Acquisition of Relative Trajectories of Surrounding Vehicles using

GPS and DSRC based V2V Communication with Lane Level

Resolution

Zhiyuan Peng

, Shah Hussain

, M. I. Hayee

and Max Donath

University of Minnesota Duluth, Duluth, MN 55812, U.S.A.

University of Minnesota Twin Cities, Minneapolis, MN 55455, U.S.A.

Keywords: Connected Vehicle Technology, V2V Communication, DSRC, Relative Trajectory, Relative Distance

Accuracy.

Abstract: Due to the anticipated benefits of connected vehicle technology, the Intelligent Transportation Systems Joint

Program Office (ITSJPO) of the US Department of Transportation continues to emphasize the need for

dedicated short range communication (DSRC) based vehicle-to-vehicle (V2V) and/or vehicle-to-

infrastructure (V2I) communication to enhance driver safety and traffic mobility. To take full advantage of

connected vehicle technology in most safety applications, precise vehicle positioning information is needed

in addition to V2V communication. Many techniques, such as vision- or sensor-based systems and

differential GPS receivers, can obtain the precise absolute position of a vehicle at the expense of cost and

complexity. However, some critical safety applications such as merge-assist or lane-change-assist systems

require only the relative positions of surrounding vehicles with lane-level resolution so that a given vehicle

can differentiate the vehicles in its own lane from the vehicles in adjacent lanes. We have adopted a simple

approach to acquire accurate relative trajectories of surrounding vehicles using standard GPS receivers and

DSRC-based V2V communication. Using this approach, we have conducted field tests to successfully

acquire relative trajectories of vehicles traveling in multiple lanes towards a merging junction with an

accuracy less than half of the lane width. The achieved accuracy level of the relative trajectory was

sufficient to differentiate vehicles traveling in adjacent lanes of a multiple-lane freeway.

1 INTRODUCTION

The Intelligent Transportation Systems Joint

Program Office (ITSJPO) of the US Department of

Transportation (USDOT) continues to be committed

to the use of dedicated short range communication

(DSRC) for active safety applications using vehicle-

to-vehicle (V2V) and/or vehicle-to-infrastructure

(V2I) communication due to its designated licensed

bandwidth, fast network acquisition, and low latency

(USDOT factsheet a, b). A USDOT research report

estimates that V2V communication has the potential

to help drivers avoid or mitigate 70 to 80 percent of

vehicle crashes involving unimpaired drivers, which

could help prevent thousands of deaths and injuries

on roads every year (Harding, J., Powell, et al 2014,

NHTSA USDOT factsheet). To take full advantage

of the potential safety benefits of connected vehicle

technology, relative trajectories of the surrounding

vehicles with lane-level resolution are needed in

addition to V2V communication (D. Jie and M. J.

Barth, 2008). Accurate positioning information with

lane-level resolution can enable many vehicular

safety applications (e.g., freeway merge-assist, lane-

change-assist, and lane-departure warning systems),

which could potentially help avoid many crashes (S.

Ammoun, F. Nashashibi et al. 2007, D. Desiraju, T.

Chantem 2015). According to one study, 36 percent

of the freeway accidents analyzed occurred on

entrance ramps, and another study reported that 20–

30 percent of total truck accidents nationwide occur

on or near ramps (A.T. McCartt et al. 2004, Bruce

N. Janson et al. 1998). Similarly, in 1991, lane-

change accidents accounted for approximately 4

percent of all police-reported crashes that occurred

in the United States; in 1999, those accidents rose to

9 percent (Basav Sen et al. 2003, G. M. Fitch et al.

2009). Another report that analyzed crash data from

2005 to 2007 concluded that 11 percent of vehicles

242

Peng, Z., Hussain, S., Hayee, M. and Donath, M.

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution.

DOI: 10.5220/0006304202420251

In Proceedings of the 3rd International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2017), pages 242-251

ISBN: 978-989-758-242-4

involved in an accident had failed to stay in the

proper lane (NHTSA: Report to congress, 2008).

An important technological milestone in the

development of a lane-change or merge-assist

application is to acquire the relative positions of

surrounding vehicles in real time. Accurate

positioning information can be obtained using either

sensor-based systems or Global Navigation Satellite

Systems (GNSSs). Both approaches have their

limitations. Sensor-based systems utilize vision- or

laser-based sensors to acquire the relative positions

of surrounding vehicles (D Chun.; K, Stol. 2012,

Abdelfatah, W.F., et al. 2011, Qingquan Li et al.

2014, H. Zhao et al. 2009). However, environmental

factors such as weather, variable lighting conditions,

absence of line-of-sight (LoS), or worn out road

markings can adversely affect the performance of

these systems (A. Bansal et al. 2014). Similarly,

GNSS-based technologies such as Global

Positioning System (GPS) cannot predict the

position of a vehicle with lane-level accuracy

without using a correction or augmentation system

e.g., differential GPS technology, inertial sensors,

gyroscope, and/or high-resolution maps (R. Toledo-

Moreo et al. 2007, N. Mattern et al. 2010, R. G.

García-García et al. 2007, J. Juang et al. 2015, S.

Rezaei and R. Sengupta 2007). Furthermore, the

deployment of either sensor-based or GPS-based

system requires sophisticated hardware and

software, resulting in increased complexity and

higher overall costs.

The above-mentioned techniques can obtain the

precise absolute position of a vehicle at the expense

of cost and complexity. However, some critical

safety applications such as merge-assist or lane-

change-assist systems require only the relative

positions of surrounding vehicles with lane-level

resolution to allow a given vehicle to differentiate

the vehicles in its own lane from the vehicles in

adjacent lanes (N. Alam et al. 2013). Therefore, in

the approach presented in this paper, we have

focused on acquiring the relative trajectories of

surrounding vehicles using standard GPS

receivers—without any additional correction

system—and DSRC-based V2V communication.

Our approach to acquire relative trajectories is

based on the fact that a major part of GPS

positioning error, caused by atmospheric effects, is

highly correlated over a vast geographical area (J.

Farrell, T. Givargis 2000, FHWA factsheet).

Therefore, multiple GPS receivers of the same kind

on different vehicles in close proximity tend to have

a similar atmospheric error at a given time. The

common atmospheric error could be canceled out to

obtain a more accurate estimate of the relative

distance between any two vehicles as compared to

the absolute position of each vehicle. Utilizing this

approach, we have successfully acquired relative

trajectories of vehicles traveling in multiple lanes

toward a merging junction with an accuracy of less

than half of the lane width using DSRC-based V2V

communication and standard GPS receivers. The

accuracy of the acquired relative trajectory was

sufficient to differentiate vehicles traveling in

adjacent lanes of a multiple-lane freeway.

The next section of this paper will describe the

concept of relative GPS positioning among

surrounding vehicles. The following section will

discuss the results from field tests to statistically

characterize the accuracy of the relative trajectories

of multiple vehicles using standard GPS receivers.

In the next section, the results from field tests to

acquire relative trajectories of surrounding vehicles

with lane-level accuracy using DSRC-based V2V

communication will be discussed. The last section

will summarize the conclusions.

2 CONCEPT OF RELATIVE GPS

POSITIONING AMONG

SURROUNDING VEHICLES

Our approach utilizes standard GPS receivers and

DSRC-based V2V communication to acquire the

relative trajectories of surrounding vehicles. The

absolute position accuracy of a standard GPS

receiver is in the range of 3–5m (William J. Hughes,

2014). This means that a GPS receiver can estimate

the position of a vehicle within a circle with a radius

of 3–5m, as shown in Figure 1a, where the true

position of the vehicle at a given time is shown by a

green dot and the red dot shows the estimated

position by the GPS receiver. The error vector from

the true position to the estimated position represents

the GPS position error. The total GPS position error

is a combination of multiple errors resulting from

different sources. Generally, the combined GPS

position error is a result of three major errors:

mechanical error, satellite ephemeris error, and

atmospheric error.

The mechanical GPS error is caused by inherent

noise or clock jitter of the crystal oscillator used in

the GPS receiver, thermal effects, manufacturing

differences, and residual mathematical error due to

quantization and rounding (D. K. Schrader 2013, R.

B. Langley 1997). Satellite ephemeris error is due to

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution

243

Figure 1: Conceptual GPS receiver error model of a single GPS receiver showing ranges of different GPS error types for (a)

a stationary vehicle at a single time instance and (b) a moving vehicle at three adjacent time instances.

the fact that the expected orbital positions of the

GPS satellites that the GPS receiver needs to

estimate its own position, could be different than

actual satellite positions. Atmospheric error, the

most significant portion of the combined GPS error,

is caused by atmospheric effects that cause the GPS

signal to bend while it travels through the

atmosphere. Of all three errors, mechanical error is

the only one that can vary randomly from one GPS

receiver to another at any given time. It can also

vary in the same GPS receiver with each subsequent

position estimate over time. On the other hand, both

ephemeris and atmospheric errors do not vary

significantly for multiple GPS receivers in close

geographical and temporal proximity. This is

because atmospheric disturbances will remain the

same over a wide geographical area and do not

rapidly change with time (J. Farrell, T. Givargis

2000, FHWA factsheet). Similarly, ephemeris error

will remain almost the same for the satellite

constellation used by GPS receivers in close

proximity to each other (Ahmed El-Rabbany 2002).

Theoretically, a GPS-estimated position can be

anywhere in the larger circle as shown in Figure 1a,

representing the range of combined GPS error.

However, after a GPS receiver gets locked to certain

satellites to estimate its position, its subsequent

position estimates will not randomly vary over the

entire large circle because atmospheric and

ephemeris errors will remain the same for a

considerable period of time. On the other hand,

mechanical error can randomly vary in every new

position estimate in any GPS receiver. The size of

mechanical error is comparatively much smaller

than the other two errors, which is highlighted by the

relative sizes of the two circles in Figure 1a.

Therefore, subsequent estimates of the same position

by a given GPS receiver will remain confined to a

smaller circle shown in the Figure 1a, representing

the range of mechanical error.

In addition to the three errors described above,

multipath error can significantly degrade the

position estimation accuracy for any GPS receiver.

Multipath error occurs when GPS signals arrive at

the receiver antenna through multiple paths as a

result of reflections from surrounding objects (e.g.,

high-rise buildings or overhead bridges) (T. Kos et

al. 2010). Multipath error is significant in urban

areas where a roadway is surrounded by high-rise

buildings. However, in rural and suburban areas,

multipath error can be negligibly small and the

significant errors are mechanical, ephemeris, and

atmospheric, as described above.

Figure 1a illustrated GPS receiver errors in static

conditions. When such a GPS receiver is placed in a

moving vehicle, it can be used to acquire a vehicle’s

VEHITS 2017 - 3rd International Conference on Vehicle Technology and Intelligent Transport Systems

244

trajectory by periodically estimating its position.

This concept is illustrated in Figure 1b, where three

adjacent GPS positions of a fast-moving vehicle on a

freeway (with minimal multipath error) are shown as

red dots. Each adjacent estimated position will vary

only within the small circle (the mechanical error

range) as opposed to randomly changing over the

larger circle because the atmospheric and ephemeris

errors will remain the same for each estimate.

Consequently, the trajectory obtained by the GPS

receiver may vary randomly, but the maximum

variations will be limited to the zigzag pattern

shown in Figure 1b. The mean trajectory obtained by

the GPS receiver (shown by the red dashed line) will

have an offset from the true trajectory (shown by the

green dashed line), but it will be a fixed offset and

its size will be determined by the magnitude of net

atmospheric and ephemeris error. Furthermore, the

variance of the trajectory obtained by the GPS

receiver will be determined by the magnitude of the

mechanical error of the GPS receiver, which is

generally small in size.

Similar to the trajectory of a single vehicle,

which can be obtained by a GPS receiver with a

small variance, the relative trajectories of multiple

vehicles in close proximity that have their own GPS

receivers can also be obtained with comparable

variances. Two practical scenarios involving

multiple vehicles—merging and changing lanes on

freeway—are depicted in Figure 2 (left side). In both

scenarios, the relative trajectories of surrounding

vehicles, if accurately known, can be beneficial in

the development of traffic safety applications. Using

the GPS error model described above, the relative

positions of three vehicles obtained by GPS

receivers are shown in Figure 2 (right side) at a

given time. The estimated GPS position of each

vehicle (shown by red dots) will have the same

offset from the true position because the net

atmospheric and ephemeris error remains the same

for all three vehicles—provided they are equipped

with GPS receivers of the same model. Therefore,

the relative distance between any two vehicles in

Figure 2: Concept of relative GPS accuracy: (a) Lane-merging scenario (b) Lane-changing scenario.

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution

245

both scenarios calculated from the estimated

positions of the GPS receivers on the two vehicles

will have a small variance determined by the

mechanical errors of the GPS receivers. An accurate

estimate of relative distance between any two

vehicles at a given time can lead toward an accurate

estimate of the relative trajectories of those vehicles

with respect to each other. The accuracy of the

relative trajectories needs to be high enough for use

in a potential safety application, such as a lane-

merge or lane-change-assist system, where it is

necessary to determine if a neighboring vehicle is in

the same or adjacent lane.

3 CHARACTERIZATION OF THE

GPS RELATIVE DISTANCE

ACCURACY

The relative trajectories of surrounding vehicles can

be obtained for any given vehicle on the road

provided it can receive the estimated GPS positions

of the neighboring vehicles. We used DSRC-based

V2V communication to exchange position

information among surrounding vehicles that had

standard GPS receivers, which allowed GPS position

data from neighboring vehicles to be processed in

any vehicle to obtain relative trajectories.

Before conducting field tests to obtain relative

trajectories of multiple vehicles on the road, the

relative distance accuracy of the standard GPS

receivers built in to the DSRC devices needed to be

characterized to determine if it is sufficient to

distinguish the neighboring vehicles in the same or

adjacent lanes. Therefore, we statistically

characterized the relative distance accuracy of the

GPS receivers built in to the DSRC devices and later

used the same devices to acquire the relative

trajectories of multiple vehicles using DSRC-based

V2V communication. The built-in GPS receivers use

a UBlox LEA-6 chipset, which is specified as having

a ± 2m absolute position accuracy with 50 percent

circular error probability (CEP). Using these GPS

receivers, we have been able to achieve the relative

distance accuracy of ± 0.5m with 95 percent CEP in

our field tests.

We conducted field tests to statistically evaluate

the accuracy of the relative distance obtained by the

built-in GPS receivers of the DSRC devices. We

installed antennas for three DSRC devices on top of

one vehicle at locations A, B, and C, as shown in

Figure 3. A top view of the vehicle used for the field

tests is shown in Figure 3a, and Figure 3b is a top-

view schematic of the vehicle showing the three

antenna locations (A, B, and C). The three locations

formed a right-angle triangle with two shorter legs

of length 1m each. We drove the equipped vehicle

on I-35 near Duluth, MN, in a round trip between

exit #239 and #242 at a speed of about 70 MPH

(speed limit) while continuously acquiring GPS

position data in all three devices at the rate of 10 Hz.

We repeated the round trip six times, exchanging

the positions of the antennas at locations A, B, and C

after each trip and using all six possible

permutations of the three devices. Each round trip

produced three distinct sets of acquired GPS

positions (one for each GPS receiver at location A,

B and C) in terms of longitude and latitude at

distinct time intervals synchronized with the GPS

satellite time. There were more than 12,000 GPS

points in each of the three sets of data (i.e., a net 20

minutes’ worth of data with 10 Hz GPS acquisition

rate). We then processed the data from all three

DSRC devices to calculate three distances (AB, BC,

and AC) for each set of three GPS points acquired at

the same time because the clock of each GPS

Figure 3: The top view of the vehicle used for the field tests with (a) pictorial view and (b) schematic view, showing three

installed antennas and their relative locations.

VEHITS 2017 - 3rd International Conference on Vehicle Technology and Intelligent Transport Systems

246

Figure 4: Average calculated distances of segments AB, BC, and AC. The histogram of each segment length is shown

beside the segment. The average angle ∠ABC is 87.8 degrees.

receiver was synchronized with the GPS satellite.

The calculated average distances of AB, BC, and

AC were 1.15, 1.16, and 1.6m, with standard

deviations of 0.21, 0.20, and 0.24m, respectively, as

shown in Figure 4. The calculated average distances

of AB, BC, and AC are shown in Figure 4 where a

circle with a 0.25m radius is drawn at each location

(A, B, and C) to indicate the spread of the calculated

relative distance because the standard deviation of

each calculated distance is less than 0.25m. The

variation of the relative distances of AB, BC, and

AC is within ± 0.5m most of the time (>95%), as

illustrated in the histogram of each segment in

Figure 4. Furthermore, the histograms show that the

maximum spread of each relative distance is within

a ± 0.6m limit (1.2m total spread), which is still less

than half of the lane width, and therefore, is

sufficient to differentiate vehicles on adjacent lanes.

Although the specified absolute position

accuracy of each GPS receiver used was ± 2m with

50 percent CEP, the relative position accuracy

between any two GPS receivers was much improved

because the net ephemeris and atmospheric error in

absolute position was similar in all three GPS

receivers and was therefore cancelled out in the

relative distance calculation.

In our approach to characterize relative distance

accuracy, we used standard GPS receivers of the

same hardware and firmware model. This was

necessary because the post processing of the GPS

signal may vary among different GPS chips being

used on different DSRC devices. The processing

Figure 5: (a) The schematic diagram of calculated headings of the two GPS receivers at locations A and B, and (b) the

histogram of the differential heading.

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution

247

algorithm may also be different among different

versions of firmware on the same kind of GPS chip.

Furthermore, the GPS receiver’s field of view is

wide enough to receive signals from more than three

or four GPS satellites, which is the minimum

number of satellites required for two-dimensional or

three-dimensional position calculation, respectively.

In such scenarios, unless the post-processing

algorithm of multiple GPS receivers is designed to

lock to the same set of satellites, it is not guaranteed

that the atmospheric and ephemeris errors will

remain the same in each GPS receiver—thereby

adversely affecting the relative distance accuracy.

We experienced this phenomenon only twice during

our early field tests when the offset of at least one of

the three GPS receivers used was different from the

others, indicating that this particular GPS receiver

locked to a different set of satellites. In the built-in

GPS receivers of our DSRC devices, we did not

have any access to modify the GPS receiver

firmware to make it lock to a particular set of

satellites. However, we did not experience this

phenomenon in any of our subsequent field tests,

including the tests described in this paper.

We also evaluated the directional accuracy for

each of the GPS receivers in this field test. We took

two consecutive GPS positions (100msec apart in

time) for each of the two GPS receivers at locations

A and B and calculated individual headings for both,

as shown in Figure 5a. Figure 5b shows the

histogram of difference in headings of the GPS

receivers at positions A and B for all available data

points, covering six possible pairs of three distinct

GPS receivers at two locations (A and B). The

average and standard deviation of the differential

heading is -0.003 degrees and 0.26 degrees,

respectively. Both GPS receivers were traveling in

the same direction, so the differential heading was

expected to be zero. The results show that a standard

GPS receiver can estimate the direction of travel

with an accuracy of a quarter of a degree which is

sufficient for use in a safety application e.g., a lane-

change or merge-assist application. This is because

a quarter of a degree mismatch between the actual

and expected direction of travel of a vehicle

traveling at 60 MPH will cause a displacement error

of about 11cm in its expected position after one

second.

Figure 6: A typical scenario from field tests showing relative trajectories of three vehicles around a merge junction of a two

lane freeway (I-35). The lower part of the Figure is the zoomed version of a smaller area in upper part showing accuracy of

the acquired relative trajectories.

VEHITS 2017 - 3rd International Conference on Vehicle Technology and Intelligent Transport Systems

248

4 RELATIVE TRAJECTORY

ACQUISITION USING DSRC-

BASED V2V

COMMUNICATION

After statistically characterizing the relative distance

accuracy for the built-in GPS receivers of the DSRC

devices, we acquired relative trajectories of multiple

vehicles using DSRC-based V2V communication.

We installed three DSRC devices with built-in GPS

receivers on three separate vehicles that were

programmed to transmit and receive DSRC-based

Basic Safety Messages (BSMs). Using those

vehicles, we conducted field tests to demonstrate the

acquisition of accurate relative vehicle trajectories

traveling in different lanes.

We conducted the field tests around Exit #239

on I-35 in Duluth, MN, which is a two-lane freeway.

One of the vehicles waited on the entrance ramp of

Exit #239 to merge on the freeway while the other

two vehicles travelled on the freeway toward the

merging junction on two separate but adjacent lanes.

When the two vehicles approached the merging

junction, the vehicle waiting at the entrance ramp

started to receive DSRC messages from the vehicles

on the main freeway. Upon receiving the first

message, the vehicle started to move and merged

onto the freeway while continuing to receive DSRC

messages from the two vehicles on the main

freeway. The vehicle on the entrance ramp logged

all of the received DSRC messages. This data was

later analyzed to obtain relative trajectories of all

three vehicles. We repeated the tests at least 12

times; each time, the acquired relative trajectories of

the vehicles were accurate enough to identify each

vehicle in its own lane.

One typical scenario of the field tests is shown in

Figure 6, where the acquired relative trajectories of

three vehicles are drawn in three different colors: red

for the vehicle traveling on the entrance ramp and

blue and green for the vehicles traveling on the main

freeway in two adjacent lanes. The relative

trajectories are superimposed onto Google Maps to

establish a frame of reference. A zoomed-in version

of the relative trajectories near the merge junction is

also shown in Figure 6, illustrating that lane-level

accuracy can be achieved using the built-in standard

GPS receivers of the DSRC devices.

To measure the range of the V2V

communication during the field tests, we calculated

the distance between the vehicles on the main

freeway and the vehicle on the entrance ramp when

that vehicle received the first DSRC messages from

each of the two vehicles on the main freeway. The

measured DSRC ranges for the DSRC devices on

the two vehicles in the test scenario of Figure 6 were

182 and 312m, respectively. In the rest of the tests,

the DSRC range typically varied between 200–

300m. The specified DSRC range is >500m (D.

Jiang 2006) when a clear line of sight is available,

but the actual achieved range (200 – 300m) was

reduced due to some natural growth around the

merge junction that caused some loss of signal

strength.

Although the relative trajectories obtained in the

field tests have lane-level accuracy, these trajectories

were obtained by post-processing GPS data acquired

through DSRC-based V2V communication during

the field tests. In the future, we plan to integrate the

post-processing algorithm within the DSRC device

of the vehicle on the merging ramp to acquire the

relative trajectories in real time. Using the real-time

trajectories, speed, and direction of travel

information from the relevant vehicles, we can

estimate a safe merge time cushion that could

potentially be used as an important parameter to

develop a merge-assist application.

We define the merge time cushion as the time it

will take for a vehicle in the rightmost lane of the

freeway to arrive at the merging junction after the

vehicle on the entrance ramp has received the first

BSM from this vehicle. The merge time cushion for

the field test result of Figure 6 was estimated to be

between 9 and 10 seconds, as illustrated in Figure 7,

Figure 7: A field test scenario showing the relative trajectories of three vehicles with time stamps.

t=0s

t=5s

t=7s

t=9s

t=10s

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution

249

where yellow lines represent the relative positions of

all three vehicles at a given time. The time stamp t =

0s in Figure 7 indicates the time when the merging

vehicle received the first BSM from the vehicle in

the rightmost lane of the freeway. Similarly, the time

stamp t = 9s indicates the time when the vehicle in

the rightmost lane of the main freeway arrives at the

merging junction, giving the merging vehicle a

merge time cushion of 9 seconds.

5 CONCLUSIONS

In this paper, we have presented a simple approach

to acquire accurate relative trajectories of

surrounding vehicles using standard GPS receivers

and DSRC-based V2V communication. Using this

approach, we have demonstrated that relative

trajectories of the surrounding vehicles can be

achieved with lane-level resolution. We conducted

field tests to successfully acquire the relative

trajectories of vehicles traveling on multiple lanes

toward a merging junction with sufficient accuracy

to distinguish two vehicles on separate or adjacent

lanes of a multiple-lane freeway. However, we

obtained the relative trajectories by post-processing

GPS data acquired through DSRC-based V2V

communication during our field tests. In the future,

we plan to integrate the post-processing algorithm

within the DSRC device of the vehicle to acquire the

relative trajectories in real time.

ACKNOWLEDGEMENTS

The authors wish to acknowledge those who made

this research possible. The study was funded by the

Roadway Safety Institute, the University

Transportation Center for USDOT Region 5, which

includes Minnesota, Illinois, Indiana, Michigan,

Ohio, and Wisconsin. Financial support was

provided by the United States Department of

Transportation’s Office of the Assistant Secretary

for Research and Technology (OST–R).

REFERENCES

Intelligent Transportation Systems - DSRC: The Future of

Safer Driving Fact Sheet. 2017. Intelligent

Transportation Systems - DSRC: The Future of Safer

Driving Fact Sheet. [ONLINE] Available at:

http://www.its.dot.gov/factsheets/dsrc_factsheet.htm.

[Accessed 23 February 2017].

Intelligent Transportation Systems. 2017. Connected

Vehicle Challenges: Potential Impact of Sharing the

5.9 GHZ Wireless Spectrum. [ONLINE] Available at:

Http://www.its.dot.gov/cv_basics/pdf/CV_basics_DS

RC_factsheet.pdf. [Accessed 23 February 2017].

Harding, J., Powell, G., R., Yoon, R., Fikentscher, J.,

Doyle, C., Sade, D., Lukuc, M., Simons, J., & Wang,

J. (2014, August). Vehicle-to-vehicle communications:

Readiness of V2V technology for application. (Report

No. DOT HS 812 014). Washington, DC: National

Highway Traffic Safety Administration.

U.S. Department of Transportation. 2017. Vehicle to

Vehicle Communication Fact Sheet National Highway

Traffic Safety Administration, US Department of

Transportation. [ONLINE] Available at:

Http://www.nhtsa.gov/pdf/v2v_fact_sheet-

02032014.pdf. [Accessed 23 February 2017].

D. Jie and M. J. Barth, 2008. Next-generation automated

vehicle location systems: Positioning at the lane level,

IEEE Trans. Intell. Transp. Syst., vol. 9, no. 1, pp. 48-

57, 2008.

S. Ammoun, F. Nashashibi and C. Laurgeau, 2007, An

analysis of the lane changing manoeuvre on roads: The

contribution of inter-vehicle cooperation via

communication, IEEE Intelligent Vehicles Symposium

pp. 1095-1100, June 2007.

D. Desiraju, T. Chantem. K. Heaslip, 2015. Minimizing

the Disruption of Traffic Flow of Automated Vehicles

During Lane Changes", Intelligent Transportation

Systems, IEEE Transactions, pp. 1249 - 1258 Volume:

16, Issue: 3, June 2015.

A.T. McCartt et al., 2004. Types and characteristics of

ramp-related motor vehicle crashes on urban interstate

roadways in Northern Virginia”, Journal of Safety

Research, vol.35, 2004, pp. 107- 114.

Bruce N. Janson, Wael Awad, Juan Robles, Jake

Kononov, Brian Pinkerton, 1998. Truck Accidents at

Freeway Ramps: Data Analysis and High-Risk Site

Identification, Journal of Transportation and Statistics,

January 1998, pp. 75 – 92.

Basav Sen, John D. Smith, and Wassim G. Najm. 2003.

Analysis of Lane Change Crashes. DOT-VNTSC-

NHTSA-02-03 DOT HS 809 571, March 2003.

G. M. Fitch, S. E. Lee, S. Klauer, J. Hankey, J. Sudweeks,

and T. Dingus, 2009. Analysis of Lane-Change

Crashes and Near-Crashes DOT HS 811 147, June

2009.

National Highway Traffic Safety Administration. 2008.

National Motor Vehicle Crash Causation Survey:

Report to Congress. DOT HS 811 059, July 2008.

D Chun; K, Stol., 2012. Vehicle motion estimation using

low-cost optical flow and sensor fusion, Mechatronics

and Machine Vision in Practice (M2VIP), 2012 19th

International Conference, pp. 507 - 512, Nov. 2012.

Abdelfatah, W.F., et al., 2011. "2D Mobile multi-sensor

navigation system realization using FPGA-based

embedded processors," Canadian Conference on

Electrical and Computer Engineering (CCECE), 2011,

pp. 1218-1221.

VEHITS 2017 - 3rd International Conference on Vehicle Technology and Intelligent Transport Systems

250

Qingquan Li et al., 2014. A Sensor-Fusion Drivable-

Region and Lane-Detection System for Autonomous

Vehicle Navigation in Challenging Road Scenarios,

Vehicular Technology, IEEE Transactions on,

Volume: 63, Issue: 2, pp. 540 – 555, Feb. 2014.

H. Zhao, M. Chiba, R. Shibasaki, X. Shao, J. Cui and H.

Zha, 2009. A laser-scanner-based approach toward

driving safety and traffic data collection, IEEE Trans.

Intell. Transp. Syst., vol. 10, no. 3, pp. 534-546, 2009.

A. Bansal, H. Badino and D. Huber, 2014. Understanding

how camera configuration and environmental

conditions affect appearance-based localization,

Intelligent Vehicles Symposium Proceedings (IV),

2014 IEEE, pp. 800-807.

R. Toledo-Moreo, M. A. Zamora-Izquierdo, B. Ubeda-

Minarro, and A. F. Gomez-Skarmeta, 2007. High-

integrity IMM-EKF-based road vehicle navigation

with low-cost GPS/SBAS/INS, IEEE Transaction on

Intelligent Transportation Systems., vol. 8, no. 3, pp.

491-511, Sep. 2007.

N. Mattern, R. Schubert, and G. Wanielik, 2010. High-

accurate vehicle localization using digital maps and

coherency images, in Proc. IEEE Intell. Vehicles

Symposium, La Jolla, CA, 2010, pp. 462–469.

R. G. García-García, M. A. Sotelo, I. Parra, D. Fernández,

and M. Gavilán, 2007. 3D visual odometry for GPS

navigation assistance, in Proc. IEEE Intell. Vehicles

Symposium, Istanbul, Turkey, 2007, pp. 444–449.

J. Juang; C. Lin., 2015. A Sensor Fusion Scheme for the

Estimation of Vehicular Speed and Heading Angle,

Vehicular Technology, IEEE Transactions on, pp.

2773 - 2782 Volume: 64, Issue: 7, July 2015.

S. Rezaei and R. Sengupta, 2007. “Kalman filter-based

integration of DGPS and vehicle sensors for

localization,” IEEE Trans. Control Syst. Technol., vol.

15, no. 6, pp. 1080–1088, Nov. 2007.

N. Alam, A. T. Balaei and A. G. Dempster, 2013. Relative

positioning enhancement in VANETs: A tight

integration approach", IEEE Trans. Intell. Transp.

Syst., vol. 14, no. 1, pp. 47-55, 2013.

J. Farrell, T. Givargis, 2000. Differential GPS Reference

Station Algorithm-design and Analysis. In IEEE

Transactions on Control Systems Technology, Vol. 8,

No. 3, May 2000, pp. 519-531.

High Accuracy-Nationwide Differential Global

Positioning System Program Fact Sheet - FHWA-RD-

03-039. 2017. High Accuracy-Nationwide Differential

Global Positioning System Program Fact Sheet -

FHWA-RD-03-039. [ONLINE] Available at:

https://www.fhwa.dot.gov/publications/research/operat

ions/03039/. [Accessed 23 February 2017].

William J. Hughes. July 2014. Federal Aviation

Administration, Global Positioning System (GPS)

Standard Positioning Service (SPS) Performance

Analysis Report #86 Technical Centre NSTB/WAAS

T&E Team.

D. K. Schrader, 2013. Combining Multiple, Inexpensive

GPS Receivers to Improve Accuracy and Reliability",

Sensors Applications Symposium (SAS), 2013 IEEE.

pp. 33-37.

R. B. Langley, 1997. GPS receiver system noise, GPS

World, vol. 8, no. 6, pp. 40–45, 1997.

Ahmed El-Rabbany, 2002. Introduction to GPS: The

Global Positioning System. 2002, pp.29.

T. Kos, I. Markezic and J. Pokrajcic, 2010. Effects of

multipath reception on GPS positioning performance,

ELMAR, 52nd International Symposium EL MAR-

2010, pp. 399-402.

D.Jiang, V.Taliwal, A.Meier and W. Holfelder, 2006.

Design of 5.9 GHz DSRC-based Vehicular Safety

Communication, IEEE Wireless Communications,

Vol. 13, No 5, October 2006, pp.36-43.

Acquisition of Relative Trajectories of Surrounding Vehicles using GPS and SRC based V2V Communication with Lane Level Resolution

251