Evaluating the Effects of One-Way Traffic Management on Different Vehicle Exhaust Emissions Using an Integrated Approach

,


Introduction
Environmental and energy issues derived from transportation emissions become increasingly severe and have attracted increasing attention from the transportation managers and practitioners in Chinese metropolis.e excessive emissions due to transportation have caused negative impacts on the air quality, public health, and climate [1].It is reported the tra c pollutants including Carbonic Oxide (CO), Hydrocarbon Compounds (HC) and Nitrogen Oxides (NO x ), from 2016 to 2017, the CO, HC and NO x emissions stemmed from transportation were 34.293, 42.20 and 57.78 million tons in China, respectively, which are not ignorable, in tra c emissions, the vehicles contribute 87.7% for CO emission, 84.1% for HC emission, and 92.5% for NO x emission [2].Compared with o -peak hours, fuel consumption during peak hours increased by 10%, and CO, HC and NO x emissions from cars increased by 20% [3].
It is widely acknowledged that reducing tra c delays, the number of acceleration and deceleration events with stopand-go tra c are bene cial for reducing vehicle emissions.Some studies have been conducted to address the in uences of tra c management measurements on tra c emissions and to propose tra c operation methods that can reduce tra c emissions.Stemley [4] pointed out that in most urban tra c environments, one-way tra c could reduce con ict points and was bene cial for reducing the tra c delays and the occurrence of accidents.However, in some cases of one-way tra c networks, drivers must bypass the blocks to reach their destinations and might increase the travel distances.Nagurney [5] studied the relationship between travelers' route choice and emissions under various tra c demands and road network structures.ey indicated that the rationality of the road network structure played an important role in reducing the traffic emission.Results showed that a reasonable road network configuration might increase travel distances but would not increase traffic emissions.On the contrary, the unreasonable road network structure may increase the overall emissions even although the travel demand was set to be smaller.Frey et al. [6] used portable instruments to collect the emission data of 20 vehicles of 4,000 km for investigating the key factors affecting vehicular emissions and to assess the impacts of coordinated traffic changes on traffic emissions.e results demonstrated that traffic coordination was one of the major factors affecting traffic emissions, and traffic design that led to increasing acceleration frequency would directly raise the traffic emissions.Coelho et al. [7] examined the relationship between emissions and vehicular dynamics and found that the traffic delays at intersections would boost the number of unexpected parking and thus result in increases in CO, HC and NO x emissions by approximately 15%, 10% and 40%, respectively.ey also indicated that more stops of vehicles would create more stops for the following cars and thus traffic control settings should consider the impacts of vehicle stops on traffic emissions.Moreover, Coelho et al. [8] quantified the impacts of urban single-lane roundabouts on traffic emissions.ey reported that traffic emissions were correlated with speed and vehicle queue length when the roundabouts were congested.
e accelerations and the number of collision points were strongly linked to the total emissions of CO, HC and NO x at the intersections.Ahn et al. [9] employed the GPS data and micro-simulation tools to study the impact of route selections on the traffic emissions.e results showed that the high load operation of the vehicle engine was one of the main reasons for the increase in emissions and faster but longer routes did not always reduce emissions.Optimizing route selections and reducing emissions should be done simultaneously to ensure the multiple optimization objectives.Zegeye et al. [10] put forward a model-based traffic flow control approach for reducing both travel time and emissions in a traffic network.e control strategy was examined by simulation experiments and the results indicated that both reduction in emission and travel time could be achieved by properly defining the optimization of control strategy.However, they reported that the control strategy focusing on the reduction of travel time alone might not reduce the emissions simultaneously.Coelho et al. [11] used field measurements and traffic flow simulations to study the effects of stop-and-go behavior of vehicles in crowded traffic on traffic emission.e results showed that stopping led to a surge in CO and Carbon Dioxide (CO 2 ) emissions.Traffic disruption contributed to the largest proportion of traffic emission and account for more than 55% of CO emissions and more than 20% of total CO 2 emissions.Pandian et al. [12] established a traffic flow model including vehicle dynamics, road configurations and traffic flows to study the impacts of traffic characteristics on emissions.ey pointed out that vehicle exhaust emissions near the traffic intersections largely depended on fleet speed, deceleration rates, queuing time in idle modes, red signal time, acceleration rates and queue length.ese characteristics have cumulative effects on traffic pollutant emissions, but the most likely factors affecting the emission at intersections were not declared.Madireddy et al. [13] combined the emission model VERSIT with the microscopic traffic simulation tool Paramics to study the impact of vehicle speed control on traffic emissions and the effects of traffic signal coordination on traffic flow in the residential areas of Antwerp, Belgium.e results demonstrated that when the speed limit reduced from 50 km/h to 30 km/h, CO 2 , and NO x emissions reduced by 25%.Traffic signal coordination could promote traffic flow and reduced CO 2 and NO x emissions by 10%.Nasir et al. [14] conducted vehicle exhaust pollutant emissions tests concerning HC, CO, CO 2 , particulate matter (PM) and NO x under traffic conditions of free-flow conditions, moderate congestion and severe congestion.ey indicated that the shortest path was not the path with the least emissions.Traffic emission had a strong correlation with average speed, traffic congestion, stops and the fastest route.Chen et al. [15] used on-site detectors to record the vehicle's operating speed and estimated HC, CO and CO 2 emissions using a micro-emission model based on the vehicles' speed and trajectory.ey explored the impacts of traffic conditions on vehicle activities and emissions and found that the emission trends of individual vehicles were basically consistent with the emission trends of traffic flows when the traffic flow was stable.Jamshidnejad et al. [16] used microscopic simulation so ware SUMO as a platform to investigate the relations between road congestions and traffic emissions.ey proposed a common framework for integrating traffic flows and emissions models to generate mesoscopic integrated flow-emission models.
eir empirical results showed that the mean and standard deviations for CO, HC and NO x relative errors using the proposed model were less than 2% and 1.6%, respectively.Meneguzzer et al. [17] used experimental vehicles equipped with a portable emission measurement system to study the CO, CO 2 and NO x emissions at the roundabouts and signal control intersections.e results show that the NO x emissions at the roundabout were higher than the signal-controlled intersections, while CO 2 and CO emissions presented the opposite principles.
One-way traffic management is one of the useful strategies for alleviating the traffic pressure and reducing saturation in the single-direction of main roads [18].Some studies have assessed the impacts of one-way traffic management on traffic network performances and indicate that it is beneficial for traffic network service capabilities and reducing conflict points [19].Nevertheless, to our best knowledge, scarce studies have analyzed the potential effects of one-way traffic management on different vehicular exhaust emissions [20].One-way traffic management somehow decreases the delays in saturated roads which is helpful for reducing emissions, but also forces vehicles to bypass and increases driving distances of some vehicles that may lead to more emissions.e quantitative evaluations regarding the overall effects of one-way traffic management on traffic network emissions are lacking.Moreover, microscopic emission models and traffic simulation tools are generally applied for traffic emission assessment [14], since it is o en not feasible to evaluate the environmental effects of traffic management measures based on the trial-and-error field experiments.However, the majority of studies using the microscopic traffic simulation with instantaneous emission models in the assessment, did not calibrate and examine the emission models based on the real world emission data that consider vehicle dynamics and di erent vehicle standards in di erent regions [21]. is study stands in the wake of the literature to contribute to the state-of-art studies by investigating the in uences of the typical one-way tra c management on di erent vehicle exhaust emissions (CO, HC, NO x ) in the urban tra c networks using an integrated method.e combination of microscope tra c simulation platform and an instantaneous emission model (Vehicle Speci c Power) calibrated with eld emission data in Chinese city contexts, is developed and employed for the assessment.e di erent exhaust emissions in the intersections, road segments and network levels before and a er the implementations of oneway tra c management are compared to comprehensively understand the e ects of one-way tra c management on different vehicle exhaust emissions.e rest of this paper is structured as follows: Section 2 describes the eld emission data collection, instantaneous emission model calibration process and the development of microscope tra c simulation. e analysis results are presented in Section 3. Lastly, Section 4 provides discussions and concluding remarks of the ndings.

Study Area and Emission Data Collection. It is crucial for tra c emission researches to develop reliable and trustful emission models. is study uses on-site emission measurement equipment to collect eld emission data for constructing a speci c instantaneous emission model in
Chinese city contexts.e study area is "Fengxian" district located in the south of Shanghai, China.Figure 1 shows the roads in the study area.e speed limit in the area is 60 km/h.In the north, this area is bounded by the Nanting Highway, which is an urban main road with three lanes in each direction.In the south, the area is bounded by Jiangnan Road that is dual three-lane.e eastern part of the area is bounded by the Jianghai Road with two-way in each direction and parking spaces on some sections.e western boundary of the area is Cai Chang Road, with two lanes in each direction and parking spaces on the roadsides.e rest of the roads located inside the area are two-way lanes.e surrounding areas are all residential and commercial areas.e average speed of vehicles during the morning peak and evening peak hours in the studied area is only 20 km/h-30 km/h.e experimental vehicles choose small vehicles, mainly including Volkswagen Long Yat, Harvard SUV and other household models according to our research goals.eir gasoline fuel emission standards are China National IV with an engine displacement of 1.6 l-2.0 l. e OBEAS-3000 Portable Emission Tester (PET) system is used to continuously monitor the instantaneous emission of HC, NO x , CO, velocity and acceleration rates of the vehicles during experiments.e PET system is composed of a Siemens E-BOX PC gas analyzer, a vehicle parameter On-Board Diagnostic (OBD) instrument, Global Positioning System (GPS), power control units and a notebook computer for overall control and data recording.e dynamic regimes including the accurate position, speed and acceleration of the experimental vehicle are obtained through the GPS system.e simultaneous vehicle exhaust emissions (CO, HC, NO x ) are detected by the OBEAS-3000 system.e OBEAS-3000 system has high sampling accuracy (every 0.1 s), but we accumulated and stored the overall data in every 1 s in actual process for the sake of reducing the variances of data.e nal collected outputs contain diverse data like the amount of instantaneous exhaust emissions as well as corresponding vehicle dynamic characteristics (e.g., locations, speeds and accelerations).e testing setup of the data collection system is shown in Figure 2. Finally, 79032 observations of emission and vehicle dynamic data were obtained.Figure 3 shows the interface of the tail gas emission collection and record information.

Emission Model and Calibration.
e Vehicle Speci c Power (VSP) model is used as the instantaneous emission model in this study.VSP is the instantaneous power of a unit mass vehicle, in units of kW/t [22].Vehicle transient emissions are closely related to values of VSP.VSP considers the wheel rotation resistance, the aerodynamic drag work, and the increased power required by overcoming the internal frictional resistance and the mechanical loss power of the drive train.Wyatt [23] provides detailed information about VSP model establishment.
e VSP value is related to speed and accelerations and can be approximately expressed by the Equation (1).
In the formula, v is velocity of vehicle. is acceleration.grade is the road gradient.g is the acceleration of gravity, 9.81 m/s 2 .
(1) VSP = v × 1.1 + g × grade + 0.132 + 0.000302v 3 .e statistical results of our empirical eld data show that the VSP value in the range of [30,−30] covers 97.33% of the overall VSP values, which is over the 95% con dence levels of the empirical data.erefore, the VSP interval range is set to [30,−30].e frequencies of VSP intervals derived from our empirical emission data are shown in Table 1.
In order to accurately quantify the VSP-based emission rates by fully using the eld emission data, we divide the VSP values by a step size of 2 kW/t to generate the BIN intervals BIN .
Speci c power partitioning simpli es the computational process and highlights the di erences in emission rates across di erent VSP intervals.Di erent studies used di erent numbers of intervals and speci c power intervals.is study deduces the VSP internals based on the collected eld data using the interval division principles proposed by Frey [24].
e internal de nitions and process are as follows:   requirements of one-way tra c management. e overall one-way tra c organization was determined depending on the tra c delays of intersections and the one-way ow ratios.Six roads in the area were changed to be one-way.e nal one-way tra c management strategy is illustrated as Figure 4. e simulation results of Synchro show that the tra c performances of the studied network were obviously improved in terms of reducing the delay in intersections and travel time in the road segments.e corresponding scenario a er oneway tra c management was established in the VISSIM simulation platform as well [28].e VISSIM can output the detailed data including the speeds and accelerations of each vehicle during the simulation period.Based on the speeds and accelerations of vehicles, the instantaneous VSP of each vehicle can be calculated by Equation (1).A erwards, the instantaneous emission rates of the three pollutant emissions (CO, HC, and NO x ) can be obtained based on the calibrated relationships between VSP and instantaneous emission rates in Table 2. e overall emissions in the road sections or in the intersections can be Taking advantages of the data collected by the eld experiments, the calibrate the relationships between the VSP BIN and emission rates of di erent vehicle exhausts.e VSP values of the experimental vehicle at each second can be calculated using Equation (1) based on the recorded velocity and accelerations.Simultaneously, the corresponding emission rates of the vehicle exhausts were detected by OBEAS-3000 system, which can link the instantaneous emission rates with the VSP values.e emission rates of the same VSP interval were averaged to get a representative value of the emission rate in the VSP interval.e nal calibrated results can be summarized in Table 2. e results construct the relationships among vehicle characteristics, VSP values and corresponding emission rates of di erent exhausts.

Scenarios Constructions.
Two scenarios are established in the study.e baseline scenario is the original tra c networks with revealed tra c demand in the study area.Field demand survey was conducted to collect the actual road network geometry data (e.g.road length, slope, number of lanes, lane width, signal timing, properties of intersections and detailed tra c ow data) during the peak hours in the study area for reappearing the tra c conditions in simulation.Based on the collected data, the baseline scenario was established in the VISSIM [25] simulation platform.e comparison scenario is the same tra c network and tra c demand a er implementations of one-way tra c management.It is important to decide how the one-way tra c management should be implemented.e Synchro [26] tra c so ware was used to simulate the road network and determine which road should be changed to one-way on the basis of quantitative tra c analysis.Synchro can output the delay time and service level in the intersections.e simulated results are shown in Table 3.It can be found that several intersections including intersection 1, 11 and 12, had large tra c volume with ignorable tra c delays.According to the eld tra c ow data in the study area, the one-way ow ratios of many road sections are larger than 1.2 in morning and evening peak hours as shown in Table 4. e design of the one-way tra c management needs to consider duality principle [27].As per the principles of urban one-way tra c organization GAT_486-2004, road sections with a road width less than 10 m and a one-way ow ratio over 1.2, meet the (2)  5. e values in Table 5 are the average instantaneous emission rates of all vehicles crossing the intersections.e results show that the instantaneous CO emission rates of vehicles in the sixteen intersections a er one-way tra c management are markedly reduced as compared to those without one-way tra c management. e decreasing ratio ranges from 6.88% to 66.9% and has a mean value of 20.46%.e instantaneous HC emission rates reduce 7.46-67.15% in di erent intersections and decrease by 21.29% on average a er one-way tra c management.For the NO x emission, the results indicate that the one-way tra c management cuts down the instantaneous NO x emission rates in the intersections by 21.06% on average with a range of 7.47-66.89%.e results of statistical paired T-test show that the improvement of one-way tra c management on reducing the three pollutants are signi cant at the con dence level of 99%.e results demonstrate that the implement of one-tra c management is indeed bene cial for decreasing the three vehicle exhausts in the intersections. is may be ascribed to the fact that one-way tra c management can improve the average speed, alleviate the con icts and reduce the delays in the intersections which are helpful in reducing exhaust emissions.

In uences on the Speed and Emissions in the Road
Sections.On one hand, the implementation of one-way tra c may force the vehicles in the opposite directions to give up the shortest driving path and choose some other routes to bypass.erefore, one-way tra c will inevitably increase the bypass distance of some vehicles, which would contribute to the increasing energy consumption and exhaust emissions during the road sections.On another hand, the implementation of obtained by accumulating the instantaneous emissions over time crossing through the intersections and road sections.

In uences on the Emissions in the Intersections.
e emissions of the CO, HC and NO x in the intersections, road sections and the whole studied network before and a er the implementation of one-way tra c management are compared.e results about emissions in the intersections  ).e instantaneous emission rates of vehicles crossing the road sections before and a er one-way tra c management are shown in Table 7. From the perspectives of all road sections, the CO, HC and NO x emission rates decrease 22.19% from 902.14 mg/s to 701.96 mg/s, 23.38% from 40.98 mg/s to 31.40 mg/s, and 26.29% from 277.45 mg/s to 204.50 mg/s, respectively.ere are some changes in the average speeds in the road sections and thus in the average travel times of road sections.

Regional Overall Emissions.
e overall emission rates of the three vehicle exhausts in the study area including the intersections and road sections before and a er the one-way tra c management are calculated and summarized in Table 8.A er one-way tra c management, the overall CO emission rate in the study network decreases 21.34% from 1784.79 mg/s to 1403.93 mg/s, the overall HC emission rate reduce 22.29% from 80.85 mg/s to 62.83 mg/s and NO x emission rate decreases 23.77% from 536.53 mg/s to 409.00 mg/s.

Emission Regression Analysis.
e tra c discharge in the one-way tra c implementation area can re ect the impact of one-way tra c on urban tra c emissions.A er oneway tra c management is implemented in the area, tra c one-way tra c management enables vehicles on one-way lanes to travel in a single direction and reduces the interferences with the opposite vehicles for alleviating the con icts, which is bene cial for increasing travel speed, travel stability of the vehicles and the operating e ciency of the entire road network. is aspect implies that one-way tra c management is helpful for reducing the exhaust emissions in road sections.
e overall e ects of one-way tra c management on emissions in the road segments need comprehensive evaluations.
e input tra c demands in the simulation scenarios before and a er one-way tra c management are set to be the same.Based on the length and the tra c volumes in the road sections, the bypassing distance can be calculated as follows: In the formula, is bypass tra c volume, Δ is bypass distance, is travel bypassing distance a er one-way tra c management, is travel distance before one-way tra c management. e length, the tra c volumes and the number of bypassing vehicles in each road section a er one-way tra c management are shown in Table 6.e results show that the one-way tra c management leads to the detour of reverse tra c demands.e total number of bypassing vehicles is 1562 per hour a er one-way tra c management.From the macroscopic analysis, the overall detour distance of the road network reaches 702.90 kilometers, and the average bypassing distance per vehicle is about 0.45 km. e average travel speeds in most road sections are improved a er one-way tra c management T 5: Emission rate of three pollutants before and a er one-way tra c management in intersections.
Note: e di erence is equal to the emission a er setting one-way tra c management subtracted by the emission before one-way tra c management.decrease, but the emissions increase.rough the analysis of the relationship between vehicle bypass distance, vehicle velocity, traffic flow and traffic emissions in Fengxian District, although one-way traffic causes a certain detour, it further improves the road capacity and increases the traffic volume in the area.It also increases the velocity of the vehicle, decreases traffic congestion and travel time.One-way traffic implementation has obvious effects on reducing emissions from urban transportation.It is of great significance to the implementation of Shanghai's energy-saving and low-emission transportation strategies.
flow and vehicle velocity increase during peak hours, but emissions of CO, HC, and NO x decrease.e relationships among traffic flow, velocity and emission rates is shown in Figure 5. Before the one-way traffic is implemented, the vehicle velocity in the area is less than 20 km/h, the CO maximum emission is more than 1650 mg/s, the HC maximum emission is more than 90 mg/s, the NO x maximum emission is more than 550 mg/s.A er the implementation of one-way traffic, the peak hourly velocity in the area is more than 30 km/h, and the traffic volume reaches a maximum of 13000 pcu/h, but the emissions decrease with the increasing of the vehicle speed.Before the road design flow is reached, the traffic flow increases as the vehicle velocity increases, and the emissions decrease as the vehicle velocity increases.A er reaching the road design flow, the vehicle velocity and traffic volume

Conclusions
is study explores the in uences of the typical one-way tra c management on di erent vehicle exhaust emissions (CO, HC and NO x ) using an integrated method.Two scenarios of a road network (i.e., before and a er one-way tra c management) are established in the microscope tra c simulation platform VISSIM based on real road con gurations and tra c demands in peak hours.An instantaneous emission model (Vehicle Speci c Power) calibrated with eld emission data in Chinese city contexts is employed to assess the instantaneous emissions of vehicles.e di erent exhaust emissions (CO, HC and NO x ) in the intersections, road segments and network levels before and a er the implements of one-way tra c management are statistically compared for quantifying the e ects of one-way tra c management on di erent vehicle exhaust emissions.e ndings can be summarized as follows: (1) the CO, HC, NO x emission rates in the intersections a er one-way tra c management is signi cantly reduced by 20.46%, 21.29% and 21.06% as compared to those without one-way tra c management; (2) one-way tra c management decreases the CO, HC, NO x emission rates in the road sections by 22.19%, 23.38% and 26.29%, respectively; (3) the overall emission rates of CO, HC and NO x in the regional level decline by 21.34%, 22.29% and 23.77% due to one-way tra c management; (4) the implementation of one-way tra c management is benecial for improving the travel speeds of vehicles in the road sections and reduce the overall emissions even though some vehicles have to detour.e results provide insights into the potential e ects of one-way tra c management on tra c emission principles and contribute to the management concerning development of environment-friendly sustainable transport network.

F 1 :
Testing area and the driving path.(Images are from google map.)(1) e contribution rate of vehicular emissions is balanced in each VSP interval, and the emission sharing rate is the ratio of the emissions in each VSP interval to the total emissions.(2)e di erence of emission rates in di erent VSP internals should be obvious.(3)Adjacent VSP values are assigned to the same or adjacent BIN intervals.

F 2 :
e experiments, emissions testing setup and schematic.

T 3 :
Tra c conditions of intersections simulated in Synchro.

F 5 :
c it y (k m /h ) (c) Relationship among emissions, tra c ow and velocity.(a) CO emission analysis.(b) HC emission analysis.(c) NO x emission analysis.

T 6 :
Bypassing vehicles and speed in the road sections before and a er one-way traffic management.: the difference is equal to the emission a er setting one-way traffic management subtracted by the emission before.: the difference is equal to the emission a er setting one-way traffic management subtracted by the emission before. NoteNote