Вы находитесь на странице: 1из 4

IOSR Journal of Computer Engineering (IOSR-JCE)

e-ISSN: 2278-0661, p- ISSN: 2278-8727Volume 16, Issue 2, Ver. V (Mar-Apr. 2014), PP 127-130
www.iosrjournals.org
www.iosrjournals.org 127 | Page

On Road Safety and Security

Abhishek Malik
#
, Jatin Sharma
#
, Animesh Verma
#
, Sachin Kumar
#
,
Anuj Kumar
*

#(
Department of Electronics & Comm. Engg., Meerut Institute of Technology, Meerut, Uttar Pradesh, India)
*
(Department of Electronics & Comm. Engg., M.Tech ,Indian Institute of Technology, Roorkee, Uttarakhand,
India)

Abstract :This paper demonstrates a novel car-following model focused on passenger comfort , In modern
society due to high traffic accidents , high crime rates the feeling of insecurity and threat is increasing, records
in the database show that the number of new cars being bought and so is the cars being stolen locally and
internationally has increased rapidly . this model focuses on two key points first is car safety .the methods
considered in this model are different from the conventional method used till date. Car safety model focuses on
anti directional schemes ie. the first scheme evidences on the cars rear view features , second considers the
cars backlight features and third on using mms and database technology. Driver comfort and safety model
focuses on acceleration of the controlled car by monitoring space headway and velocity of the controlled car.
Fiction coefficient is a key factor in this model.
Keywords: Brake comfort, car control model, car following, driver behavior, performance, Efficiency

I. INTRODUCTION
By using Auto braking system technology we are achieving many advantages. Car safety and comfort
has always been a major issue in public. There are a lot of car security and comfort system that had been
produced lately but the outcome is still disappointing. To built a realistic , efficient and a user friendly technique
we need to widen up a database by involving real life traffic. One of the most usual approaches for vehicle
detection is identifying the vehicle in our visual range or may be from images or videos. But with the increasing
car manufactures there has been a increase in wide range of vehicle colours, sizes poses, orientation, shape and
designs. While the conventional related to driver safety using speed sensors were not as effective or efficient in
controlling accidents and on road traffic jams so the ORSS model is robust and quite effective in dealing with all
these problem. Progress has been made in all the schemes [1].

A. Model for comfortable breaking based on space and velocity
At time T preceding car applying emergency brake , at time T+ following car driver realiser to apply brake
following car brake according is a. when the following car stop at T+ , The displacement of following and
preceding car is D and s .


Fig. 1. Car-following model based on space and velocity

With the help of diagram

s+s = D + L. (1)
For safe and comfortable braking model car must stop before any collision occurs this particular moment, the
space headway L is zero, and the displacement of the following car can be calculated by

2aD = v
2
. (2)
From (1), (2), and L = 0, we derive
2a(s+s) = v
2
. (3)
On Road Safety And Security
www.iosrjournals.org 128 | Page

[1]Preceding car displacement s depend on magnitude of brake acceleration and friction coefficients and driver
intension.
Minimum stopping distance
smin = v
2
L
2g (4)

Here, vL is the preceding cars velocity, is the friction coefficient of the road surface, and g is the acceleration
of gravity.
This can be described by the formula

2a s + v
2
= v
2
. (5)


2g

According to (5), if s, v, and are known values, then the acceleration a can easily be calculated. If there is an
acceleration value known to make people feel uncomfortable, we can then avoid such discomfort by limiting a
cars acceleration


Fig. 2. Relationship between s and v at different values of a.

below that value. Fortunately, researchers have discovered the relationship between subjective comfort and
acceleration.
From (6), we know that there will be a different acceleration a for each unique velocity v and different distance s
for each value of . To calculate the acceleration, we modify (5) into

a = v
2

2 s +v
2
. (6)


2g

Here, v represents the following cars velocity, s represents the distance between the two cars, is the
coefficient of friction, and a is the deceleration of the following car [2] .
For every unique value of , we can deduce the relationship between s and v by using a diagram. For
example, Fig. 2 illustrates the specific curves formed when we graph the relationship between the brake distance
s and velocity v at various accelerations a at the fixed of 0.8. To avoid a collision, the distance between two
cars should always be larger than zero [3] .

In Fig. 2, amax is the maximum deceleration of the car, and it can be calculated according to

amax = xp g (7)
ac =2 m/s
2
(8)
smin =v
2
/2g (9)
s =v
2
/2ac (10)
acmax =2 m/s
2
(11)
vmin =20 km/h = 5.56 m/s (12)
vmax =173 km/h = 48.06 m/s (13)
On Road Safety And Security
www.iosrjournals.org 129 | Page
0.2 < < 0.9 (15)



Fig. 3. Relationship between s, v, , and a.

B. Four key Modules for car Detection and Recognition Systems Are :-

Fig. 4. Overview of the proposed car recognition system.

The proposed car detection and recognition system is composed by four main modules:
1. CAR SEGMENTATI ON- unknown cars are divided into section from background. A angled rear view image of
car need to be extracted.
2. CAR FEATURE EXTRACTI ON- we know most of the cars are similar at the lower 1/4
th
section, while the
difference in upper 3/4
th
, characterizing the car in the upper 3/4
th
section. As we know the back light of all
the cars are usually having red colour as their main and more relevant colour. So to find the regions present
in the segmented cars image. Then coming on to the licence plate and outer feature etc. (colour as can not
be change at a short instance).
3. CAR DATA BASE PROCESSI NG- the data base of the car is constantly updated with adding up of new cars and
their rare view features etc. so as to properly mat with the unknown cars.
4. CAR RECOGNI TI ON- a comparison is made between a testing car with these in the date base combination of
shape similarity, lights etc.
Critical comfortable acceleration A
c
is 2m/s
2
[2].
The table given above represent different zones which our model signifies [4].

II. PRACTICAL APPLICATION
[4] At IST parking, 18 cars entered. A video footage was taken of all the cars with a database of about 40 cars
which had been from random storage during certain time period. Video recording was done in different days but
with similar illumination conditions. No special care was taken when acquiring the tests and database sets, in
terms of illumination, reflections or foreign shadows reduction.

On Road Safety And Security
www.iosrjournals.org 130 | Page
III. Results
44% Success Rate Due to Scheme(1) i.e., rear car features.
72% Success Rate Due to Scheme(2) i.e., back light recognition.
While a combination provides 89% success rate i.e., 16 out of 18 cars.
Car colour recognition resulted in 100% success rate.

IV. Future Prospects
Finally, the novel methods proposed in this paper can be explored as easy to use techniques for
automatic car recognition contributing to the work in progress vehicle recognition research area.

V. Conclusions
One contribution of this paper is the fact that we were able to find a brake comfortable car-following
model for longitudinal acceleration, whereas most former car-following models only consider vertical vibration.
Although scheme(1) is 44% successful and scheme(2) is about 72% successful but the combination gives a
success rate of 89%. It is robust in mature . As we know the conventional methods of detection are the licence
plate recognition but the drawback is that it could be easily replaced while on the other hand, the external
features are hard to change in a short while and they can even be monitored with a low video quality. The result
have been acquired with 1
st
scheme in a time duration of 1 minutes 20 sec/car while with 2
nd
scheme it is just 10
sec. The computer used in Intel core 2 6400@2.134GHZ processor and 2giga byte RAM. This technique is
ready for use with work in progress and scope for improvement.
Another contribution is that we were able to determine a comfortable brake acceleration for our model
in this paper. By comparing the discomfort caused by vertical vibration with the discomfort caused by
longitudinal acceleration, a comfortable brake acceleration is determined according to standard international
criteria. Although its value varies with each different situation, it can be adapted to any specific environment in
the real world [5] .

REFERENCES
[1] Y. Umemura, Driver behavior and active safety (overview), R D Rev. Toyota CRDLSpecial Issue Driver Behavior and Active
Safety, vol. 39, no. 2, pp. 18, 2004.
[2] R. W. Rothery, Car following models, in Traffic Flow Theory, N. Gartner, C. J. Messer, and A. K. Rathi, Eds., Transportation
Research Board (TRB), ch. 4, Special Rep. 165, 1999.
[3] M. F. Aycin and R. F. Benekohal, Comparison of car-following models for simulation, Transp. Res. Rec., no. 1678, pp. 116127,
1999.
[4] R. N. Janeway, Vehicle vibration limits to fit the passenger, SAE J., vol. 56, pp. 4849, Aug. 1948.
[5] V. Cossalter, A. Doria, S. Garbin, and R. Lot, Frequency-domain method for evaluating the ride comfort of a motorcycle, Vehicle
Syst. Dyn.: Int. J. Vehicle Mech. Mobility, vol. 44, no. 4, pp. 339355, Apr. 2006.
[6] G. J. Shen, Z. Q. Liu, and L. J. Meng, Analysis of the longitudinal dynamics test on No. T41/42 train from West Beijing Station to
Xian Station, Roll. Stock, vol. 40, no. 11, pp. 89, Nov. 2000.
[7] U.S. Department of Transportation, Compendium of Executive Summaries from the Maglev System Concept Definition Final
Reports, pp. 4981, Mar. 1993. DOT/FRA/NMI-93/02.
[8] M. Y. Bian, K. Q. Li, D. F. Jin, and X. M. Lian, A velocity control strategy for vehicular collision avoidance system, in Proc.
IEEE Int. Conf. Mechatronics Autom., Jul. 29Aug. 1 2005, vol. 4, pp. 18271830.
[9] C. G. Wallman and H. Astrom, Friction Measurement Methods and the Correlation Between Road Friction and Traffic Safety. A
Literature Review. Linkping, Sweden: Swedish Nat. Road Transp. Res. Inst., 2001, pp. 2223.
[10] Y. F. Liu, Researches on driver behavior modeling, Ph.D. dissertation, Zhejiang Univ., Hangzhou, China, Sep. 2007.
[11] S. Shi, Z. Qin, and J. Xu, Robust Algorithm of Vehicle Classification, International Conference on Software Engineering,
Artificial Intelligence, Networking, and Parallel/Distributed Computing, 2007, pp. 269-272.
[12] Y. Du and N. Papanikolopoulos, Real-Time Vehicle Following Through a Novel Symmetry-Based Approach, IEEE International
Conference on Robotic & Automation, 1997, pp. 3160-3165, vol.4.
[13] A. Santhanam and M. M. Rahman, Moving Vehicle Classification Using Eigenspace, Proceedings of the 2006 IEEE International
Conference on Intelligent Robots and Systems, 9-15 Oct. 2006, pp. 3849-3854.
[14] X. Ma, and W. Grimson, Edge-based rich representation for vehicle classification, Proceedings of the Tenth IEEE International
Conference on Computer Vision (ICCV05), 2005.
[15] David Anthony Torres, More Local Structure Information for Make- Model Recognition, Dept. of Computer Science, University
of California at San Diego, taken from http://www-cse.ucsd.edu.
[16] Milos Stojmenovic, Real Time Machine Learning Based Car Detection in Images with Fast Training, Springer-Verlag, March
2006.
[17] N. Otsu "A Threshold Selection Method from Graylevel Histograms, IEEE Transactions on Systems, Man, and Cybernetics, vol. 9,
no. 1, Jan. 1979, pp. 62--66.
[18] D. Santos, J. Fazenda, P. Correia, Using Gait to Recognize People, Proc EUROCON The International Conference on
Computer as a Tool , Belgrado, Servia & Montenegro , Nov. 2005.
[19] C. Paulo, P. Correia, Automatic Detection and Classification of Traffic Signs, WIAMIS2007, Santorini, Greece, June 2007.
[20] P. J. Phillips, H. Moon, S. A. Rizvi, and P. J. Rauss, The FERET Evaluation Methodology for Face-Recognition Algorithms,
IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 10, October 2000, pp. 1090-1104.

Вам также может понравиться