JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.

ORG

Implementing new automation and programmable devices in controlling of traffic flow with use of passive defense approaches
Jalal Nakhaei
Department of Art and Architecture, Science and Research Branch Islamic Azad University; Tehran, Iran

Abstract: The goal of this paper is to state and evalua te the differences in gap accept ance observations between left lane and right lane change, and experiment overall aggressiveness by the means of right lane change behaviors and use of electrical instruments for reaching this goal, furthermore we use Digital Signal Processing on our controlling cameras to be able to distinguish different behaviours of drivers. Also, in this paper we evaluate the decision making process of drivers, we do this work with use of electrical sensors for accumulating some data and clarifying and processing them and finally with use of cumulative distribution functions of driver lane change behaviours from the observed field data. These experiments are performed for drivers using I-20 in Gran d Prairie, Texas with the roadside controlling cameras and some other electronical controlling instruments which were amounted near the intersection of I-20 and Great Southwest Blvd. Our experiments and evaluations demonstrates, that the whole ratio of right lane change observations to left lane change observations was close to 3 to 1.
Keywords: Electronic devices, roadside control camera, lane change, digital image processing, electronic control systems.

INTRODUCTION:

The impact of lane change maneuvers is fundame ntal to m icroscopic traffic flow theor y. I n microscopic model both the sy stem entities and their interactions are described at a high level o f detail. For example, a lane-change maneuver at this level could invoke the car-following law for the subject vehicle with respect to its current leader, then with respect to its putative leader and its putative follower in the target lane, as well as representing other detailed driver de cision processes. The duration of the lane- change maneuver can a lso be cal culated (TRAFFIC SIMULATION BY EDW ARD LIEBERMAN18 AJAY K. R ATHI). There are sever al lane change algorithms in microscopic simulation models aiming at si mulating a se ries of decisi onmaking processes during lane change movement. The agreement between the simulated and field lane change behaviors is an im portant factor fo r simulation models to replicate the re al world conditions. A recent study (Godthelp and Shumann 1994) found errors between speed desired and maintained to vary from -0.3 to -0.8 m/sec in a lane change maneuver; driv ers tended to lose velocity when they made such a maneuver. In an article published by Kouichi Sumida et. al. - Mitsubishi Motors, it is measured the changes of heart rate in order to analyze the drivers' stress while they tried high speed lane-change on the stationary driving simulator. In this paper they concluded that the reasons of the changes were mainly dependent on the following two ite ms. (1) mental pressure caused by the coming task of lane-change, (2) mental disturbance by the task to keep the vehicle inside the lane. By using the mental pressure and disturbance, they could evaluate the stability of the vehicles in lane-ch ange maneuver. The purpose of this paper is to experiment that there is difference in gap acceptance observations between left lane change and right lane change maneuver, and evaluate the overall aggressiveness of right lane change behaviors. In this paper, the driver's decision-making processes are evaluated using cumulative distribution functions of driver lane change behaviors from the observed field

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

data. These evaluations are perfor med for drivers using I-20 in Grand Prairie, Texas roadside cameras mounted near the intersection of I-20 and Great Southwest Blvd.

with the

Due to the difficulty of tra cking many vehicles over time and space, we colle cted lane change maneuver data recorded from a Texas Department of Transportation roadside camera located near the intersection of I-20 and Great Southwest Blvd. The type camera cannot be mentioned due t o the security concerns. This approach eli minates the need for high-resolution maps ac curate enough to capture the exact positions of vehicles in individual lanes. Five video seg ments from two days were analyzed in this project. The time periods analyzed were: Monday, October 20, 2008 o 9:00 am to 9:30 am o 4:00 pm to 4:15 pm o 4:30 pm to 4:50 pm Thursday, October 23, 2008 o 3:00 pm to 4:00 pm o 4:30 pm to 4:50 pm A combination of data from all five time periods were examined together, and then the seg ments were segregated according to Peak an d Off-peak periods. The 9am and 3pm seg ments were considered in the off-peak analysis, and the 4pm and 4:30pm segments were included in the peak period analysis. The tim e periods were segregated in this m anner to determ ine if lane change behavior varies between low volume and high volume traffic conditions. The determination of gap size s were accomplished by recording the elapse d time betw een a vehicle moving into a particular lane (the lead vehicle) passing a fixed point on the freeway and the vehicle directly behind the lead vehi cle (the trailing vehicle) in that lane passing the same point. Tim es were recorded using a h and-held stop watch, and gap values were carried out to tenths of a second. For this project, the west abutment of the Interstate 20 overpass over Gr eat Southwest Blvd. proved to be an ideal fixed poi nt at which to base the measurements. This determination was performed for both right lane and left lane change. The vehicle that made lane change was not tracked after the lane change. Due to the small numbers of lane changes observed by large tractor-trailer trucks and motorcycles during the observation times, the behavior of these vehicles was not considered in this study. Also gap lengths greater than five seconds were not measured to prevent excessive gap lengths from skewing the data.

Data Collection Approach:

Data Collection and Results:

The gap acce ptance observations for this project we re analyzed in two parts: As a whole by combining all observations made over all five time periods co nsidered, and individually by segregating the observations into peak and off-p eak time periods. To help account for poss ible error in the data collecti on procedure stemming from the operation of the stop watch, the observations were grouped into 0.2 second classes for analysis. For each case, the num ber of observations, average gap size, and sta ndard deviation of the ga p size were computed. These same statistics were calculated for the raw (ungrouped) observations for comparison purposes. Summary tables of the grouped data are shown in Tables 1 through 3, and the statistics computed from the ungrouped observation data are presented in Table 4. The number of observed gaps that fell within each class wa s compiled and tabulated, a nd histogram plots of the frequencies of observations in each cl ass were drawn for right and left lane change movements. It is interesting to note that the shapes of the histograms, shown in Figures 1 and 2, resemble that of an exponentia l function that gap accep tance distributions commonly represent. To prove that the observed gaps we re exponentially distributed, a Chi-Squared test was performed. The critical gap para meter was estimated for e ach case by computing the gap length that fell in the middle of the observations. Theoretically, half of the drivers would accept

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

this gap, an d half of t he drivers would reject it. Not surprisingly, the critical gap for left lan e change movements was found to be larger than that for right lane changes. Expected frequencies were calculated by multiplying the total number of observations for each lane change direction by the cumulative probability function, and then compared with the observed frequencies. For both the right and left lane change movements, the error between obser ved and expected frequencies was found to be within t he threshold chi-squared error limits, indicating that the observed gaps were exponentially distributed. Results of the Ch i-Squared tests are presented i n Table 5. Plots of the observed cumulative frequencies are shown in Figures 3 and 4. Finally, hypothesis tests were performed to determine if drivers did, indeed, accept shorter gaps for right lane changes than for left lane changes. For the three conditions cons idered the Offpeak Periods, Peak Periods, and Combined Observations a one-sided hypothesis test was run on the combined mean gap lengths and variances of th e left and right lane change gap observations. The null (h0) hypothesis stated that the gap lengths for right lane changes were, on average, about the same or greater than th ose for the left lane changes, and the al ternative (h1) hypothesis stated that the average gap lengths for right lane changes were less than those for left lane changes. The tests were performed at a 5% level of s ignificance, which is the common significance level used for traffic studies. Since this st udy used va riances that were estimated from observed data a nd not true or universally accepted variances, the hypothesis tests were per formed using the Student s t distribution. A po oled variance was found by taking the weighted average of t he observed variances, and this was used to compute the test statistic, t*. This test statistic was compared with the t-value from the Students t table at a 5% le vel of significance and degrees of freedom equal to the sum of both sample sizes minus 2. If the absolute value of t * was found to be less than that of the tabulated value, the null hypothesis was confirmed. Conversely, if the absolute value of t* was greater than the tabul ated value, the null hypotheses was rej ected. For this study , it was concluded that, overall, the average g aps accepted by drivers making right lane changes are statistically less than those making left lane cha nges. Of the tw o time periods considered, this was also found to be true during the off-peak conditions. For the peak conditions, however, the average gaps of left-lane changes were found to be statistically equal to or greater than those of right hand maneuvers. For the periods where the both the right and left lane change sample sizes were large (>30), which was the case for the Combined and Off-peak Peri ods, hypothesis tests were also conducted as if the variances were known. This is acceptable because, according to the Central Limit Theorem, large sample sizes normally produce estimated variances that are close to the true variances. In this case, the test statistic (Z*) was found using a pooled value of the measured variances, and this was compared with the Z value from the Standard Normal at a 5% level of significance. For both cases in which this m ethod was performed, its c onclusion matched that of the assumed variance method. The null hy pothesis was rejected, and th e average gap lengths of right lane ch ange maneuvers was found t o be statistically less than those of left lane change maneuvers. All hypothesis tests are shown in the Appendix to this report.

Conclusions and Discussion:

The first thing that was no ticed when collecting the data for this paper was the disproporti onate number of right lane change movements observed compared to left lane change movements. The overall ratio of right lane change observations to left lane change observations was close to 3 to 1, and the disparity was more than 4 to 1 during the Peak Period observations. This is likely due, in part, to the l ocation of the camera from which th e video was taken for this project. As stated previously, the camera is located alongside westbound Interstate 20 at its interchange with Great Southwest Parkway, which is less than a mile upstream of the Interstate 20 / State Highway 360 interchange. It is reasonable to think that a l arge number of vehicles would access the right lane in this location in preparation to exit t o SH 360, and the observations reflected this. The disparity

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

may also stem fro m the idea that left lane ch anges are mainly convenience movements made for speed of travel purposes, while right lane ch anges are necessity movements made, in large part, for navigational purposes to a ccess exit ramps. The increase in vehicle volume during peak periods results in a lower availability of com fortable gaps in whi ch to change lanes, and t his condition would likely reduce the number of convenience lane change movements more than it would the necessity lane change movements. The second item that stood out in the fi ndings was that the difference in gap lengths in the two lane change movements was most prevalent in the Off-peak Period, and not so much in the Peak Period. The average gap length observed for right lane changes was shorter than that for left lane changes during all periods considered. Interesti ngly, however, this difference was alm ost one half second in the off-peak tim es, and it reduced to less than a quarter second during the peak times. The overall results indicated that, overall, there is a statistical difference between the gap acceptance lengths for right and left la ne changes, but most of the discrepancy appears to o ccur during periods of lighter traffic. This may be due to the fact that, since there are fewer vehicles on the r oad during the of f-peak, there are a greater number of long gaps f or drivers to ch oose from. What then occurs is the reverse of what happens during the peak times and described in the previous paragraph. The lighter volume allows a greater amount of freedom for vehicles to move within the freeway , which results in a greater number of convenience (i.e. left) lane change movements, and these movements occur with longer gap lengths. Meanwhile, the nece ssity (i.e. right) lane changes continue to occur at relatively high rates and with shorter gap lengths. While the scope of this study was rather li mited by its use of one camera to collect data from a single location, we believe that it does shed som e light on driver behavior regarding lane change maneuvers during varying traffic conditions. It would be interesti ng to expand the scope of this study to ot her locations through out the Metroplex a nd even t o other metropolitan locations to determine if the behavior patterns described herein translate to other locations as well.

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

10

Tables and Figures

Table 1

STATISTICSFROMGROUPEDOBSERVATIONS CombinedPeriods
GapInterval (seconds) 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 < 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 4.80 4.80 Mark Mi 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50 1.70 1.90 2.10 2.30 2.50 2.70 2.90 3.10 3.30 3.50 3.70 3.90 4.10 4.30 4.50 4.70 4.90 Mi2 0.010 0.090 0.250 0.490 0.810 1.210 1.690 2.250 2.890 3.610 4.410 5.290 6.250 7.290 8.410 9.610 10.890 12.250 13.690 15.210 16.810 18.490 20.250 22.090 24.010 : : : LeftLaneChanges Frequency fi 0 0 2 5 7 6 7 4 3 5 1 1 1 1 2 3 2 1 1 1 2 2 4 0 0 61 2.04 1.249 f i Mi 0.000 0.000 1.000 3.500 6.300 6.600 9.100 6.000 5.100 9.500 2.100 2.300 2.500 2.700 5.800 9.300 6.600 3.500 3.700 3.900 8.200 8.600 18.000 0.000 0.000 124.300 fiMi2 0.000 0.000 0.500 2.450 5.670 7.260 11.830 9.000 8.670 18.050 4.410 5.290 6.250 7.290 16.820 28.830 21.780 12.250 13.690 15.210 33.620 36.980 81.000 0.000 0.000 346.850 RightLaneChanges Frequency fi 0 1 7 15 18 27 20 13 9 12 7 3 6 4 6 2 3 2 1 3 2 1 3 1 2 168 1.71 1.035 f i Mi 0.000 0.300 3.500 10.500 16.200 29.700 26.000 19.500 15.300 22.800 14.700 6.900 15.000 10.800 17.400 6.200 9.900 7.000 3.700 11.700 8.200 4.300 13.500 4.700 9.800 287.600 fiMi2 0.000 0.090 1.750 7.350 14.580 32.670 33.800 29.250 26.010 43.320 30.870 15.870 37.500 29.160 50.460 19.220 32.670 24.500 13.690 45.630 33.620 18.490 60.750 22.090 48.020 671.360

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

11

Table 2

STATISTICSFROMRAWOBSERVATIONS CombinedPeriods
LaneChangeDirection: SampleSize: nL: AvarageGapSize: L: Left 61 2.02 sec. sec. nR: R: Right 168 1.71 sec. sec.

StandardDeviation: L: 1.242

R: 1.032

OffPeakPeriod
10/20/200809000930 10/23/200815001600

LaneChangeDirection: SampleSize: nL: AvarageGapSize: L:

Left 43 2.05

sec. sec.

nR: R:

Right 94 1.64 sec. sec.

StandardDeviation: L: 1.329

R: 1.063

PeakPeriod
10/20/200816001615 10/20/200816301650 10/23/200816301650

LaneChangeDirection: SampleSize: nL: AvarageGapSize: L:

Left 18 1.94

sec. sec.

nR: R:

Right 74 1.79 sec. sec.

StandardDeviation: L: 1.033

R: 0.993

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

12

Table 3 - Chi Squared Test Results

LeftLaneChangeManeuvers
GapLength Ti (sec)
0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 4.40 4.60 4.80 5.00

RightLaneChangeManeuvers
ErrorIndex (EiOi) /Ei
7.274 13.681 15.531 12.311 7.507 4.824 2.216 1.609 1.371 0.540 0.761 0.944 1.079 1.167 0.922 0.472 0.290 0.263 0.225 0.181 0.055 0.001 0.187 0.144 0.111
2

Observed Frequencies Oi
0 0 2 7 14 20 27 31 34 39 40 41 42 43 45 48 50 51 52 53 55 57 61 61 61

Expected Probability p(gTi)=1e

0.119 0.224 0.317 0.398 0.470 0.533 0.589 0.638 0.681 0.719 0.753 0.782 0.808 0.831 0.851 0.869 0.885 0.898 0.910 0.921 0.931 0.939 0.946 0.953 0.958
Ti/

Expected Frequencies Ei=p(gTi)N

7 14 19 24 29 33 36 39 42 44 46 48 49 51 52 53 54 55 56 56 57 57 58 58 58 (hat) (table)
2 2

Observed Frequencies Oi
0 1 8 23 41 68 88 101 110 122 129 132 138 142 148 150 153 155 156 159 161 162 165 166 168

Expected Probability p(gTi)=1e

0.137 0.255 0.357 0.445 0.521 0.586 0.643 0.692 0.734 0.770 0.802 0.829 0.852 0.872 0.890 0.905 0.918 0.929 0.939 0.947 0.954 0.961 0.966 0.971 0.975
Ti/

Expected Frequencies Ei=p(gTi)N

23 43 60 75 87 98 108 116 123 129 135 139 143 147 149 152 154 156 158 159 160 161 162 163 164 (hat) (table)
2 2

ErrorIndex (EiOi) /Ei

22.975 40.832 44.997 35.789 24.685 9.434 3.700 1.987 1.430 0.423 0.239 0.376 0.186 0.142 0.015 0.027 0.009 0.008 0.019 0.000 0.003 0.002 0.045 0.053 0.110 187.486 197.064
2

73.668 77.931

Critical Gap = = n= l= = =

1.58 61 1 59 5%

Since 2(hat) is less than the Critical Gap = = 1.36 threashold error 2(table), n = 168 exponential distribution accurately l= 1 models the observed gap = 166 = 5% acceptances for left-handed lane change maneuvers.

Since 2(hat) is less than the threashold error 2(table), exponential distribution accurately models the observed gap acceptances for right-handed lane change maneuvers.

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

13

Figure 1- Typical View from TxDOT Video Camera

Figure 2

Figure 3

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

14

Figure 4

Figure 5

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617

JOURNAL OF COMPUTING, VOLUME 4, ISSUE 10, OCTOBER 2012, ISSN (Online) 2151-9617 https://sites.google.com/site/journalofcomputing WWW.JOURNALOFCOMPUTING.ORG

15

References:
[1]

[2]

[3]

[4] [5] [6]

[7]

[8]

[9]

[10]

[11]

Mehdi Darbandi Appl ying Kalman Filtering in so lving SSM estimation probl em by the means of EM algorithm with considering a practical example; published by the Jo urnal of Com puting Springer, 2012; USA. Mehdi Darbandi; Com parison between miscellaneous platforms that presen t for cloud c omputing and accreting the security of these platforms by new filter; published by the Journal of Computing Springer, 2012; USA. Mehdi Darbandi; New an d novel technique in designing electromagnetic filter for elim inating EMI radiations and optimization performances; published by the Journal of Computing - Springer, 2012; USA. Mehdi Darbandi; Appraising the role of cloud co mputing in daily life and presenting new solutio ns for stabilization of this technology; published by the Journal of Computing - Springer, 2012; USA. Mehdi Darbandi; Cloud Computing make a revolution in econ omy and Information Technolo gy; published by the Journal of Computing - Springer, 2012; USA. Mehdi Darbandi; Consi dering the high im pact of gettering of silicon on fabrication of wafer designing and optimize the designing with new inn ovative solutions; published by the Jo urnal of Computing Springer, 2012; USA. Mehdi Darbandi; Devel oping concept of elect romagnetic filter design by considering new parameters and use of mathematical analysis; published by the Journal of Com puting - Springer, 2012; USA. Mehdi Darbandi; Is the cloud computing real or hype Affirmation momentous traits of this technology by proffering maiden scenarios; published b y the Journal of Com puting Springer, 2012; USA. Mehdi Darbandi; Measurem ent and collation o verriding traits of co mputer networks an d ascertainment consequential exclusivities of cloud co mputing by the means of Bucy filtering; published by the Journal of Computing - Springer, 2012; USA. Mehdi Darbandi; Unabri dged collation about multifarious computing m ethods and outreaching cloud computing based on innovative procedure; published by the Journal of Computing - Springer, 2012; USA. Mehdi Darbandi; Scrutiny about all s ecurity standards in cloud com puting and present ne w novel standard for security of such networks; published by the Journal of Co mputing - Springer, 2012; USA.

Biography:
Jalal Nakhaei:
Science and Research B ranch Islamic Azad University; Teh ran, Iran. His research areas are In formation Technology (IT), Secu re Spaces, and Structures.
is currently PhD. Student in Art and Architecture Engineering in

2012 Journal of Computing Press, NY, USA, ISSN 2151-9617