Supply Chain Management Journal

Transport Costs Optimization Using Up-To-Date Road Traffic Information Valentin IORDACHE Angel Ciprian CORMOŞ Florin Codruț NEMŢANU Politehnica University of Bucharest valentin.iordache @upb.ro Vicente Ramon TOMAS LOPEZ Universitat Jaume I Castellon

Abstract Activities like delivery of goods or passenger transport represents today a necessity and all the actors involved are pursuing their goals in reducing costs or effects of these activities. Minimizing traveling time and arriving in time at the destination, reducing fuel consumption and respecting schedules when is needed are benefits of future route planning systems. Achieving these benefits would not be possible without incorporating up-to-date road traffic information obtained from vehicle-to-vehicle and vehicle-to-infrastructure communication systems. In this paper we present a method of using traffic information, received on board of vehicles, to optimize calculation of a minimum cost route, and we propose a vehicle-infrastructure and vehicle-to-vehicle traffic data exchange system architecture to support the route planning systems. Keywords: intelligent transport systems, traffic data, route planning, V2I, V2V

Introduction One of the basic needs of humans was and is moving from one place to another. Currently, a transport process assumes an energy consumption, that, most of the times, cannot be reused or recovered, which leads to occurrence of costs. Transport costs are a monetary measure of what one provider must pay to produce transport services [J.P. Rodrigue et al., 2009]. Each individual or organization must take certain decisions when choosing a transport route so that its total cost is as low as possible. Any cost benefit of a degree of variability and uncertainty. Road transport costs may vary depending on various factors such as location, time, vehicle status, etc. Some factors can be determined with fairly good accuracy, for example the estimation of traffic on certain road sections at certain times of day based on previous traffic records, while others such as weather conditions or accidents are difficult or impossible to predict. Cost estimates are influenced by the limited sources of data, complex

32

modeling required often or entry into new research areas. Costs can be variable or fixed. For vehicles, variable costs, or marginal, are those that change with traveled distance, such as travel time cost, fuel consumption cost and vehicle ageing cost. Fixed costs are not affected by vehicle use and can be, for example, taxes (purchase, registration, insurance, etc.), although some may vary depending on vehicle age (insurance fee). The real costs of using a vehicle are not always perceived correctly. Often we give more weight to immediate costs such as the fuel and parking fees, underrating incidental costs generated by insurance fees, maintenance and repair or vehicle depreciation in value. Another type of cost partitioning can be done depending on the orientation of costs, i.e. internal, supported directly by the passenger, and external, that may affect him or not indirectly. Depending on the addressed problem type some costs may be external when one is considering only one vehicle or may be internal when taking into account a group of affected vehicles (i.e. the congestion cost).

2012, Volume 3, Number 1

Supply Chain Management Journal

Using a vehicle has multiple effects on both the driver or passenger, through visible travel costs supported by them, and on the other road participants, the environment or society in general. Costs of using a vehicle with average characteristics are illustrated in Figure 1.

Note that the highest costs are internal type, affecting directly the traveler and they are the most important ones when determining the cost of travel, while external costs, although low, are important when summed.

Figure 1. Values of costs generated by vehicle use

(source: [T.A. Litman, 2009]) Drivers make decisions on routes to follow considering some of the costs listed above (mainly internal) but also other factors that may determine their reduction: − fuel costs, are the most obvious and most important in a journey and are heavily dependent on the speed of travel and vehicle conditions (load, tire pressure, technical condition, etc.); − short-term parking costs are influenced by the availability of parking locations and cost of vehicle parking along the travel route; − travel time costs, taking into account the value of time; − vehicle maintenance costs: vehicle wear may increase depending on the traffic characteristics or road condition; − route comfort degree: a route affected by congestion, for example, increases the driver fatigue, thus increasing the risk of accidents; − risk of accidents: routes with high risk of accidents can be established based on statistics and dangerous areas can be determined;

2012, Volume 3, Number 1

− congestion degree, which may cause significant variations in fuel consumption and travel time or even in the driver's comfort state; − weather conditions, may cause an increase in travel costs when they are poor by increasing travel time, fuel consumption and the risk of accidents. In conclusion, travel cost may include a variety of direct and indirect costs can be influenced by a variety of factors, but the most obvious cost that the driver immediately feels is the fuel cost. An effective way to decrease this cost is represented by an accurate planning of travel routes, taking into account traffic conditions and vehicle's operating parameters when traveling on a specific route. Data transmission on board of a vehicle involves the implementation of dedicated communication systems, so called vehicular communications. Although their main purpose was initially to increase traffic safety and reducing accident effects, gradually, they found their role in avoiding incidents and finding the best travel routes by

33

Supply Chain Management Journal

processing traffic data, thus achieving economies in fuel consumption and time spent in traffic. Vehicular communications are those communications involving participants in traffic and is in fact a sum of communication between one vehicle and other vehicles (V2V) or between one vehicle and road infrastructure (V2I/I2V). They can provide information that helps to create an extended virtual horizon to allow drivers to benefit from an intelligent travel route planning taking into account existing road traffic. Outside the vehicle we can be obtained traffic information that can be historical, in real time or estimated, and within it information about the vehicle status and operating parameters. All these can be used when preparing the travel route, immediately before departure or during the journey. V2I communications allows the driver to access traffic information services that are able to provide currently various combinations of traffic data (volume, occupancy, average speed, travel time, etc.). V2V communications allows detecting traffic incidents (congestion, accidents, road-works, etc.) so that periods of time when traffic on certain road segments is difficult can be avoided. Intra-vehicle communications can be used to determine unique characteristics that define each particular vehicle, the state of the environment surrounding the vehicle at a time (road status and weather conditions, travel speed, vehicle load, fuel consumption, etc.) or allow communication with the driver to display or retrieve data. Given the above, determining the optimal route in terms of cost involves analysis of a large amount and variety of data, which is virtually impossible to be done by the driver. Thus, it is necessary to implement an integrated system to manage and process data using various algorithms.

1. Real time route planning systems The first generation of planning systems calculate their routes using static data, such as distance, legal speed limits, road types and average fuel

34

consumption. Advanced systems are able to incorporate the current traffic data in real time, allowing them to adapt to actual traffic conditions. A classification of such systems may be in the way that traffic data is provided on board of vehicle, which can be centralized or decentralized systems [W. Dong, 2011]. The difference between them is that the centralized systems require a traffic control center to provide information, being reliable and accurate but limited to a certain coverage area, while decentralized systems are based either on estimated traffic data pre-loaded into system memory or on data collected from other vehicles in traffic through V2V communications. Advantage of the latter is that they don't require constant communication with an information center or to pay a subscription fee to access the information, but with an obvious disadvantage in terms of accuracy. Principles, applications, and implementation of such systems have been addressed in numerous scientific papers. In [T.N. Tong and C.O. Lam, 1992] and [J. Wahle et al., 2001] authors propose solutions for travel routes planning systems using current and historical traffic data for their calculation and taking into consideration the effects of space-time traffic distribution over the travel times on a road network subject to congestion, and traffic data obtained from simulations along with real-time data. The beneficial effect of using traffic information was shown in [J. Fawcett & P. Robinson, 2000], the authors developing the Java programming environment, software that has managed to minimize traffic delays due to congestion. Also a web application was developed in [J. K. Ambrose et al., 2009], which was used to calculate several alternative routes using a probabilistic model based on historical traffic data. A methodology used in a route planning subsystem was presented in [J. Guzolek & E. Koch, 1989] using different optimization criteria, incorporating a variety of information including real-time traffic data, planning routes efficiently and quickly.

2012, Volume 3, Number 1

Supply Chain Management Journal

An implementation and testing of an advanced route guidance and information called LISB [J.M. Sparmann, 1991] was performed in Berlin, in the main intersections being installed transmitters and receivers to capture the measured travel times of vehicles and submit their recommendations on optimal routes, results being promising. In [T. Yamashita et al., 2004] the authors proposed a mechanism for sharing information on travel route of a vehicle with other vehicles in traffic, the study result is improved traffic efficiency for both individual users and whole system level. An approach to a similar idea, using communication systems between vehicles was presented in [T. Fukuda et al., 2002] with good results in urban areas. Regarding the current implementations of route planning systems using traffic data, there are some companies that provide equipment and especially necessary traffic information for accessing this type of services, and their number is growing larger. Volvo Navigation System (VNS) and the system used in Mercedes-Benz trucks are two examples of systems that use proprietary traffic data supplied via RDS-TMC2 and can plan routes based on traffic incidents. These systems however would require the existence of the service, plus they supply only a small number of incidents detected, not reflecting the exact traffic conditions. In Poland, Geosystems and Aqurat make available to drivers a planning route application called AutoMapa with two options that take into account traffic parameters, LiveDrive! and SmartRoutes. LiveDrive! option allows to acquire traffic incidents information and updating internal databases using a mobile Internet connection, while SmartRoutes option uses traffic conditions estimates based on statistical data provided by users of the road [AutoMapa, 2012]. Garmin has brought to market Garmin Traffic service that uses HD Radio technology to transmit traffic information in digital format, data being transmitted in conjunction with analog

AM or FM radio signals. Traffic data is collected from two billion sources, owners of Garmin devices, cell phones, reports of incidents, RDS-TMC sources, historical data provided by NAVTEQ, traffic sensors, etc. [Garmin, 2012]. TomTom offers its users a paid service, TomTom HD Traffic, using a GPRS connection to send on board of vehicles traffic information updated every 3 minutes. TomTom dedicated equipment can be both portable or embedded in a vehicle offering a wide variety of options and features. To obtain traffic information, in addition to conventional data sources and data supplied by vehicles equipped with TomTom devices, they implemented a data collection technology called CFCD 3 that uses signals from in-vehicle GSM phones involved in active calls, to locate their position [TomTom, 2012]. Current systems obtain traffic data, almost exclusively, by using communications with road infrastructure. The approach in this paper is an integrated route planning system that uses multiple-vehicle communications systems (V2I, V2V and intra-vehicle) to get more data on traffic conditions that may influence the travel route (fuel consumption, speed and travel time, traffic incidents, weather conditions, etc.). On this basis, the system will be able to accurately plan travel routes with minimum cost.

2

3

Radio Data System - Traffic Message Channel

2012, Volume 3, Number 1

2. Reducing costs through route planning optimization Road transport network can be represented as a graph. A graph G is an abstract representation of a set of objects, some of which are connected by links. Objects are called nodes (e.g. roads intersections or other significant points on the road), and connections between them are called arcs (road segments between nodes). Determination of minimum cost travel routes thus involves building a graph and use an algorithm to extract the route according to certain conditions imposed, using solutions for the problem of finding least cost path. Determination of least Cellular Floating Car Data

35

Supply Chain Management Journal

cost path r(s,d) assumes the analysis of all possible paths between source s and

destination d.

Figure 2. Graph example

Considering the cost of a route as being the fuel consumption and that it is directly influenced by the speed of travel, we establish that route planning requires four types of data, as a minimum necessary: • fuel consumption for all travel speeds; • length of each road segment; • travel time on each road segment; • average travel speed on each road segment. In addition to these parameters, other traffic data or data related to the

vehicle can be taken into account, which can influence the route, travel speed or fuel consumption. 2.1. systems

Data

provided

by

V2I

Real-time traffic data (RTTD) provides a concrete traffic image and it is used more and more in various areas of Intelligent Transport Systems. This kind of data is collected by sensor networks from the road level and is sent to users in the shortest possible time.

Figure 3. Information on traffic conditions

(source: http://pems.dot.ca.gov) The main traffic information is provided by V2I communications from Traffic Management Centers. Data provided by a wide variety of sources is collected and processed by an integrated

36

traffic management information system. TMC integrates the data collected, stores the information in databases and provides the necessary interface for using them. A good example is

2012, Volume 3, Number 1

Supply Chain Management Journal

California's Freeway Performance Measurement System operated by the California Department of Transportation and the University of California, Berkeley, who make traffic data available to any type of user, free of charge. In addition to TMC's, private data traffic providers exists, which usually combines information obtained from TMC's with information from other sources or from own sources (own sensors network). The most important of them, at international level, are INRIX, NAVTEQ, and TomTom. The main interest information, to plan a travel route, that can be obtained through V2I communication is: average travel speed, travel time, information on traffic incidents. In addition, facilities can be provided for services such as obtaining information on points of interest, parking spot reservation, to pay certain fees, etc. 2.2. systems

Data

provided

by

V2V

Route planning systems using traffic data provided by road infrastructure are effective as long as the vehicle is in contact with sources of data. Lack of these sources leads to uncertainty of data that the system has at its disposal. Informing the driver on traffic conditions along the route and recommending an alternative route, so that he can avoid periods of time when traffic on certain road segments is difficult, can be done using V2V communications. Thus, by creating peerto-peer networks we can get near-by information from oncoming vehicles or real time information from other vehicles able to communicate with the infrastructure by ad-hoc networking. Data that can be transmitted using V2V communications can be

obtained using in-vehicle sensors and they can provide important information about traffic incidents. A big advantage of this type of communication is the possibility of locating incidents with high precision, and knowing exactly when they occurred. A road incident is defined as a planned or unplanned event that affects the traffic flow of a road [S. Vaneet et al., 1994]. Incident types include congestion, accidents, road-work, damaged or abandoned vehicles, road obstacles. Traffic incidents adversely affect road level of service, drastically reduce its capacity and sometimes represents a danger to road users. Incident detection is the process of determining their presence and location and it is based on data from invehicle systems and from other vehicles in traffic, data analysis being accomplished by using automatic incident detection systems (AID). Next we describe three types of incidents that can be detected by in-vehicle sensors. Congestion is mainly characterized by low travel speeds, stopand-go's, frequent stops, reduced distances between vehicles. Their detection can be done using information provided by several types of systems and sensors on board of vehicles involved in congestion: speed sensor, deceleration sensor, measuring the distance between vehicles with ACC systems or pre-crash systems, etc. Data can be transmitted periodically using V2V communication systems to other vehicles. In this way, any vehicle can record flow characteristics of oncoming traffic in an internal database and can determine, using algorithms, the presence of a congestion in traffic, by dividing a segment of road in three areas (Figure 4).

Figure 4. Dividing a segment of road in three areas

2012, Volume 3, Number 1

37

Supply Chain Management Journal

Accidents detection can be done using information provided by several types systems and sensors on board of the vehicle involved in an accident: airbag system, pre-collision system, acceleration sensors, rotation sensors, emergency lights. However, additional data are needed, to know if the accident is affecting the traffic flow. Closure of certain roads or lanes, or road-works detection can be achieved in several ways: − through other information channels, such as systems using RDS-TMC, when location and duration of work are precisely known; − using beacons placed in locations affected by road-works to transmit the information to other vehicles through communications over short distances. In this case, also, location and duration of work are precisely known; − using individual detection systems based on automatic recognition of road signs, on images provided by on board cameras. In this case we only know the precise location of the affected area. 2.3. Data vehicle systems

provided

by

in-

route, when received from other vehicles in traffic. Measuring vehicle speed is necessary because fuel consumption depends largely on it. As speed increases, the power required to overcome air friction forces increase rapidly, adding, also, the growing power needed to defeat the resistance forces of tire rolling. Increasing the power generated by the engine means increasing fuel consumption. In addition, depending on the vehicle speed we can determine the travel time at each moment. Vehicle load degree has a direct impact on fuel consumption, weight gain resulting in increasing vehicle tire rolling resistance, thus increasing fuel consumption. For freight vehicles, determining the load can be performed using sensors which measure the load on each axle, or by weighing the vehicle with its external systems. For passenger vehicles we can use sensors to determine the presence of them on seats or the user can manually enter in the system the number of passengers. In-vehicle sensors can measure instant fuel consumption and remaining quantity in the tank to determine the maximum travel distance and recommend refueling stops. 3. System architecture

Due to its performances, the vehicle itself can be considered a significant source of information about factors that can influence the process of route planning. Unique characteristics that define each particular vehicle, or the environment state surrounding the vehicle, can be determined using data from in-vehicle sensors and systems. Travel speed depends heavily on road state. The presence of water, snow or ice reduces safe driving conditions and consequences are obvious, decrease of travel speeds. Warning drivers of vehicles on road state is done through implementation of on board systems that can monitor road condition. Information may be used to update the corresponding average speed on that road segment, when the information is obtained from its sensors, or on other road segments that are part of the travel

38

Being established what type of data can be used by the route planning system, to allow a more precise determination of the minimum cost route, we can establish a general system architecture, both for the vehicle and at the infrastructure level. A vehicle-infrastructure communication system, taking into account the needs of a dynamic route planning system, must include the following components (Figure 5): − On Board Equipment (OBE), available inside the vehicle and responsible with the data exchange between vehicle's route planning system and road infrastructure. − Road Side Equipment (RSE), acts like a gateway between vehicles and the traffic data and service provider's systems. Can, also,

2012, Volume 3, Number 1

Supply Chain Management Journal

collect data from sensors along the road. − Road Side Unit (RSU), responsible for the data exchange with the RSE's and the decision making process.

− Traffic Management Center (TMC), the supplier of real time traffic data.

Figure 5. Components of a V2I communication system

Complexity of the system depends on many factors. For example, we can choose for a full coverage of a road segment, which can be done with a large number of RSE's, or we can choose certain distances between 2 RSE's, depending on the necessity of the traffic data exchange, and minimize the number of them needed. How many RSE's we need to cover a certain area depends on the communication technology used. Suitable technologies for data exchange with moving vehicles can be WiMAX, LTEor DSRC. WiMAX and LTE works over tens of kilometers and DSRC only over hundreds. Covering distance may be influenced by road and land

2012, Volume 3, Number 1

characteristics. In areas whit heavy traffic there may be a need for more RSE's in the same location to serve the large number of vehicles. It is necessary to have a RSE before every location that can influence the travel route, like gas station, parking, intersection or highway exit. Also, critical locations like those with high risk of accidents (tunnels, bridges, etc.) or congestion should be covered with RSE's to rapidly detect incidents and alert incoming vehicles. The proposed system architecture for data exchange between vehicles and the road infrastructure is presented in Figure 6.

39

Supply Chain Management Journal

Figure 6.Vehicle to road infrastructure data exchange system architecture

An intra-vehicle communication system, that takes into account the needs of a dynamic route planning system, must include the following components: − V2I and V2V capable communication equipment (DSRC); − gateway to the vehicle's communication bus (CAN, LIN), to allow communication with invehicle subsystems; − GPS unit to allow vehicle localization;

a route planning system that can retrieve data from in-vehicle subsystems and sensors; − sensors and subsystems to measure various traffic parameters; − on board computer able to process and store information; − a human-machine interface (HMI) that allows the driver to interact with the system. The proposed system architecture for intra-vehicle data exchange is presented in Figure 7. −

Figure 7.Intra-vehicle data exchange system architecture

40

2012, Volume 3, Number 1

Supply Chain Management Journal

Conclusion Whether we talk about transporting passengers or freight, each vehicle driver or operator seeks to minimize transport costs. One of the most important is the cost of fuel consumption. A method of reducing it is the optimal planning of travel routes using real-time traffic information so as to avoid possible adverse traffic conditions. Current dynamic route planning systems are based solely on information obtained from road infrastructure, mainly due to the fact that V2V communication systems are still in early stages of testing and implementation. V2I communications has the great disadvantage of the need to develop infrastructure and sensor networks to cover the road network, in order to accurately monitor road conditions. Installation, management and maintenance are costly, traffic information is limited to roads covered by the sensor systems and to avoid loading the environment of communication, transmissions are made from time to time, which reduces system efficiency. V2V communication systems eliminate these disadvantages by using point-to-point communications and ad hoc networks between vehicles, to form a picture of traffic conditions. The system proposed in this paper optimizes the travel route, and obtain the lowest cost in terms of fuel consumption by using real-time traffic data provided by road infrastructure, the other vehicles in traffic and the host vehicle. It is also considered using data from host vehicle's systems to know and estimate the actual fuel consumption (travel costs), for different values of travel speed and varying load. With all this information, planning travel routes with minimum cost can be made with a high degree of precision and confidence. References Ambrose, J.K., Bukovsky, D.J., Sedlak, T.J., Goeden, S.J. (2009) Developing a travel route planner accounting for traffic variability, in Systems and Information

2012, Volume 3, Number 1

Engineering Design Symposium, SIEDS '09, Charlottesville, VA, p. 264-268 Automapa Traffic (2012) [Internet] http://www.automapa.pl Dong, W. (2011) An overview of invehicle route guidance system, in Proceedings of Australasian Transport Research Forum 2011, Adelaide, Australia Fawcett, J. Robinson, P. (2000) Adaptive routing for road traffic, IEEE Computer Graphics and Applications, vol. 20, no. 3, p. 4653 Fukuda, T., Takefuji, K., Ikemoto, Y., Hasegawa, Y. (2002) Dynamical route-planning for vehicles based on global traffic information and communication, in Proceedings of The IEEE 5th International Conference on Intelligent Transportation Systems, 2002, pg. 538-543 Garmin Traffic (2012) [Internet] http://www8.garmin.com/traffic/ Guzolek, J., Koch, E. (1989) in Vehicle Navigation and Information Systems Conference, 1989, Toronto, Canada, p. 165-169 Litman, T.A. (2009) Transportation Cost and Benefit Analysis: Techniques, Estimates and Implications, Second Edition, Ed. Victoria Transport Policy Institute Rodrigue, J.P., Comtois, C., Slack, B. (2009) The Geography of Transport Systems, Second Edition, New York, Ed. Routledge Sparmann, J.M. (1991) Benefits of dynamic route guidance systems as part of a future oriented city traffic management system, in Vehicle Navigation and Information Systems Conference 1991, Volume 2, p. 839-847 TomTom International BV. (2012) How TomTom’s HD Traffic™ and IQ Routes™ data provides the very best routing, White paper Tong, T.N., Lam, C.O. (1992) A dynamic route guidance system based on historical and current traffic pattern, in Proceedings of the Intelligent Vehicles '92 Symposium, Detroit, MI, p. 365-369

41

Supply Chain Management Journal

Vaneet, S., Koppelman, F.S., Flannery, C.P., Schofer, N., Bhandari, J.L. (1994) Duration and Travel Time Impacts of Incidents - ADVANCE Project, Northwestern University, Evanston, IL, Technical Report TRF-ID-202 Wahle, J., Annen, O., Schuster, Ch., Neubert, L., Schreckenberg, M. (2001) A dynamic route guidance system based on real traffic data,

42

European Journal of Operational Research, vol. 131, no. 2, p. 302308 Yamashita, T., Izumi, K., Kurumatani, K. (2004) Car navigation with route information sharing for improvement of traffic efficiency, in Proceedings of The 7th International IEEE Conference on Intelligent Transportation Systems, 2004, p. 465-470

2012, Volume 3, Number 1