Embodiment
In conjunction with
accompanying drawingfurther describe the specific embodiment of the present invention.
Embodiment one: combine
fig. 1,
fig. 2understand present embodiment, the chain-circuit time delay dynamic prediction method considering V2V in the VANETs of intersection towards highway described in present embodiment, set up chain-circuit time delay forecast model (Extended Link Duration Prediction, the ELDP) model of expansion;
In order to how long accurately predicting two automobiles can link, chain-circuit time delay model must overcome following three challenges: (1) it must solve automobile in the impact of intersection turning on chain-circuit time delay, (2) it must adjust and adapt to the change of the speed of a motor vehicle, (3) it must transport solution lamp on the impact of chain-circuit time delay.In order to solve first challenge, the present invention considers automotive steering angle in ELDP model.In order to solve second challenge, the present invention proposes the distribution of the relative velocity of use two car instead of instantaneous velocity and calculates and expect the chain-circuit time delay of the garage that will occur.In order to solve the 3rd problem, the present invention's probability be automobile in crossing modeling, then calculate the chain-circuit time delay because traffic lights causes.
In order to analyze automobile and traffic lights to the impact of two workshop chain-circuit time delays, the present invention considers turning to of crossing in automobile city and highway environment respectively.
fig. 1the all steering modes giving automobile on the way crossing comprise turning left and turn right.
The present invention exists
fig. 2in give in detail two cars crossing turn to comprise turn left and turn right.Arrow points out the direction of running car, and r is the radius of wireless transmission, d
irepresentative be when each second predict chain-circuit time delay time two cars between initial distance, d
i<r, i=1,2,3 ..., n, ELDP predicts that the lower prediction of hypothesis is by the chain-circuit time delay of crossing below: i) automobile travels until link disconnection along the tangential direction of automobile real trace, ii) until when link disconnects, the speed of two cars keeps invariable.When a link is broken, the distance between two cars is radio communication radius r.Above-mentioned hypothesis is also applied to predicting the chain-circuit time delay prediction of keeping straight on, and that is, the angle between tangent line and real car track is zero degree.
The relative distance of two cars when step one, calculating link disconnect; Step 2, estimate two cars relative velocity distribution; Step 3, predict the chain-circuit time delay of two cars based on step one and step 2, specific as follows: step 3 one, to calculate correlative factor needed for two car chain-circuit time delays, comprise the relative velocity v of two cars, the initial distance d between two cars, the travel direction of two cars; Step 3 two, on the basis of step 3 one, calculate the chain-circuit time delay of two cars.
Embodiment two: combine
fig. 3~
fig. 5understand present embodiment, present embodiment and embodiment one unlike: when the calculating link described in step one disconnects, the detailed process of the relative distance of two cars is: step sets up a plane right-angle coordinate one by one, using the position of an automobile as true origin
as Fig. 3in shown in (a), assumed vehicle can move forward at any angle, and this angle is the angle between speed of a motor vehicle direction and transverse axis.Angle change, from 0 ° of degree to 360 °, that is can be any direction in a coordinate system.Speed of a motor vehicle direction is the true directions that automobile on the way travels, and the value of speed is
in order to analyze automobile chain-circuit time delay, illustrate automobile n1 and automobile n2 moves forward along α and β respectively, α and β is the angle between moving vehicle direction and x-axis,
as Fig. 3in shown in (b), α, β ∈ [0,2 π], calculates according to geometric vector Vector triangle
fig. 3in the relative velocity vector of two cars in (b)
define new coordinate system with automobile n1 position O' for true origin, automobile n1 velocity vector is
automobile n2 velocity vector
move in parallel
obtaining relative velocity vector is:
vector
with
coordinate be:
Wherein v
1for the speed of automobile n1, v
2for the speed of automobile n2;
coordinate equation (2) represents,
Order:
Equation (2) equation (4) represents again:
direction at first quartile, the second quadrant, third quadrant or fourth quadrant,
value equals
as Fig. 4shown in.
Two cars move forward along different directions and respective speed respectively,
as Fig. 3in shown in (b) coordinate system, according to relative motion principle, assumed vehicle n1 is motionless, and automobile n2 then moves forward with the relative velocity of two cars, sets up new coordinate system using the position of automobile n1 as true origin, and new coordinate system is set up in former coordinate system translation
as Fig. 5shown in, the true origin of new coordinate system is O'.
fig. 5in, automobile travels in new coordinate system, and automobile n1 and the coordinate of automobile n2 in former coordinate system are (x
1, y
1) and (x
2, y
2), so automobile n2 coordinate in new coordinate system is D (x
2-x
1, y
2-y
1), all following derivations are all carried out in new coordinate system.The relative velocity of automobile n1 and automobile n2 is
another vector
perpendicular to vector
intersection point is E, v
1and v
2initial distance be d, d
vthat initial point O' is to vector
vertical range, automobile n2 is with relative velocity
the position D disconnected from a D to link
0travel, if distance O'D
0just in time equal radio communication radius r, the link between automobile n1 and automobile n2 will disconnect, and so automobile n2 drives to a D from a D
0time be link trip time, the chain-circuit time delay of therefore prediction in each second equals automobile n2 and drives to D from a D
0time used.
The relative distance DD of two cars when step one two, link disconnect
0; Provide two straight line l
1and l
2standard equation, according to the principle of straight line standard equation, adopt a some D (x
2-x
1, y
2-y
1) and straight line l
1direction vector
set up straight line l
1standard equation as follows,
straight line l
1perpendicular to l
2, therefore, the direction vector dot-product of two straight lines for equaling 0,
so straight line l
2direction vector
equation (7) can be represented as,
utilize true origin O'(0,0) and straight line l
2direction vector
set up straight line standard equation,
Simultaneous equations of the present invention (5) and (8),
Therefore put E coordinate can obtain:
point is to the vertical range d of straight line
vcan be calculated as follows by the computing formula of space mid point to the distance of straight line: vector
be expressed as equation (10)
Vector
be expressed as equation (11),
So distance ED
0be respectively calculated as follows according to triangle Pythagorean theorem with ED,
Embodiment three: combine
fig. 6understand present embodiment, present embodiment and embodiment one or two are unlike the relative distance DD of two cars when the link described in step one two disconnects
0be divided into four quadrants specifically to calculate, wherein, the computation process of first quartile is: if B>0 and A>0,
at first quartile, be divided into the following two kinds situation: (1)
fig. 6in (a) large than some E of horizontal ordinate of giving a D, that is,
So DD
0=ED
0-ED (16)
(14) and (15) are substituted into (16) formula, then (16) formula is derived as follows,
(2)
fig. 6in (b) some D of providing very coordinate is less than the horizontal ordinate of some E, that is,
so the present invention can obtain DD
0value, DD
0=ED
0+ ED (18)
Embodiment four: combine
fig. 7understand present embodiment, one of present embodiment and embodiment one to three are unlike the relative distance DD of two cars when link described in step one two disconnects
0in the computation process of the second quadrant be: if B<0 and A>0,
be at the second quadrant, be divided into the following two kinds situation:
(1)
fig. 7in (a) horizontal ordinate of giving a D less than some E, that is
so the present invention can calculate DD by equation (20)
0value,
DD
0=ED
0-ED (20)
Equation (20) is derived as follows again,
(2)
fig. 7in (b) horizontal ordinate of providing a D than horizontal ordinate large of some E, that is
So
DD
0=ED
0+ ED (22), equation (22) is expressed as following formula again,
Other step is identical with one of embodiment one to three.
Embodiment five: combine
fig. 8understand present embodiment, one of present embodiment and embodiment one to four are unlike the relative distance DD of two cars when link described in step one two disconnects
0in the computation process of third quadrant be: if B<0 and A<0,
at third quadrant, be divided into the following two kinds situation: (1)
fig. 8in (a) point out that the horizontal ordinate pen point E's of a D is little, that is
So DD
0value is derived as:
(2)
fig. 8in (b) point out that the horizontal ordinate of the ratio E of a D is large,
so
Embodiment six: combine
fig. 9understand present embodiment, one of present embodiment and embodiment one to five are unlike the relative distance DD of two cars when link described in step one two disconnects
0in the computation process of fourth quadrant be: if B>0 and A<0,
in fourth quadrant, be divided into the following two kinds situation:
(1)
fig. 9in (a) show that the horizontal ordinate of a D is larger than E, that is,
So
DD
0=ED
0-ED (26)
(2)
fig. 9in (b) show that a D horizontal ordinate is less than E, that is,
So
DD
0=ED
0+ED (28)
To sum up, if
in first quartile and fourth quadrant, so solve DD
0the method of value be identical, if
at second and third quadrant, the present invention adopts said method to calculate DD
0value.
Embodiment seven: one of present embodiment and embodiment one to six unlike the detailed process of the relative velocity distribution of: estimation two car described in step 2 are: in document [37], [38], [39], in, research shows speed of a motor vehicle Normal Distribution.Assuming that the speed v of an automobile is a stochastic variable and Normal Distribution, v ~ N (μ, σ
2), its probability density function (PDF) is:
In fact, when running car is on road time, speed is a vector, that is,
therefore, speed not only has value but also has direction, and velocity vector value is
the direction of velocity vector is along α angle, and this angle is the angle between vehicle traveling direction and transverse axis.Because refer to speed of a motor vehicle Normal Distribution above, be the value Normal Distribution of velocity vector in fact, namely
normal Distribution, because
therefore, equation (30) is derived as (31),
About the speed of any two cars,
with
so the relative velocity of two cars is
So
normal Distribution, the probability density function (PDF) of two car relative velocities represents with equation (32),
Estimate the distribution of the relative velocity between any two cars, need the sample gathering relative velocity, such as, the sample set of the vector of a relative velocity
so the set of its vector value is
they can from N (μ, the σ of a normal distribution
2) obtain in member, utilize maximum-likelihood method to estimate μ and σ
2, maximization likelihood function lnL (μ, the σ of this method
2), can represent with following formula
ask about μ and σ
2derivative and produce Maximum-likelihood estimation,
Wherein
be the estimation of μ, n is the number of sample,
be the velocity amplitude of i-th sample, it is also called sample mean, because it is the arithmetic mean of all samples. because
estimating without inclined (uniformly minimum variance unbiased, UMVU) of same minimum variance, its normal distribution [45], such as
therefore,
standard error be with
proportional, such as, sample set is larger, and evaluated error is less.Estimate σ
2, use sample variance s
2, its square root s is called sample standard deviation.
Wherein, n is the number of sample,
be the velocity amplitude of i-th sample, wherein
for sample variance,
Because it is sample
variance.Use s
2(instead of
) estimate σ
2, because s
2be unbiased esti-mator and
biased estimator [45].In order to avoid storer overflows, record only has the sample of the some of relative velocity, therefore only has the sample of nearest 5s to be considered for estimated parameter μ and σ.Relative velocity can be obtained
obey following normal distribution:
Embodiment eight: one of present embodiment and embodiment one to seven unlike: the correlative factor required for the calculating two car chain-circuit time delay described in step 3 one, is specially:
Assuming that absolute and relative speed Normal Distribution, the chain-circuit time delay in two workshops connected arbitrarily can be calculated by the distance relatively travelled.The chain-circuit time delay in two workshops is regarded as a stochastic variable T.Depend on following 3 points: 1) the relative velocity v of two cars, 2 distribution height of this stochastic variable) initial distance d between two cars, 3) their relative travel direction.Because perhaps these three factors can change in two blockchain termination process, forecast model must adjust prediction principle adaptively and calculate result accurately.
In order to determine the distribution of T, first the present invention introduces the concept of relative operating range L.Suppose that two cars travel in the same way, if the speed of a motor vehicle of rear car is larger than the front truck speed of a motor vehicle, L becomes r+d, and r is communication radius, otherwise L=r-d.When two car Facing Movements time, if they are mutually away from, L=r-d.Otherwise, L=r+d because they towards each other the other side travel.Relatively increasing gradually apart from inner after having crossed engagement point, each other just mutual away from.
Embodiment nine: combine
fig. 10 understands present embodiment, one of present embodiment and embodiment one to eight unlike: the process of the chain-circuit time delay of calculating two car described in step 3 two is: the distribution function CDF of stochastic variable T is as follows,
Wherein,
it is the value of relative velocity vector between two cars.
In above-mentioned equation, both sides are to t differentiate, and the probability density function of stochastic variable T is:
Because the value of relative velocity vector
normal Distribution, equation is write as following formula again,
In formula, μ and σ is mean value and the standard deviation of relative velocity vector.Therefore, the chain-circuit time delay of expection can calculate with following formula,
Because the value of relative velocity vector
normal Distribution, almost the speed of 99% is to the value spending vector
be distributed in [μ-4 σ, μ+4 σ] scope, so the value of definition velocity vector
possible maximal value and minimum value are respectively μ+4 σ and μ-4 σ, and therefore, the interval of definite integral is reduced to from [0, ∞] [L/ (μ+4 σ), L/ (μ-4 σ)].Finally, according to the result of step one and step 2, the chain-circuit time delay predictor formula without traffic lights is:
Two cars are calculated with the chain-circuit time delay of free position at highway by formula (40).
At a time, automobile n is supposed
iand n
jconnect.Based on car speed vector
with
the distribution of value and automobile n
iand n
jbetween relative distance, the chain-circuit time delay between two cars can be estimated.Suppose that the moving window of each car is 5, so automobile n
jpreserve the speed sample in its nearest 5 seconds
neighbours' automobile that these 5 speed samples will be broadcast in next beacon period.Because automobile n
iat n
jradio transmission range in, these speed samples will be by
nireceive.Contrast with its speed sample, n
irelative velocity can be calculated
wherein,
k=1,2 ...., 5.According to 5 speed samples
automobile n
iand n
jbetween the distribution of relative velocity vector value can be estimated according to equation 34 and 35.In each beacon, also contains the positional information of an automobile, so n
iand n
jbetween initial distance L
ijcan calculate too.Use estimated value μ
ij, σ
ijand L
ijreplace the μ in equation 40, σ and L, n
iand n
jbetween expection chain-circuit time delay calculate.
If perhaps the speed of an automobile can brake suddenly due to automobile above and decline dramatically or overtake other vehicles due to it and increase.Unexpected change like this in relative velocity can cause the great variety of chain-circuit time delay result.In order to avoid such problem, EMA (Exponential Moving Average) method is adopted to anticipate the speed of an automobile.
In above formula
at the processed velocity amplitude of t,
the value of the instantaneous velocity vector in this moment.Because
velocity vector value sample from 0 to the linear combination of t, and from
arrive
normal Distribution, so
also Normal Distribution.In other words, the forecast model that the present invention proposes stands good in the velocity amplitude sample by the process of EMA method.
Automobile n
iand n
jthe process of chain-circuit time delay prediction strictly follows flow process
fig. 10, order:
According to the analysis of model E LDP, as automobile n
ireceive one from automobile n
jbeacon, calculate the relative velocity between them and relative distance, then estimate mean value and the standard deviation of relative velocity, finally, the chain-circuit time delay in calculation equation 40.
Simulating, verifying of the present invention is as follows:
Because moving vehicle model is a key factor affecting ELDP model accuracy in VANETs, we produce the movement locus of automobile by VISSIM simulator [46], this simulator is the traffic simulator of a microcosmic and the analysis tool be widely used, and is used for design and assess various traffic system.VISSIM is multiple functional traffic simulator, provides friendly user modeling interface, has the ability for the modeling of mass transportation transportation network, a quite detailed aspect is analyzed the interaction of garage, such as, change or overtake other vehicles, and garage and traffic system.VISSIM can obtain about the detailed status variable information in the time scale of each car even than more accurate [47] of wonderful level.They can simulate the street network on surface, and certainly by road, interchange etc., stop or traffic control intersection.
When the chain-circuit time delay of the garage that prediction two connects, ELDP considers three principal elements: the distribution of relative velocity, traffic lights and motor turning.In order to evaluate the performance of ELDP model, our two scenes reflect the impact of above 3.First, in highway scene, the metastable speeds of automobile, how about the distribution that we analyze relative velocity will affect the accuracy of model.The second, in City scenarios, because traffic lights causes car speed frequent variations, how we affects predicting the outcome of model if analyzing traffic lights.Finally, we also have rated the impact of motor turning on model prediction accuracy.At highway and City scenarios, the relative moving direction in two workshops also greatly can affect the performance of ELDP model.When two cars towards each other the other side travel or mutually away from time, the relative distance between them can keep correspondingly reducing or increasing, and therefore ELDP model can predict the chain-circuit time delay between them exactly.But, when two cars travel in the same way with very close speed, followed by automobile perhaps overtake other vehicles, perhaps can not overtake other vehicles, ELDP model does not temporarily predict until relative velocity becomes larger than threshold value.
A. highway scene
Just
as Fig. 1shown highway scene
, tablelattice 1 are configured with the analog parameter .T type crossing of all highway scenes in the centre of road, and the angle between vehicle traveling direction and transverse axis is α, α ∈ [0,2 π].In order to understand relative movement direction and motor turning to the impact of ELDP model, we divide automobile in groups to according to travel direction: (1) automobile travels in the same way on the road of keeping straight on, that is, angle between two car travel directions is 0 the latter π, (2) two cars are Facing Movement on the road of keeping straight on, that is, angle between the travel direction of a car and x-axis is 0 or π, angle β between another car and x-axis is π or 0, and they do not change travel direction and pass through crossing, (3) two cars travel in the same way before crossing, automobile by not changing travel direction during crossing until link disconnects, but, another automobile has redirect on another road in crossing, steering angle is
(4) two cars Facing Movement before crossing, but, an automobile does not change travel direction until link disconnects, and another automobile redirect on another road in crossing, steering angle
in each second, each automobile collect it with the speed of neighbours, and calculate their relative velocities and relative distance, it also records the angle between its oneself and transverse axis.Then, based on the model proposed in these parameters and Section three, the link experiment of a prediction is calculated.From track of vehicle file, we find the link trip time of a pair automobile of any connection easily, and this will be counted as the time of real two cars connections.
Form 1
(1) automobile travels in the same way on the road of keeping straight on
We select two cars travelled in the same way randomly, and depict their time dependent average velocity
as Fig. 1in 1 shown in (a), from this
figureon, it is fast that we clearly find out that average relative velocity changes before 10s, but no longer change after 10s, and this shows that the relative velocity of two cars is fast in these two phase change.
fig. 1in 1, (b) gives the fluctuation tendency of relative velocity, and the change before and after same 10 is obvious, and before 10s, fluctuation tendency is large, and it shows automobile on a highway owing to not having the impact of traffic lights, and its travel speed is steady comparatively speaking.
fig. 1in 1, (c) provides two blockchain roads and started to connect in about first second and disconnect at 78s, in whole simulation process, prediction time delay and simulation time delay two curves very close, this shows that ELDP model can provide and predicts the outcome accurately.We further define predicated error, and predicated error is that the difference of predictions and simulations time delay is divided by simulation time delay.
fig. 1in 1, (d) gives the distribution function of the predicated error of ELDP, to the predicated error of the chain-circuit time delay of automobile, ELDP shows that the error predicted the outcome of more than 95% is less than 0.4% to this.
Let us surprisingly we randomly drawed 10 to the automobile travelled in the same way and depict ELDP prediction average forecasting error
as Fig. 1in 1 shown in (e), from column
in figurewe can find out that all average forecasting errors are less than 8%.This accuracy enough supports great majority application.From the accuracy affecting chain-circuit time delay prediction, u is principal element, and L, σ are secondary causes, there is no turn to owing to travelling on the road of keeping straight on, so the change of angle between two cars and transverse axis is then almost very little.
(2) Facing Movement on the straight road in High-speed Circumstance
If two automobiles travel in the same way on the road of keeping straight on, perhaps their relative distance increase or reduce, and this depends on their relative velocity, and such as relative velocity is larger, and relative distance is longer.But, due to two car Facing Movements, before they meet, the distance between two cars continues to reduce until zero; After meeting, the distance between two cars continues to increase.This makes model prediction accuracy higher, and in order to prove ELDP model prediction accuracy, we are the Stochastic choice automobile of a pair Facing Movement from simulation.
fig. 1in 2, (a) and (b) depict average velocity and the standard deviation of relative speed, as we can see, having almost no change of the relative velocity of two cars, due to Facing Movement, relative velocity ratio in the same way large, because keep straight on, so the speed that relative velocity is two cars is simply added, the mean value of relative velocity is greater than 33 as we can see.The standard deviation of relative velocity is zero, and it shows that the speed of two cars does not almost fluctuate.Show that the speed that in highway environment, two cars travel is relatively stable again.
Because these two nodes are Facing Movement, so they drive towards the other side each other until meet, and then mutual away from, their relative distance continued to reduce before they meet, and continued to increase until link disconnects after meeting.
fig. 1in 2, (c) shows that chain-circuit time delay is connected to disconnection from link and continues to reduce, and this is rational.Contrast with situation in the same way, the tie-time of these two cars is shorter, then increases very soon again this is because their relative distance reduces rapidly, and they have rolled communication range each other away from.?
fig. 1in 2 in (c), prediction time delay and real chain-circuit time delay very close to each other, we depict the distribution function of ELDP predicated error further
as Fig. 1in 2 shown in (d), see as us, the error that predicts the outcome of about 74% is less than 2%, and almost the error of 90% is less than 4%, travel, so the impact of the change ELDP of the angle of vehicle driving trace is extremely secondary cause due on the road of keeping straight on.
We have extracted again the automobile of other 9 pairs of Facing Movements, put together with this and give average forecasting error bar shaped giving 10 pairs of nodes together with example
figure is as Fig. 1in 2 shown in (e).From
in figurefind out, the overall predicated error of Facing Movement is less than what travel in the same way.Most of predicated error is less than 2%.This means that ELDP model can predict chain-circuit time delay more accurately for the automobile of Facing Movement provides.
(3) two cars travel in the same way before crossing, and a car turns in crossing, and another car does not change direction.
fig. 1in 3, (a) and (b) give average relative and did not change in time before 80s, but sustainable growth after this, because a car is at intersection turning, speed changes, so relative velocity correspondingly changes, this variation tendency is seen and is drawn in (b), relative velocity sustainable growth after 80s.?
fig. 1in 3, (c) gives chain-circuit time delay and the simulation time delay of ELDP prediction, two cars connect when 1s and disconnect at 90s, predictions and simulations time delay is very close to each other, and this shows that ELDP can predict the chain-circuit time delay of two cars exactly, even if an automobile is at intersection turning.The error distribution function of ELDP prediction is presented, from this in (d)
in figurecan find out that the error predicted the outcome of more than 85% is less than 10%.
That we have randomly drawed again 10 pairs of nodes and made average forecasting error column
fig. 1in 3 shown in (e), just
as in figureshown in, most of average forecasting error is less than 5%.
(4) two cars Facing Movement before crossing, a car turns in crossing, and another car does not change direction running.
Because an automobile is at intersection turning, average velocity declines in time,
as Fig. 1in 4 shown in (a), due to before a car 7s at intersection turning, so the relative velocity of two cars changed soon before 7s, relative velocity fluctuates smaller after 7s, this
fig. 1in 4, (b) sees and draws.From
fig. 1in 4, (c) finds out, the link of two cars connects from about 1s, disconnects at 16s because two cars Facing Movement and by behind crossing before crossing, a motor turning on another road, so the tie-time is shorter.From being connected to disconnection, closely, this shows that, for this situation, ELDP can predict chain-circuit time delay exactly for prediction time delay and simulation time delay,
fig. 1in 4, (d) gives the error distribution function of prediction chain-circuit time delay, from
in figure, we can find out, the error that predicts the outcome of 88% is less than 4.2%.We adopt the average forecasting error of 10 pairs of random nodes to make
fig. 1the bar shaped of (d) in 4
figure, as we expect, the error of 90% is less than 4%.Results of these predictions show Facing Movement and less in the predicated error of the automobile of intersection turning with a car.In a word, even if automobile at intersection turning to another road, ELDP model also can predict the chain-circuit time delay of any two cars under highway scene exactly, and therefore, the angle between vehicle traveling direction and x-axis is not the principal element of impact prediction error.Find out from this situation, average relative velocity is greater than 33m/s, exactly because larger average velocity determines the accuracy of prediction.The standard deviation of relative distance and relative velocity is not the principal element of impact prediction error.And, little than in the same way of the predicated error that we find Facing Movement, this is rational, because the distance between two cars is faster than change in the same way, and relative velocity ratio in the same way large, even if having a car at intersection turning, result is also like this.Therefore, ELDP model can predict the chain-circuit time delay that two cars under highway scene travel with any direction (such as, intersection turning, change) exactly.
The aggregate analysis of the chain-circuit time delay of C.ELDP prediction
In order to understand how ELDP model perform along with different parameters, we depict has different u, the chain-circuit time delay of σ and L
as Fig. 1shown in 5.At this
in figure, u ∈ (0,20) and σ ∈ (0,5) L is set to 50m, 100m, 150m and 200m.From this
in figurewe can find out, when L and σ is fixed, chain-circuit time delay increases along with the reduction of u.Particularly when u ∈ (0,10) time, chain-circuit time delay change greatly.When u and L fixes, chain-circuit time delay reduces along with σ and increases.When σ level off to 0 time, altering a great deal of chain-circuit time delay.If u and σ is constant, chain-circuit time delay increases along with L and increases.As μ >10, no matter how σ and L changes, and the change of chain-circuit time delay can be very little.We find that the change of u is larger than other two parameter influence chain-circuit time delays.
In addition, the angle (α, β) between traval trace and transverse axis exists
fig. 1directly do not discussed in 5, but, when we calculate three parameter (u, σ, L), be we consider α, β.
Conclusion: LDP model is expanded as ELDP model, utilizes distribution instead of the instantaneous velocity of relative velocity in model.Except relative velocity, model considers vehicle headway, the impact of traffic lights and the angle of crossing's automobile turning.Based on this model, it is that an automobile dynamically can predict it oneself and the chain-circuit time delay of neighbours' automobile that actual solution is designed object.In all parameters, average relative velocity is the most important factor affecting chain-circuit time delay, therefore, estimates that this parameter becomes of crucial importance exactly.In order to avoid the impact of unexpected velocity variations, we use EMA method to speed sample process.Analog result shows that the chain-circuit time delay of ELDP model prediction VANETs is applicable, actual.Particularly can predict the chain-circuit time delay of highway scene very exactly.Because each automobile only needs to gather and and speed sample in its nearest 5s of neighbours' Car sharing.This research network overhead is little.ELDP model after expansion can be employed the chain-circuit time delay predicting that any two cars are turned at urban environment and highway environment crossing.Our next step plan is the accuracy proving ELDP by real data.
The present invention shows the speed strictly Normal Distribution of automobile; Therefore, use the relative velocity of two cars instead of instantaneous velocity to predict the chain-circuit time delay of the expection of garage.
The present invention extends chain-circuit time delay forecast model (the Link Duration Prediction proposed in document [3], LDP) function, after expansion, model is chain-circuit time delay forecast model (the Extended Link Duration Prediction of expansion, ELDP), such model can predict the chain-circuit time delay between each automobile and its all neighbours' automobiles practically, the present invention is mainly similar to the distribution of the relative velocity of vehicle, by considering that the chain-circuit time delay of expection is predicted in the impact of workshop initial distance, traffic lights and steering angle.In addition, present invention employs EMA (Exponential Moving Average) method to carry out processing speed sample and solve the problem that the speed of a motor vehicle changes suddenly.Result proves, each car only needs the sample gathering nearest 5s to realize the accuracy of chain-circuit time delay prediction.
The model that the present invention proposes is first model for the chain-circuit time delay real-time estimate in highway scene and City scenarios between any two nodes, automobile is considered, so this model is more suitable for real highway and City scenarios in the distance factor of intersection turning, traffic lights, relative velocity and time dependent two garages.
List of references of the present invention is as follows:
[1]T.Taleb,E.Sakhaee,A.Jamalipour,K.Hashimoto,N.Kato,and Y.Nemoto,“A Stable Routing Protocol to Support ITS Services in VANET Networks,”Vehicular Technology, IEEE Transactions on,vol. 56,no. 6, pp. 3337–3347, 2007.
[2]W. Viriyasitavat, F. Bai, and O. Tonguz, “Dynamics of Network Connectivity in Urban Vehicular Networks,” Selected Areas in Communications, IEEE Journal on, vol. 29, pp. 515–533, Mar.2011.
[3]Xiufeng Wang,Chumeng Wang,Gang Cui and Qing Yang“Practical Link Duration Prediction Model in Vehicular Ad Hoc Networks”,International Journal of Distribute Sensor Networks.
[4]R. Miller, Q. Huang, “An Adaptive Peer-to-Peer Collision Warning System,” In Proc. IEEE Vehicular Technology Conference (VTC), May 2002, pp. 317–321.
[5]M. Torrent-Moreno, J. Mittag, P. Santi, and H. Hartenstein, “Vehicleto-Vehicle Communication:Fair Transmit Power Control for Safety-Critical Information,” In IEEE Trans. Veh. Technol., V. 58,Issue 7,September 2009, pp. 3684–3707.
[6]A. Nandan, S. Das, G. Pau, and M. Gerla, “Co-operative downloading in vehicular ad-hoc wireless networks,” In Proc. IEEE Conference on Wireless On-demand Network Systems and Services(WONS), January 2005, pp. 32–41.
[7]O.K. Tonguz, M. Boban, “Multiplayer Games over VANET: a New Application,” Ad Hoc Networks Journal (Elsevier), Vol. 8, No. 5, July 2010.
[8]T. Nadeem, S. Dashtinezhad, C. Liao, and L. Iftode, “Traffic view: Traffic data dissemination using car-to-car communication,” In ACM Sigmobile Mobile Computing and Communications Review,Special Issue on Mobile Data Management, Vol. 8, Issue 3, July 2004, pp.6–19.
[9]C. Sommer, A. Schmidt, R. German, W. Koch and F. Dressler,“Simulative Evaluation of a UMTS-based Car-to-Infrastructure Traffic Information System,” In IEEE Workshop on Automotive Networking and Applications (AutoNet), December 2008.
[10]O. Jetter, M. Killat, J. Mittag, F. Schmidt-Eisenlohr, J. Dinger, and H. Hartenstein, “Simulative Analysis of Vehicle-to-X Communication considering Traffic Safety and Efficiency,” In: W.E. Nagel,D.B. Krner, M. Resch (Eds.): “High Performance Computing in Science and Engineering,” Springer,December 2009.
[11]J. Nzouonta, N. Rajgure, G. Wang, and C. Borcea, “VANET Routing on City Roads using Real-Time Vehicular Traffic Information,” In IEEE Trans. Veh. Technol., Vol. 58, No. 7, September2009, pp. 3609–3626.
[12]Alsharif N, Cespedes S and Shen X.M,“iCAR: Intersection-based Connectivity Aware Routing in Vehicular Ad hoc Network.”In ICC, pages 1736-1741.2013.
[13]M.Hu, Z.Zhong, H.W and M.Ni, “Effect of Fading Channel on Link Duration in Vehicular Ad Hoc Networks,” in WCNC, 2013 Proceedings IEEE, (Shang hai, China.), pp. 1687–1691, Apr. 2013.
[14]K.-H. Jung, W.-S. Lim, J.-P. Jeong, and Y.-J. Suh, “A link contact duration-based routing protocol in delay-tolerant networks,” Wireless Networks, vol. 19, no. 6, pp. 1299–1316, 2013.
[15]J.Boleng,W.Navidi and T.Camp, “Metrics to enable adaptive protocols for mobile ad hoc networks”,in Proceedings of the international conference on wireless networks(ICWN02),2002,pp.293-298.
[16]M.Boban,G.Misek,and O.Tonguz,“What is the best achievable QoS for unicast routing in VANETs”in GLOBECOM Workshops,2008 IEEE,Nov 2008,pp.1–10.
[17]H. Luo and D. Laurenson, “Link-duration-oriented route lifetime computation for aodv in manet,” in Wireless Communications and Signal Processing (WCSP), 2010 International Conference on,Oct 2010, pp.1–4.
[18]W. Yuen, S. Chen, K. Huang, and H. Yoshiba, “Link duration based routing protocol for multihop ad hoc networks,” Aug. 23 2007, US Patent App. 11/357,867. [Online]. Available: http://www.google.com/ patents /US 20070195702
[19]Y.-T. Wu, W. Liao, C. Tsao, and T.-N. Lin, “Impact of node mobility on link duration in multihop mobile networks,” Vehicular Technology, IEEE Transactions on, vol. 58, no. 5, pp. 2435–2442,Jun 2009.
[20]C. Tsao, Y.-T.Wu,W. Liao, and J.-C. Kuo, “Link duration of the random way point model in mobile ad hoc networks,” in Wireless Communications and Networking Conference, 2006. WCNC 2006.IEEE, vol. 1,April 2006, pp. 367–371.
[21]A. B. McDonald and T. Znati, “A Path Availability Model for Wireless Ad-Hoc Networks,”Proc. IEEE WCNC 1999, Sep. 1999, pp.35-40.
[22]X. Hong, M. Gerla, G. Pei, and C.-C. Chiang, “A Group Mobility Model for Ad Hoc Wireless Networks,” Proc. ACM international Workshop on Modeling, Analysis and Simulation of Wireless and Mobile Systems (MSWiM), Aug. 1999.
[23]K.H. Wang and B. Li. “Group Mobility and Partition Prediction in Wireless Ad-Hoc Networks,” IEEE ICC 2002, Apr. 2002, pp. 1017-1021.
[24]F.Bai,N.Sadagopan,B.Krishnamachari, and A.Helmy, “Modeling path duration distribution in MANETs and their impact on reactive routing protocols,” Selected Areas in Communications, IEEE Journal on, vol.22, pp. 1357-1373, Sep.2004
[25]N. Sadagopan, F. Bai, B. Krishnamachari, and A. Helmy.Paths: analysis of path duration statistics and their impact on reactive manet routing protocols.In MobiHoc ’03: Proceedings of the 4th ACM international symposium on Mobile ad hoc networking & computing, pages 245–256, New York,NY, USA, 2003. ACM.
[26]F.Bai, N.Sadagopan, B.Krishnamachari, and A.Helmy, “Modeling path duration distributions in MANETs and their impact on reactive routing protocols,” Selected Areas in Communications, IEEE Journal on, vol. 22, pp. 1357–1373, Sep. 2004.
[27]S.Shelly and A.V.Babu, “Analysis of Life Time in Vehicular Ad Hoc Networks of Free-Flow Traffic State”,Wireless Personal Communications, Vol 75,Issue 1, pp 81-102,Mar.2014.
[28]H.Zhu,L.Fu,G.Xue,Y.Zhu, M.Li and L.Ni, “Recognizing Exponential Inter-Contact Time in VANETs”, in INFOCOM, 2010 Proceedings IEEE,pp.1-5,Mar.2010.
[29]S.Shelly and A.V.Babu, “Probability Distribution of Link Life Time in Vehicular Ad Hoc Networks” , in Proceedings of the 2013 International Conference on Advances in Computing,Communications and Informatics, ICACCI 2013, pp:1418-1423, 2013.
[30]G. Yan and S. Olariu, “A Probabilistic Analysis of Link Duration in Vehicular Ad Hoc Networks”,Intelligent Transportation Systems, IEEE Transactions on, vol.12, pp.1227-1236,Dec.2011.
[31]J. Yoon, M. Liu, and B. Noble, “Random waypoint considered harmful,” in INFOCOM 2003.Twenty-Second Annual Joint Conference of the IEEE Computer and Communications. IEEE Societies,vol. 2, March 2003, pp. 1312–1321 vol.2.
[32]S. Cho and J. Hayes, “Impact of mobility on connection in ad hoc networks,” in Wireless Communications and Networking Conference,2005 IEEE, vol. 3, March 2005, pp. 1650–1656 Vol. 3.
[33]E. Y. Hua and Z. J. Haas, “Path Selection Algorithms in Homogeneous Mobile Ad Hoc Networks,” ACM IWCMC ‘06, July 2006, pp. 275–280.
[34]Z.J.Haas and E.Y.Hua,“Residual link lifetime prediction with limited information input in mobile ad hoc networks,”in Proc.INFOCOM 2008, Phoenix,AZ,Apr. 2008.
[35]R. Korsnes, K. Ovsthus, F. Y. Li, L. Landmark, and O. Kure, “Link lifetime prediction for optimal routing in mobile ad hoc networks,”MILCOM 2005, Atlantic City, NJ, Oct. 2005
[36]E.Hua and Z.Haas, “An algorithm for prediction of link lifetime in MANET based on unscented kalman filter ,” Communication Letters,IEEE,vol.13 pp. 782-784, Oct. 2009.
[37]M. Nekovee, “Modeling the spread of worm epidemics in vehicular adhoc networks,” in Vehicular Technology Conference, 2006. VTC 2006-Spring. IEEE 63rd, vol. 2, May 2006, pp. 841–845.
[38]M. L. Sim, M. Nekovee, and Y. F. Ko, “Throughput analysis of wi-fi based broadband access for mobile users on the highway,” in Networks,2005. Jointly held with the 2005 IEEE 7th Malaysia International Conference on Communication.,2005 13th IEEE International Conference on, vol. 1, Nov2005, pp.21–26.
[39]W.Sun, H.Yamaguchi, K.Yukimasa, and S.Kusumoto, “Gvgrid: A qos routing protocol for vehicle ad hoc networs,” in Quality of Service,2006.IWQoS 2006.14th IEEE International Workshop on.IEEE,2006,PP.130-139.
[40]M.Rudack, M.Meincke, K.Jobmann, and M.Lott, “On dynamical aspects of inter vehicle communications(ivc),”in Vehicular Technology Conference,2003.VTC 2003-Fall.2003 IEEE 58th,vol.5.IEEE, 2003,pp. 3368- 3372.
[41]M. Rudack, M. Meincke, and M. Lott, “On the Dynamics of Ad-Hoc Networks for Inter Vehicle Communications,” In Proc. International Conference on Wireless Networks (ICWN), 2002.
[42]M.M. Artimy, W. Robertson, and W.J. Phillips, “Connectivity in intervehicle ad hoc networks,”In Proc. of IEEE Canadian Conference on Electrical and Computer Engineering, Vol. 1, 2004, pp.293–298.
[43]M.M. Artimy, W.J. Phillips, and W. Robertson, “Connectivity with static transmission range in vehicular ad hoc networks”, In Proc.Communication Networks and Services Research Conference(CNSR),2005, pp. 237 –242.
[44]M. Hu, Z. Zhong, H. Zhu, M. Ni, and C.-Y. Chang, “Analytical modeling of link duration for vehicular ad hoc networks in urban environment,” in Proceeding of the Tenth ACM International Workshop on Vehicular Inter-networking, Systems, and Applications, ser. VANET’13. New York, NY,USA: ACM, 2013, pp. 61–70.
[45]K. Krishnamoorthy, Handbook of Statistical Distributio- ns with Applications, ser. Statistics: A Series of Textbooks and Monographs. Taylor & Francis, 2010.
[46]K. Shafiee, J.B. Lee, V. C.M. Leung, G. Chow, “Modeling and Simulation of Vehicular Networks,” In: Proceedings of the first ACM international symposium on Design and analysis of intelligent vehicular networks and applications, New, York, NY, USA: ACM , Miami, Florida, USA,NOV.2011,pp.77-86.
[47]D. Gettman and L. Head, “Surrogate Safety Measures From Traffic Simulation Models, Final Report, “Final Report prepared for the Federal Highway Administration, U.S. Department of Transportation, 2003.