MONTE CARLO SIMULATION

MODEL OF
INTER-VEHICULAR
COMMUNICATIONS



By,
Team 4
Abstract:
Inter-vehicle communication (IVC) is emerging in research
prominence for the interest that is generating in all major car manufacturers and for
the benefits that its inception will produce. The specific features of IVC will allow
the deployment of a wide set of possible applications which span from road safety
to entertainment and traffic management. We present an analytical model for
multi-hop connectivity of IVC in a traffic stream, in which positions of vehicles
are simulated. We study multi-hop connectivity between two equipped vehicles
subject to arbitrary distribution patterns of vehicles, market penetration rates, and
transmission ranges. After introducing Most-Forwarded-within-Range (MFR)
communication chains and node-related events, we derive a recursive model of
node and further define a number of performance measures of multi-hop
connectivity. We then apply the model to study multi-hop connectivity of IVC in
both uniform and non-uniform traffic and obtain results consistent with those in
literature.

Why the simulation?
We develop a Monte Carlo simulation model for analyzing
instantaneous inter-vehicle communication in a traffic stream. We investigate the
properties of IVC and demonstrate the success rate for information to propagate
beyond a location or the probability for two equipped vehicles to be connected can
be naturally estimated through repeating Monte Carlo simulations.

Introduction:
In recent years, inter-vehicle communication (IVC) based on wireless
communication technologies as Wi-Fi has become an intriguing option in
developing intelligent transportation systems. With a subpopulation of vehicles
equipped with wireless communication and information processing units, real-time
critical traffic information can be propagated both to relevant travelers to help
make proper choices over routes and departure times and also to traffic
management centers to help operate the traffic system more efficiently.
With equipped vehicles as communication nodes, an IVC system can be
considered as a special mobile ad hoc network When the market penetration rate is
sufficiently high, relaying information in an IVC system is through multi-hop
broadcasting. In this manner, a piece of information can be transmitted to vehicles
outside of communication range in a relatively short time. It is reasonable to
assume such multi-hop broadcasting to be instantaneous.
Further, we assume that information propagates in the fashion of most
forwarded within range (MFR). That is, a piece of information is transmitted to
the farthest equipped vehicle within the transmission range of a sender. It is the
fastest approach to sending a message.






.
Framework and Definitions:
It is assumed that k vehicles are stationed at different positions(no
overlapping) in a single straight line and all travelling in the downstream
direction.
Positions of vehicles are determined by traffic models, and the distribution
pattern of vehicles can be arbitrary (either uniform, in which the distances
between two consecutive vehicles remain the same). Although the positions
of vehicles are deterministic in IIVC, the probability for any individual
vehicle to be equipped remains the same as market penetration rate , i.e.,
the percentage of equipped vehicles among all.

Assuming whether or not any two particular vehicles are equipped is
independent, the distribution of equipped vehicles in a traffic stream can be
described by Bernoulli trials .Then for traffic stream of K vehicles, there
can be 2K different Bernoulli trials. A realization of Bernoulli trials
corresponding to a traffic stream can be represented by a sequence of 0s and
1s, where 0s stand for non-equipped vehicles and 1s for equipped
vehicles. The probability of the existence of a given Bernoulli trial can be
computed from market penetration rate.

To determine how far information can propagate in a realization of Bernoulli
trials, we can construct a so-called most forwarded within range" (MFR)
communication chain of nodes as follows. First, we denote the information
source by node 0. Then, within the transmission range R, i.e., the
maximum transmission distance at one hop, of node h (h 0), node h +1 is
the equipped vehicle farthest in the direction of information propagation.
We denote by (k;h) the event that vehicle k is node h in a communication
chain. If node h is the end node, from which information cannot be further
relayed, we denote the event by (k;h*). Then, the probability for vehicle k to
be node h of all possible communication chains is denoted by P(k;h) , and
the probability for vehicle k to be end node h by P(k;h*) . Here P(k;h) is
called node probability, and P(k;h*) end node probability. Therefore, we
only consider connectivity between information source and an equipped
vehicle. If we define P(k) as the probability for a communication chain to
stop at vehicle k ,regardless of the number of hops, then
P(k*) =P(k;h*)

The probability for information to travel to and beyond vehicle k a traffic
stream of K vehicles (excluding the information source) is given by
s(k) =P(i*), where i varies from k to K

If vehicle k is equipped, then it is connected to the information source as long as
a communication chain does not end before k, the farthest upstream vehicle of
k. Therefore, the connectivity between the information source and equipped
vehicle k equals the probability for information to travel beyond vehicle k,
(k) =s(k*)





Monte Carlo Model:
For a traffic stream with K vehicles, the flow-chart of Monte Carlo
simulations of Instantaneous inter-vehicle communication, in which each step is
carried out as follows.
Inputs for the experiments include positions of K vehicles x(k),( k =1,.,K ),
penetration rate communication range R , and the number of experiments M .
Note that the position of information source is x(0) = 0 .
In each experiment, K uniformly distributed random variables k(X), are generated
in [0,1].
If k(X) , vehicle k is equipped, and we can obtain a realization of Bernoulli
trials. Here the event when vehicle k is equipped is denoted by (k;1) .
Starting from the source, information is transmitted to the farthest equipped
vehicle within a communication range, i.e., in the manner of MFR. In this way, we
can obtain a unique MFR communication chain for each experiment.
If information travels to vehicle k with h hops ( h > 0 ), we denote this event by
(k;h) . Further, if node h is the end node of the MFR communication chain, we
denote the event by (k;h*) .
The numbers of occurrences of (k;1) , (k;h) , and (k;h*) are denoted by N(k;1)
N(k;h) , and N(k;h) , respectively. In each experiment, these quantities are added
by 1 if the corresponding events occur.
After finishing M experiments, we can compute the simulated probabilities of
P(k;1) P(k;h) , and P(k;h). Then we can use these probabilities to find P(k*), s(k)
and (k).



Simulation using Java:
The program is attached in another Word document. The program facilitates
generation of both uniform and non-uniform traffic stream. Certain parameters
such as range, penetration ratio, vehicle density are assumed and can changed
accordingly. The program requires the user to enter the car number k. P(k*), s(k)
and (k) are computed and displayed. The values were then transferred to
MATLAB to obtain suitable graphs.


Output of simulation:
Few screenshots of the output are attached below:
1. Uniform traffic stream:













2. Non-uniform traffic stream:













Analysis with graphs:
Figure 1: Vehicle density=58veh/km, penetration ratio=0.1,
Range (as mentioned in the legend)
Figure 2: Vehicle density=19veh/km, penetration ratio=0.1,
Range (as mentioned in the legend)
Figure 3: Vehicle density=19veh/km, penetration ratio=0.3,
Range (as mentioned in the legend)

The connectivity between an equipped vehicle and the information source is
shown for a uniform traffic stream with different parameters, i.e. different range,
vehicle density and penetration ratio.
One thing common to all graphs is that the connectivity increases
significantly as the Range R increases.
Comparing figures 1 and 2, for the same penetration ratio=0.1, we see that
the connectivity worsens for lesser traffic density, i.e. connectivity is poorer
in the second case when vehicle density is lesser (19 veh/km).
Comparing figure 2 and 3, for same vehicle density=19 veh/km, we see that
connectivity increases significantly with higher penetration ratios.The graph
of 1000m range in figure 3 with =0.3 has very good connectivity even at
long distance, approx. 10km. We can conclude that connectivity will be poor
initially when IVC is implemented due to low market penetration ratio.


Non-uniform Traffic Case:

Figure 4: Vehicle density=58veh/km, penetration ratio=0.1,
Range (as mentioned in the legend)

Figure 5: Vehicle density=58veh/km, penetration ratio=0.3,
Range (as mentioned in the legend)
The above two graphs are drawn between connectivity and distance for non-
uniform traffic stream. As obvious, since the vehicles are randomly spaced,
connectivity is rugged and not as smooth as the uniform case.
Comparing figures 4 and 5, we can conclude that penetration ratio plays a
huge role in the connectivity.
In figure 5, for range=1000m, it is observed that connectivity is 1 for all cars
at all distances. This is the ideal expected case.


Conclusion:
We developed a model for computing node probabilities, end node
probabilities, from which various performance measures of multi-hop connectivity
were defined. The model assumes instantaneous information propagation, a
condition that applies to situations for which market penetration is sufficiently high
so as to minimize the impact of vehicle movement. We then studied examples of
multi-hop connectivity in both uniform and non-uniform traffic and showed that
the distribution patterns of vehicles on a road can significantly affect the
performance of IVC.
The results of this study serve to better understand how well
information may propagate through IVC under certain conditions, and to help
determine properties of communication devices (in terms of communication range)
that potentially could be used to achieve certain performance of such systems.




Team Members:
Prajuwal Naidu 1PI12EC071
Sagar.A.S 1PI12EC089
Sharan.S.Patil 1PI12EC097
Shashank.D.P 1PI12EC098
Shravan Hanji- 1PI12EC102
Y.S.Sujay 1PI2EC129