path valued 35m is as wide as sevencells and the length of path to test is set as long as eighty cells(160m) and the bus stop is located beside 51 to 55 cells in xdimension The yellow area in Figures 4 and 5 represents thearea to generate pedestrian flow to cross the path and causeconflicts on cyclists

The model for the bicycle path with the harbor-shapedbus stop changes the location of bus stop compared with thepath with conventional bus stop whose bus stop occupies afew area from bicycle path and thus reduces the width of pathto five cells (25m)Themodel is illustrated in Figures 4 and 5

314 Bicyclersquos Speed Bicyclersquos speed includes lateral speedand forward speed Lateral speed is perpendicular to thestreet that occurs when bicycle changes lanes and forwardspeed is parallel to the street In this paper v stands for therepresented forward speed and vrsquo for the lateral speed Wesuppose that only the current lane and adjacent lanes areavailable for bicycle to run which means the bicycle can onlychange at most one lane at a time unit

In the CA model we assume the simulated speeds ofbicycles are discrete values the forward speed could be 1 23 4 and 5 where five (10ms) is set as the fastest speed limitthat bicycle could achieve And bicyclersquos lateral speed couldbe -1 0 and 1 The negative symbol represents the direction

32 CA Model Updating Rules In each step of simulationbicyclersquos speed and position are updated according to the


x

y

Figure 5 Illustration ofmodeling scene for harbor-shapedbus stop

basic movement schematic and the definition of speed aboveThe speed updating should be prior to position updating

321 Updating Forward Speed Normally bicycles tend toaccelerate to its peak speedHowever the acceleration processis constrained from three restrictions

(1) Bicycle Acceleration Ability According to AmericanAssociation of State Highway and Transportation Officials(AASHTO 2001) [24] the maximum acceleration rate foran adult upright cyclist is 15 ms2 Though very few officialdocuments have been made to define the acceleration ofelectric bicycle it makes intuitive sense that the accelerationspeed of electric bicycle is faster than the acceleration ofconventional bicycle As a result we set overall bicyclersquosforward acceleration ability as 2ms2 (1 cell) We assumethe acceleration only takes place when the bicycle does notchange lane

(2) The Maximal Space That Allows Bicycle to Approach Thevalue of the maximal space dnmax is selected based on thefollowing schematic If the bicycle is in the leftmost lane (orrightmost lane) we set the value dn1 (or dn3) as 0 to preventthe bicycle out of the bicycle roadway If dn2 = maxdn1dn2 dn3 then dnmax = dn2 If dn2 = maxdn1 dn2 dn3 anddn1 = dn3 then there is 50 chance that dnmax = dn1 and 50chance that dnmax = dn3 Otherwise dnmax = maxdn1dn3

In addition we set 6 as the safe distance that the bicycledoes not need to change lane even when the forward spacedn2 is the shortest distance

(3) The Maximal Speed Limit Which Is Defined as 5 inSection 314 The updating expression of a bicyclersquos forwardspeed is as follows vn(t + 1) = vn(t) + 1 if dnmax = dn2vn(t +1) = minvn(t + 1) dnmax vmax

322 Forward Speed Random Slowdowns Although mostbicycles try to run at its maximal speed for the utilizationpurpose there are some uncertain factors disturbing bicycleflow that still may cause bicycle to slightly slow down Wedefine that the probability of a bicycle to randomly slow downis Ps and when this happens we have vn(t+1) = maxvn(t+1) minus 1 1

Besides those uncertain impacts the conflict with cross-street pedestrian also may cause the deceleration of bicycleHowever from the field study not every bicycle will slowdown when they meet conflict with pedestrian on the bicyclepath some of them simply maintain the forward speed and

y

x

Figure 6 The influencing area of a pedestrian on bicycle flow

make pedestrian wait We define the probability for a bicycleto slow downwhen it is influenced by a pedestrian as Psm andthe influence area showed in Figure 6

The values of 119875119904and 119875

119904119898are calibrated later

323 Updating Lateral Speed From the former discussionthe lateral speed can only be -1 0 and 1 respectivelyrepresenting the movement of changing to the left lane con-tinuing on the current land and changing to the right laneThe movement of changing lane is dependent on the spaceheadway the bicycle will turn into the lane with the mostforward space The mechanism of this update is that bicyclestend to run on the lane with the longest space headway

If 119889nmax = dn2 v1015840 = 0 if 119889nmax = dn1 v

1015840 = 1 if dnmax =dn3 v1015840 = minus1

324 Updating Position Theway to update position ismerelyby adding updated speed to the current position of thebicycle

xn (t + 1) = xn (t) + vn (t + 1) (1)

yn (t + 1) = yn (t) + v1015840

n (t + 1) (2)

33 Simulation Rules

331 Bicycle Initialization The bicycle flow initializationconsists of specific bicycle initialization with feature ofdensity and speed distribution The traffic volume and speeddistribution could be gained from field survey Then we havethe density of bicycle flow based on the fundamental trafficflow theory k = QV

Figures 4 and 5 show the way to initial the bicycleflow The green area is the potential cell to generate bicyclewith probability Pocc which represents the probability that arandom cell in bicycle roadway is occupied This value couldbe calculated as

119875119900119888119888=

119896

119898119886119909119894119898119886119897 119889119890119899119904119894119905119910=

119896

(100015) lowast 6(3)

According to the field study the initial average speed is 6525ms for the path with harbor-shaped bus stop and the valuefor the path with conventional bus stop is 6481 ms Weassume the distribution of initial bicycle speed is normaldistribution s sim N(v 120590) and the value of 120590 is set as 06 withthe field data So the initial speed for a random bicycle in themodel is int(s2) cells

332 Pedestrian Initialization The yellow area and red areain Figures 4 and 5 show the potential cell to generate


x

y

Figure 7 Illustration of initialization of bicycle and pedestriantravel

pedestrians who are going to cross the bicycle roadway Theprobability for a potential cell to generate a pedestrian in atime unit is based on the pedestrian rate

Specific studies have found pedestrian walking speedsranging from 1253 ms to 1319 ms Considering this speedwould be reduced when pedestrians cross the bicycle paththerefore we set the pedestrian speed as a constant value of 1ms (2 cells)

333 Bicycles-Pedestrians Conflicts For a normal pedestrianin a randomly selected marked cell if there is at least onebicycle regardless of the movement of changing lane thatcould approach the cell in front of the pedestrian in a singletime unit or could approach the cell two steps before thepedestrian in a single time unit (orange area) the speed ofpedestrian becomes 1 Otherwise the speed of pedestrianbecomes 2

34 Traffic Features Detectors

341 Conflict Counter Whenever there is a bicycle runninginto the conflicting area (orange area in Figure 7) of anypedestrian we view it as an event of conflict Finally wecan obtain the conflicts number in each second and also theaverage conflicts number per second

342 Influenced Speed Detector This detector is used torecord the bicycle traffic flow performance in the area besidethe bus stop We record the average speed of bicycles in thearea of 50 to 65 in x dimension in each second and also theoverall average speed of bicycle in this area across the wholemodeling period

35 Calibration of theCAModel Most input value ofmodel isselected as the same value of the field study such as the initialspeed of bicycle and pedestrian the density of people andcyclists and the geometric size of bicycle path But the valueof the probability of a bicycle to randomly slow down Ps andthe probability of a bicycle to slow down when influenced bypedestrians Psm remain unknown

In this study the least sum squarewas used to calibrate thevalue of Ps and PsmWe setmultiple scenarios for the values ofPs and Psm varying from 01 to 1 respectively with intervalsof 01 for both An iterative procedure was then followed tofind out the optimum Ps and Psm which provide the leastsum square of differences in simulated and observed averagespeed

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2

Figure 8 Average conflict numbers per second for path withharbor-shaped bus stop

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2

Figure 9 Average conflict numbers per second for path with con-ventional bus stop

From Table 1 when values of both Ps and Psm are 08 theminimal least sum square is obtained This result is validatedby corresponding to the field data where Psm = 075

4 Result Analyses and Discussion

We selected two indexes to indicate the performance ofbicycle traffic flow the average conflicting number persecond and the average speed of influenced bicycle Trafficflow performances under different conditions could revealthe effect of alternative bus stop condition on bicycle trafficflow

41 Conflicts Number The average conflict numbers per sec-ond for different scenes are showed in Figures 8 and 9 whereblue areas indicate low conflict number while yellow indicateshigh conflict number with corresponding pedestrian densityand bicycle density (Lower conflict number is preferred)Both pedestrian density and bicycle density vary from 0025vehm2 to 02 vehm2 with interval of 0025 vehm2 Theresults show low conflicts frequency with low bicycle andpedestrian density and the conflicts number increases withthe incensement of bicycle and pedestrian density at bothpaths This finding is reasonable because the potential ofconflicts is low at free flow condition with low traffic densitywhile the possibility of conflicts increased at high traffic


Table 1 The least sum square (LSS) of speed differences with different probability of a bicycle to randomly slow down

119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density

Figure 10 Difference of average conflict numbers per secondbetween the scene with harbor-shaped bus stop and conventionalbus stop

density condition due to the increasing interactions betweenbicycles and pedestrians [3 7] It was also found that conflictsfrequency at harbor-shaped bus stop area is higher than thatat conventional bus stop area for the same bicycle and pedes-trian density This result indicates that the harbor-shaped busstop has more impact of bicycle traffic flow in reflection ofincreased conflicts which can be explained as harbor-shapedbus stop could reduce the width of bicycle path

In Figure 10 red area indicates that the average conflictnumber per second in the scene with the harbor-shaped busstop is larger than that in the scene with the conventional busstop and blue area indicates the counter relationship It canbe concluded that pedestrian density has little influence whenbicycle density is at low level When bicycle density is highpedestrian density causes significant change on the numberof conflicts as harbor-shaped bus stop has more conflictswhen pedestrian density is low but conventional bus stop hasmore conflicts when pedestrian density is high This resultis consistent with previous studies of [1 3] which showedvarious impacts of bicycle and pedestrian density on conflictsfrequency

42 Speed Performance of Influenced Bicycles Another indexof trafficperformance is the speed of those bicycles influenced

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12

Figure 11 Average influenced bicycle speed for path with harbor-shaped bus stop

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15

Figure 12 Average influenced bicycle speed for path with conven-tional bus stop

by pedestrians As showed in Figures 11 and 12 yellow arearepresents higher speed performance than blue area Theresult shows that bicycle speed is high at a low bicycle andpedestrian density level which is intuitive because bicycleat free flow can travel with a high speed It was foundthat bicycle speed decreased with the increasing bicycleand pedestrian density The result is supported by previousresearches [6 7 12] An interesting finding is that bicycle


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02

Figure 13 Difference of average influenced bicycle speeds betweenthe scene with harbor-shaped bus stop and conventional bus stop

speed at conventional bus stop area is higher than thatat harbor-shaped bus stop area for the same bicycle andpedestrian densityThe result confirms the impacts of harbor-shaped bus stop area on bicycle traffic flow which can be alsofound in [8]

As shown in Figure 13 the average speeds of influencedbicycle at the two bus stops are similar under condition ofhigher bicycle density regardless of the difference of pedes-trian However when bicycle density reduces conventionalbus stop results in higher speed of bicycle travel whenpedestrian density is low while harbor-shaped bus stop hashigher bicycle speed when the pedestrian density is higher

5 Conclusions

In this study we assumed two different effects of bus stopalong the bike path on bicycle traffic The first effect is thatsome bus stop may occupy land space from bicycle pathanother effect is the pedestrian number crossing the bicyclepath to bus stops

We set up cellular automation model to simulate theseeffects and calibrate the parametersrsquo values considering theleast sum square difference between simulated and empiricalresults

Finally we analysed the result of CA model The increas-ing density of pedestrian crossing the street would reduce thebicycle speed and cause conflicts number to increase On theother hand conventional bus stop could provide wider rangeof positive performance condition in terms of both speed andconflict number comparing to the harbor-shaped bus stopwhich limits the space for bicycle flow

Nevertheless the combination effect of bus stop types andpedestrian is much complex When bicycle density is lowthe effect of bus stop plays a little role in conflicts numberbut plays an important role in bicycle speed Under thiscondition conventional bus stop is preferred conditioning onlow pedestrian density while harbor-shaped is preferred con-ditioning on high pedestrian density When bicycle densityis high there is limited effect of bus stop on bicycle speedbut not in conflicts number Conventional bus stop is stillpreferred when pedestrian density is low and harbor-shapedbus stop performs better when pedestrian density is high

Data Availability

The data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

The authors declare that they have no conflicts of interest

Acknowledgments

The work is sponsored by the Natural Science Foundation ofGansu Province (No 1506RJZA109) And the research wasconducted by the Bluesky studio at School of TransportationSoutheast University

References

[1] J Dales P Jones R Black et al ldquoInternational cycling infras-tructure best practice studyrdquo in Proceedings of the Internationalcycling infrastructure best practice study 2014

[2] Y Guo P Liu L Bai C Xu and J Chen ldquoRed light runningbehavior of electric bicycles at signalized intersections inChinardquo Transportation Research Record vol 2468 pp 28ndash372014

[3] Y Guo Z Li Y Wu and C Xu ldquoExploring unobserved het-erogeneity in bicyclistsrsquo red-light running behaviors at differentcrossing facilitiesrdquo Accident Analysis amp Prevention vol 115 pp118ndash127 2018

[4] Y Guo T Sayed andM H Zaki ldquoEvaluating the safety impactsof powered two wheelers on a shared roadway in China usingautomated video analysisrdquo Journal of Transportation Safety ampSecurity pp 1ndash16 2018

[5] Fitzpatrick K K Hall D Perdinson and L Nowlin TCRPReport 19 Guidelines for the Location and Design of Bus StopsTRB National Research Council Washington DC 1996

[6] R Z Koshy and V T Arasan ldquoInfluence of bus stops on flowcharacteristics of mixed trafficrdquo Journal of Transportation Engi-neering vol 131 no 8 pp 640ndash643 2005

[7] J L Moura B Alonso A Ibeas and F J Ruisanchez ldquoA two-stage urban bus stop location modelrdquo Networks and SpatialEconomics vol 12 no 3 pp 403ndash420 2012

[8] X-M Zhao Z-Y Gao and B Jia ldquoThe capacity drop caused bythe combined effect of the intersection and the bus stop in a CAmodelrdquoPhysica A Statistical Mechanics and its Applications vol385 no 2 pp 645ndash658 2007

[9] X-M Zhao Z-Y Gao and K-P Li ldquoThe capacity of twoneighbour intersections considering the influence of the busstoprdquo Physica A Statistical Mechanics and its Applications vol387 no 18 pp 4649ndash4656 2008

[10] K Fitzpatrick and R L Nowlin ldquoEffects of bus stop design onsuburban arterial operationsrdquo Transportation Research Recordno 1571 pp 31ndash41 1997

[11] F Zhang Z Li and D Zhao ldquoInfluences of various types ofbus stops on traffic operations of bicycles vehicles and busesrdquoTransportation Research Board Annual Meeting 2015

[12] Dehideniya Udugamage Ranasinge Wathsala Jayanthi BunkerJonathanM ampamp Bhaskar Ashish (2017) Saturation headwayvariation at a signalised intersection approaches with a down-stream bus stop and bicycle lane In Australasian TransportResearch Forum 2017 27-29 November 2017 Auckland NewZealand


[13] M Borjesson C M Fung S Proost and Z Yan ldquoDo buseshinder cyclists or is it the other way around Optimal bus faresbus stops and cycling tollsrdquo Transportation Research Part APolicy and Practice vol 111 pp 326ndash346 2018

[14] H Botma ldquoMethod to determine level of service for bicyclepaths and pedestrian-bicycle pathsrdquo Transportation ResearchRecord no 1502 pp 38ndash44 1995

[15] Botma H Bicycle Planning in theNetherlands Presented at theHelsinki University of Technology April 1994 7

[16] Z Li W X Wang J Shan J Jin C Lu and Yang ldquoAnalysis ofBicycle Passing Events for LOS Evaluation on Physically Sepa-rated Bicycle Roadways in Chinardquo in Presented at 89th AnnualMeeting of the Transportation Research Board WashingtonDC 2010

[17] Z Li W Wang P Liu J Bigham and D R Ragland ldquoModel-ing bicycle passing maneuvers on multilane separated bicyclepathsrdquo Journal of Transportation Engineering vol 139 no 1 pp57ndash64 2013

[18] CMinh K Sano and SMatsumoto ldquoCharacteristics of passingand paired riding maneuvers of motorcyclerdquo Journal of theEastern Asia Society for Transportation Studies vol 6 pp 186ndash197 2005

[19] H Liu H Wang and Y H Feng ldquoBicycle flowmodeling basedon cellular automatardquo in Proceedings of the 2008Chinese ControlConference (CCC) pp 527ndash531 Kunming China July 2008

[20] G Gould and A Karner ldquoModeling bicycle facility operationcellular automaton approachrdquo Transportation Research Recordno 2140 pp 157ndash164 2009

[21] D Zhao W Wang C Y Li Z B Li P M Fu and X JHu ldquoModeling of passing events in mixed bicycle traffic withcellular automatardquoTransportation Research Record no 2387 pp26ndash34 2013

[22] Y J Luo B Jia J Liu W H K Lam X G Li and Z Y GaoldquoModeling the interactions between car and bicycle in hetero-geneous trafficrdquo Journal of Advanced Transportation vol 49 no1 pp 29ndash47 2015

[23] S Xue B Jia R Jiang X Li and J Shan ldquoAn improved Burgerscellular automatonmodel for bicycle flowrdquo Physica A StatisticalMechanics and its Applications vol 487 pp 164ndash177 2017

[24] AASHTO (2001) A policy on geometric design of highwaysand streets American Association of State Highway and Trans-portation Officials Washington DC

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom


Figure 1 Effects of bus stop on bicycle traffic flow

intersection approaches with a downstream bus stop andbicycle lane The results showed that buses significantly affectchange mean saturation headway where bicycles do not [12]Borjessonca et al optimised the number of bus stops andprices for car bus and cycling in the busiest inner citycorridor in StockholmThe authors found that in the welfareoptimum the bus service only requires a small subsidy due tocongestion in the bus lane crowding in the buses and extraboarding and alighting time per passenger [13]

Evaluation methods of Bicycle Level of Service (BLOS)have continually seen improvements and refinements BLOSis set up based on the perceived hindrance to user by Botmaas a function of pedestrian and bicyclist volumes path widthand speed [14] According to 2010 Highway Capacity Manual(HCM) the disturbing events of bicycles have been dividedinto passing events and meeting event and occur whenthe bicycle encounters a bicycle in the same direction andopposing direction respectively HCM also define that theweighted sum of the number passing and meeting events iscriterion used to determine level of service for bicycle paths[15] Additionally some research proposed several specialcases of BLOS evaluation model as the extended applicationof the generalized BLOS evaluation model Li et al provideda special case of BLOS evaluation model conditioning onphysically separated bicycle roadway [16 17] Minh et alconducted a field video and found that the speed differencesbetween passing and passed motorcycle were big [18]

Another area of intense research in the recent past isthe application of cellular automation model (CA model)in simulation of traffic flow This has been spurred by theincreasing presence and utility of computer used for researchThis tool is widely used to determine LOS measures forvehicle land and is expected to analyse the level of servicefor bike path construction CA model has the advantagethat it could operate computationally simply while providingrich behavior data for comparison to the empirical resultsand collection of the features of the bike flow The basicprocess for CA model is the simulation rulesrsquo establishmentand parametric calibration Lin has summarized the previousstudies and set up a set of detailed simulation definitions andupdate rules for CA model and validated the model withsimulation experiments [19] Gregory and Alex establishedCA model to analyse the service level of bicycle roadwayand calibrated the model with data [20] Zhao used the CAmethod to model the characteristics of bicycle passing eventsin mixed bicycle traffic on separated bicycle paths The CAmodel was calibrated with the use of the field data to simulate

the passing events in the mixed bicycle traffic The resultsshowed that the CA model could simulate the features ofbicycle passing events well [21] Luo et al proposed a CAmodel to simulate the car and bicycle heterogeneous trafficon urban road To capture the complex interactions betweenthese two types of vehicles a novel occupancy rule is adoptedin the proposed model Results indicated that the constantand fixed occupancy rule adopted in the previous studiesmight lead to overestimation of car flux in the heterogeneoustraffic flows with different bicycle densities [22] Xue et almodeled the bicycle flow using an improved Burgers CAmodel and the results showed that the improved model couldeffectively simulate the bicycle flow [23]

Literature review results showed the bus stops have asignificant negative impact on traffic operation and this isparticularly true for bicycle flows When a bus stop installedon a bike lane the bicycle operation may be hindered becauseof the heavy pedestrian activities at the bus stop resulting inthe disturbance of bicycle flow As such there is a need toinclude the effects of bus stops when modeling the bicycletraffic flow With the development of simulation techniquesCA model provided us with an efficient and flexible wayto capture the bicycle flow characteristics at the bus stopbike lane Within the CA model each bicycle and pedestriancan be set as one cell The parameters for each cell suchas the acceleration of bicycle can be also set in a flexibleway Consequently the interactions between pedestrians andbicycles can be simulated accurately

The objective of this study is to investigate the effects ofbus stop on the bicycle traffic flow A cellular automation(CA) model is developed for this goal Accordingly theremainder of this paper is organized as follows The nextsection specifies the data collection The CA model formodeling bicycle traffic flow is given in Section 3 Section 4introduces the CA models results and discussion Finallyconclusions are presented in Section 5

2 Data Collection and Description

Data of bicycle traffic flow on exclusive bicycle path iscollected for the later simulation work Field studies werecarried out on two sections of exclusive bicycle paths inthe center of Nanjing city West Beijing Road and LongpanRoad We selected the sections near Gulou Bus Stop onWest Beijing Road which is a conventional bus stop andthe section around a harbor-shaped bus stop Gangzicun BusStop on Longpan Road to compare and analyse the effectof alternative bus stop on bicycle traffic flow Both selectedsections are generally 35 m wide smooth bicycle path whilethe section near the harbor-shaped bus stop on LongpanRoad is 25 m wide

Cyclists are typically students recreational users andcommuters The selected bicycle paths have exclusive useHowever on certain sections of paths beside the bus stopspedestrians were allowed through to the bus stop To collectfield data traffic scenes on the bicycle path section wererecorded with video cameras by placing cameras on thepedestrian streets along the selected bicycle paths as shownin Figure 2 One hour of video was recorded on Gulou Bus





3 CA Model







2


2m

05m

y

x

d1

n(t)

d2

n(t)

d3

n(t)

a1

n(t)

a2

n(t)


x

y











x

y










y

x



The values of 119875119904and 119875






xn (t + 1) = xn (t) + vn (t + 1) (1)

yn (t + 1) = yn (t) + v1015840

n (t + 1) (2)

33 Simulation Rules



119875119900119888119888=

119896

119898119886119909119894119898119886119897 119889119890119899119904119894119905119910=

119896

(100015) lowast 6(3)




x

y










005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2








119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density





005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15




005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






3 CA Model







2


2m

05m

y

x

d1

n(t)

d2

n(t)

d3

n(t)

a1

n(t)

a2

n(t)


x

y











x

y










y

x



The values of 119875119904and 119875






xn (t + 1) = xn (t) + vn (t + 1) (1)

yn (t + 1) = yn (t) + v1015840

n (t + 1) (2)

33 Simulation Rules



119875119900119888119888=

119896

119898119886119909119894119898119886119897 119889119890119899119904119894119905119910=

119896

(100015) lowast 6(3)




x

y










005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2








119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density





005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15




005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



x

y










y

x



The values of 119875119904and 119875






xn (t + 1) = xn (t) + vn (t + 1) (1)

yn (t + 1) = yn (t) + v1015840

n (t + 1) (2)

33 Simulation Rules



119875119900119888119888=

119896

119898119886119909119894119898119886119897 119889119890119899119904119894119905119910=

119896

(100015) lowast 6(3)




x

y










005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2








119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density





005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15




005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



x

y










005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

16

14

12

10

8

6

4

2








119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density





005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15




005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




119875119904

119875119904119898

01 02 03 04 05 06 07 08 09 101 11020 10776 9985 10690 9124 8456 7893 7717 7603 708502 10027 9953 9899 7930 8531 6600 6531 6659 5045 419203 9157 8683 7987 8217 5958 6017 4855 5478 2868 342504 8270 7318 6485 5824 4722 3875 4155 3036 4089 241805 7123 6142 5962 4985 4770 2519 2227 1727 1855 057906 6015 5616 4844 3718 2655 2112 3129 1857 0831 016507 5146 4813 2370 1890 1577 1018 0468 0291 0196 009208 3281 2381 1308 0440 0681 0504 0107 0016 0048 029009 0577 0730 0080 0030 0129 0442 1025 1265 1388 24261 2167 2889 2108 2569 3108 3649 3753 4582 4581 4762

005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

3

2

1

0

-1-2-3-4-5-6

Bicy

cle D

ensit

y

Pedestrian Density





005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

32

3

28

26

24

22

2

18

16

14

12


005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bike

Den

sity

Pedestrian Density

35

3

25

2

15




005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



005

01

015

02

025

03

035

04

005 01 015 02 025 03 035 04

Bicy

cle D

ensit

y

Pedestrian Density

12

1

08

06

04

02

0

-02




5 Conclusions





Data Availability




Acknowledgments


References




























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia

















RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


modeling the effects of bus stops on bicycle traffic flow by cellular...

Documents