CN111601278B

CN111601278B - Software-defined heterogeneous Internet of vehicles access management and optimization method

Info

Publication number: CN111601278B
Application number: CN202010363894.6A
Authority: CN
Inventors: 周海波; 严潇琳; 钱博; 许云霆; 马婷
Original assignee: Nanjing University
Current assignee: Nanjing University
Priority date: 2020-04-30
Filing date: 2020-04-30
Publication date: 2023-05-05
Anticipated expiration: 2040-04-30
Also published as: CN111601278A

Abstract

The invention discloses a software-defined heterogeneous Internet of vehicles access management and optimization method, which comprises the following steps: step 1: designing a vehicle communication network architecture based on SDN to realize global management of a vehicle network; step 2: setting up a local database on a data plane, collecting vehicle information, and storing a management policy issued by a control plane; step 3: uploading the collected vehicle data to a control plane, and calculating the vehicle network performance index on the control plane; step 4: and determining the weight and benefit function of each performance index, determining the selection of the access network mode through an evolutionary game method, and issuing the selection to the vehicle on the data plane. The method for managing and optimizing the heterogeneous Internet of vehicles access by the software definition flexibly selects the vehicle access mode based on different service requirements, so that the network service quality of the heterogeneous network of the vehicle is optimized according to the requirements of the required service, the network performance is greatly improved, and the service quality of the communication network of the vehicle is improved.

Description

Software-defined heterogeneous Internet of vehicles access management and optimization method

Technical Field

The invention belongs to the technical field of Internet of vehicles, and relates to a heterogeneous Internet of vehicles access management control and access network optimization method based on SDN.

Background

As a key technology in Intelligent Transportation Systems (ITS), vehicle ad hoc networks (VANETs) may connect vehicles in road and urban areas and provide wireless communication between vehicles, between vehicles and road side infrastructure. The VANETs communication technology and the vehicle sensing capability are utilized in advanced applications, for example, the VANETs can support various vehicle services such as vehicle information exchange (such as speed, acceleration and position), security services (including cooperative collision warning, intersection collision prevention and remote vehicle diagnosis), entertainment services (such as online video, online games, online uploading and downloading), and the like. Moreover, VANETs enhance the collaboration between vehicles, pedestrians and traffic infrastructure, which will help eliminate the current 80% of road traffic accidents and help promote more intelligent, safer ground traffic systems for the automotive and telecommunications industries.

Due to the wide existence of cellular base stations and the well-established infrastructure, conventional vehicle communication methods are implemented through cellular base station forwarding. Cellular access technologies, such as LTE (Long Term Evolution ) technologies, can provide reliable and wide internet access for vehicle communications, and the coverage of base stations is larger than other modes of communication, and thus the communication distance of the vehicle is longer. As technology matures, many autonomous automobile manufacturers (e.g., GM-OnStar and BMW-connectieddrive) choose cellular access technology as an automobile internet access solution. However, there is also a problem with cellular access technology, first of all overload. According to the study published by cisco, the total number of mobile links increased from 86 to 123 million by 2022, 2017. At the same time, mobile data traffic will proliferate to 77 bytes per month. And by 2022, the average traffic per mobile user (all equipment and service plans) is expected to increase from 2.3GB per month to 13.3GB per month in 2017. At the same time, the average flow per connection increases from 1.3GB to 6.3GB. The use of cellular networks for in-vehicle internet access only exacerbates the existing overload problem and further reduces the performance of vehicle user equipment and normal cellular user equipment, and the dramatic increase in mobile services results in a severe shortage of cellular spectrum resources.

To solve the overload problem caused by the single use of the cellular network, communication between vehicles can be performed by using heterogeneous networks. The heterogeneous network differs from other communication networks in two ways: (1) There are multiple access network technologies in heterogeneous networks, and the transmission frequency bands and transmission protocols of the various access networks are different, each having advantages and disadvantages. (2) Users may access the network through different access technologies according to different needs rather than communicating in the same way only. Therefore, the heterogeneous network has obvious advantages compared with a single communication network, namely the overload problem caused by the single network is improved, the advantages of each access network can be exerted, the advantages are complemented, and the user can select the optimal access mode to communicate according to different service demands. And because the base station comprises a plurality of access networks, the base station also comprises a plurality of operator equipment, so that the base station can realize distribution and lighten the pressure of the base station. Meanwhile, the cellular base station forwards, and the D2D multiplexes the cellular spectrum and the special spectrum to carry out vehicle communication, so that the spectrum utilization rate can be increased, and the network performance can be improved. However, with different access networks, the existing conventional communication architecture makes the base station implement not only data collection and transmission functions, but also control functions, which may result in a larger device pressure ratio of the base station.

In summary, the problems now exist: (1) The overload problem of the cellular network is aggravated by singly using the cellular network for communication, and the communication quality requirement of the intelligent traffic system service cannot be met. (2) The traditional communication architecture takes a cellular base station as a control and data center, the equipment pressure ratio of the base station is larger, and the architecture using SDN can realize separation of data and a control plane, so that flexible control of network flow is realized. The significance of solving the technical problems is that: the overload problem of the cellular network is solved, the access network of the vehicle is flexibly controlled, and the service quality of the whole vehicle network system is improved.

Disclosure of Invention

Aiming at the problems existing in the prior art, the invention realizes a software-defined heterogeneous Internet of vehicles access management and optimization method.

The invention is realized by establishing a vehicle communication network management model based on software definition, and controlling modes of a vehicle access network, including three modes of base station forwarding, D2D (device-to-device) and DSRC (dedicated short distance communication), through an SDN (Software Defined Network ) controller. The network access management model based on SDN comprises a data plane and a control plane, wherein the data plane is responsible for collecting the state and network information of the vehicle, and the control plane calculates the performance index of the vehicle communication network, including transmission rate, interference, transmission delay and energy loss, according to the collected vehicle information by a reverse power control method. And then optimizing the network service quality of the heterogeneous network of the vehicle according to the obtained performance index in a control plane so as to achieve the optimal network service quality according to the requirement of the required service. And finally, issuing the obtained optimized access network mode selection result to each vehicle on the data plane. The specific optimization method comprises the following steps: and establishing a benefit function required by the game, and maximizing the benefit function by a method of evolutionarily playing the game, thereby managing the access network mode of the vehicle.

The invention discloses a heterogeneous Internet of vehicles access management and optimization method based on software definition, which comprises the following steps:

step 1: designing a vehicle communication network architecture based on SDN to realize global management of a vehicle network;

step 2: setting up a local database on a data plane, collecting vehicle information, and storing a management policy issued by a control plane;

step 3: uploading the collected vehicle data to a control plane, and calculating the vehicle network performance index on the control plane;

step 4: determining the weight and benefit function of each performance index, determining the selection of an access network mode through an evolutionary game method, and issuing the selection to a vehicle on a data plane;

wherein, the performance index obtained in the step 3 is obtained by the following steps:

step 3.1: obtaining probability density function f of distance between vehicles according to distribution condition of vehicles _V (x) And a probability density function f of the distance between the vehicle and the base station _C (x) I.e. a movement model of the vehicle;

step 3.2: the energy loss is represented by the transmission power of the vehicle, and for the vehicle communicating through the cellular network, the transmission power distribution of the vehicle is determined through a mobile model and a channel reverse power control strategy, wherein the reverse power control strategy is that the transmission power is reversely deduced by setting a receiving power threshold value

I.e. transmit power +.>

wherein ρ₀ Represents the received power threshold, h is the channel gain, X represents the communication mode selected by the vehicle, d _X Representing the communication distance of the vehicle, gamma _X Representative pathLoss index;

step 3.3: for a vehicle communicating through a cellular network, the interference and signal-to-interference-and-noise ratio SINR of the vehicle is obtained from the transmission power of the vehicle, namely:

interference to the ith vehicle:

wherein N represents a collection of vehicles, N\i represents the rest of vehicles except the vehicle I, I _D,i And I _C,i Respectively representing the interference caused by the vehicle selecting the D2D mode communication and the base station forwarding mode communication to the vehicle i, P _D,j And P _C,j Respectively representing the transmission power of the vehicle j selecting the D2D mode communication and the base station forwarding mode communication, γ _D And gamma is equal to _C Path loss index, D, respectively representing the vehicle selecting D2D mode communication and base station forwarding mode communication _ji Is the distance from the source of the interference to the receiver,

and />

Is an indication function +.>

Indicating the selection of D2D mode, ">

Indicating selection of a base station forwarding mode;

SINR of the vehicle for D2D uplink transmission is expressed as:

wherein n₀ Is the noise power;

step 3.4: for vehicles communicating through the cellular network, the SINR outage probability of the vehicle is obtained through the transmission power, and then the transmission rate, throughput and time delay of the vehicle are obtained,

the SINR outage probability of a vehicle communicating over a cellular network is:

wherein ,

representing the probability of occurrence of (·), +.>

Represents SINR outage threshold, ++>

And->

The Laplacian transformation of the random variable PDF when the D2D mode and the cellular mode are selected;

the average transmission rate of the vehicle is:

where B represents the bandwidth of the cellular network,

indicating the expectations of (, p) _o,i (x) Representing the SINR outage probability found above,

the delay is:

step 3.5: a vehicle communicating in DSRC mode, which is flattened based on the characteristics of IEEE802.11p networkEqual transmission rate and delay, for a vehicle performing DSRC transmission, using T _e 、T _s and T_c To indicate the duration of idle, the average time that channel busy was detected due to successful transmission, and the average time that channel busy was detected by the vehicle during a collision, where T _e =δ, i.e. a time gap, T _s and T_c The method comprises the following steps:

wherein E [ P ]]H=phy, the average size of the packet _hdr +MAC _hdr Is the header length of the packet, SIFS is the short inter-frame time, DIFS is the distributed inter-frame time, ACK is the time of the reception point to the transmission point acknowledgement;

therefore, the average latency of a vehicle communicating through DSRC is:

t _DS ＝(1-p _t )T _e +p _s T _s +p _c T _c ，

wherein p_s Representing the probability of any DSRC mode vehicle within the carrier sense range to successfully transmit data in one time slot, denoted as p _s ＝nτ(1-τ) ^n-1 Where n is the number of vehicles selecting DSRC mode, p _t Representing the probability that at least one transmission occurs within a slot time, p _t ＝1-(1-τ) ⁿ And τ is the average transmission probability of the DSRC transmission vehicle, p _c Representing the probability of collision of a vehicle transmission during a slot time, where p _c ＝p _t -p _s ；

The average rate of vehicle communication through DSRC is:

in the mode, a user adopts a competitive communication mode to carry out data transmission, the DSRC mode is mainly based on IEEE802.11p protocol, adopts a CSMA/CA mechanism, does not use a channel of a cellular network, and uses an independent channel, so that the channel state can be detected until the channel is idle before data is transmitted under the mechanism, and then the data is transmitted, so that collision is avoided, and the mechanism avoids interference;

the optimization method of the step 4 comprises the following steps:

step 4.1: and (3) constructing a benefit function model by the weight of the vehicle network performance index obtained in the step (3), wherein the benefit function of the vehicle i for selecting the j-th mode to communicate is defined as follows:

wherein ,

for normalized performance index vector, +.>

The normalized weight vector;

step 4.2: performing evolutionary game on each vehicle according to the benefit function model, selecting a communication mode for each player in the evolutionary game by using a strategy, and calculating the average benefit function of all players when the kth round of game is performed

Setting j=0;

step 4.3: calculating the benefit function w of each vehicle _ij Differences from the average benefit function of the previous round

a. If it is

The benefit function value of the strategy j is better than the average benefit function of the hybrid strategy of the previous round of game, and the player i selects the current strategy in the round of game;

b. if it is

The current strategy is inferior to the hybrid strategy of the previous round of game, the player i randomly selects an unselected strategy j, and then the process (4.3) is repeated until ∈>

If the ith game player has selected all strategies, selecting the strategy with the highest benefit function;

c. repeating this step for each gambler;

step 4.4: and updating the strategies of all users in each round of game until the strategies of all users are unchanged, so that evolutionary equilibrium is achieved.

Further, in step 1, the communication network architecture includes a data plane and a control plane, the data plane includes a local database for collecting data of the vehicle, transmitting the data to the control plane and receiving a management policy issued by the control plane, the control plane includes a global database and an SDN controller, calculating a vehicle communication performance index according to the collected vehicle information, and determining a network access policy.

Further, heterogeneous internet of vehicles access technologies include cellular network and DSRC access technologies, where cellular network technologies include two modes of D2D relaying and multiplexing cellular channels through base stations.

Further, in the step 2, the data plane includes a vehicle moving scene, the collected vehicle information includes status information such as speed, acceleration, position, etc. of the vehicle, and the updated policy information includes an access network control policy, that is, a manner of controlling the vehicle on the data plane to access the network.

Further, the vehicle network performance indexes in the step 3 include throughput, transmission rate, interference and energy loss

Compared with the prior art, the method for managing and optimizing the heterogeneous Internet of vehicles access by the software design provides a solution for managing the vehicle communication network, improves the overload problem caused by the fact that the vehicle uses a single cellular network by adding the heterogeneous network, and can also meet the communication quality requirement of intelligent traffic system service. Meanwhile, the SDN architecture is used, so that a data plane is separated from a control plane, and the pressure of a base station is reduced; secondly, the access management optimization method based on the evolutionary game is one of the outstanding contributions of the patent, a benefit function is established according to the weight of the performance index, and the benefit function is optimized, so that the access mode of the vehicle is optimized, and the service quality of the whole vehicle network system is improved; and thirdly, the differentiated management can be carried out aiming at the services with different service requirements, and different weights of the performance indexes can be set according to different service requirements, so that the targeted optimization can be carried out. The optimization management method of the invention can flexibly select the access mode of the vehicle based on different service requirements, so that the network service quality of the heterogeneous network of the vehicle can be optimized according to the requirements of the required service. Compared with the traditional architecture of a single network, the invention can flexibly manage the access network, greatly improve the performance of the network and improve the service quality of the vehicle communication network

Drawings

FIG. 1 is a heterogeneous Internet of vehicles scene graph based on SDN as employed by an embodiment of the present invention;

FIG. 2 is a block diagram of an implementation of a vehicle access network optimization algorithm in accordance with an embodiment of the present invention;

FIG. 3 is a graph of the benefits of three alternative options for different numbers of vehicles;

fig. 4 shows the average speed for three options for different numbers of vehicles.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the following detailed description of the embodiments of the present invention is given with reference to the accompanying drawings: the embodiment is implemented on the premise of the technical scheme of the invention, and detailed implementation modes and specific operation processes are given. It should be understood that the specific examples described herein are for illustrative purposes only and that the scope of the present invention is not limited to the following examples.

As shown in fig. 1, to implement global management of a vehicle network, the system is implemented by an SDN architecture. The cellular base station controls a forwarding plane of a network switch or a router through an OpenFlow protocol, so that a network path of a network packet is changed, and centralized control is performed through an SDN controller. The architecture of the present invention is largely divided into a data layer and a control layer. The data layer collects information mainly through the base station, and the control layer is mainly responsible for global strategies related to the access network. Including access control, traffic management, movement control, etc.

In the framework, each base station is equipped with a local database for storing collected vehicle information, including density, position, speed, acceleration, etc. The database updates the information of the various parts of the vehicle as new vehicles enter or leave the coverage area of the base station. Vehicle information collected by a plurality of base stations will be sent to an SDN controller. The global database of the control layer stores global vehicle state information and is updated periodically by each base station. The data layer comprises a vehicle mobile communication network consisting of vehicles and base stations and a local database.

The control layer controls the global vehicle mainly through the SDN controller. The SDN controller realizes quality coordination and information sharing of the vehicle communication network by separating data from a control plane. The control layer mainly comprises a global information database and decision modules, including access control, flow control, channel allocation and the like. The decision module calculates the vehicle information collected in the database and runs the programmable application on the controller in a global view of the entire service area to implement the global function. And then sending the decision result to the base station of the data layer to realize the decision.

For communication networks, two vehicles may typically relay information to each other through a base station. However, if the distance between two vehicles is short, they can transmit data in a single hop manner through D2D or DSRC without the need for a base station relay. For cellular mode and D2D mode, they share cellular spectrum. Considering the cellular spectrum, namely the frequency division duplex LTE network, the orthogonal channel group is adopted for uplink and downlink transmission. There are multiple channels in the cellular network and, similar to calculating the interference for all channels, the analysis may be limited to one channel for simplicity. I.e. D2D transmission and cellular uplink transmission, multiplexes one channel within the coverage area of the base station.

In order to simplify the analysis of the vehicle movement model, a uniform distribution model of the vehicle over the road is established, each base station communication range covering a stretch of length L which is part of the entire stretch. Taking three base station coverage areas as an example, the steady-state distribution of the vehicle positions r is approximately uniformly distributed:

since the steady-state distribution of the vehicle position is nearly uniform, which means that the vehicle is equally likely to be present at any position, the V2V transmission distance d _V2V Follows a triangular distribution, therefore d _V2V The Probability Density Function (PDF) of (c) can be expressed as:

thus, in the coverage area of a base station, the distance from the vehicle to the midpoint of the road segment

Thus, PDF f of the distance from the vehicle to the base station _C (x) The method comprises the following steps:

wherein ,l₀ For the distance of the base station from the middle point of the road section, and l _max For the furthest distance of the vehicle from the base station

For cellular networks, the communication between a cellular mode and a D2D mode is composed of path loss and shadowingIs a large scale fading effect of (a). According to this model, the decay rate is d ^-γ Where d is the distance between two vehicles transmitting data to each other and γ is the path loss index. In particular, gamma _C and γ_D The path loss index for the cellular uplink and D2D link, respectively. The communication channel follows a rayleigh fading model in which the channel gain h between any two locations follows an independent unit exponential distribution, i.e., h-exp (1).

The method adopts channel reverse power control to adjust the transmitting power of the vehicle, so that the average receiving signal power (including the uplink transmission of the base station and the D2D transmission) of the receiving vehicle reaches a certain receiving power threshold value rho after the power of the transmitting signal passes through the path loss ₀ h。ρ ₀ h is denoted as instantaneous received power.

In general, if the channel reverse power control mode and the required transmit power are greater than the maximum transmit power P _max The signal will be truncated. Thus, if the required transmit power is less than P _max The two vehicles may transmit data using single hop communications without the need to use base station relays. This requires that the distance between the two vehicles must be D _max In relation to a single hop distance threshold. Single hop distance threshold D _max By the P of the transmitting vehicle _max And a received power threshold ρ ₀ And (5) determining. I.e.

The vehicles may select single hop communication, i.e., the distance between the vehicles is less than D _max The probability of (2) is:

probability p that communication is possible only by base station relay method, which is not a potential vehicle for single hop communication _c0 ＝1-p _v0 。

The distance between the two vehicles must be D _max In this, this is referred to as a potential single hop communication vehicle pair. The potential single hop communication vehicle pair is not necessarily selectedThe D2D/DSRC mode is selected for transmitting data. They may also select the mode of the base station relay.

The probability density function of a potential single hop communicating vehicle is

For the DSRC mode, the DSRC vehicle first perceives the channel before transmitting the data due to the CSMA/CA (Carrier Sense Multiple Access with Collision Avoid, carrier sense multiple access with collision avoidance) mechanism. If a busy channel is detected, there is a random backoff, the length of which is controlled by a backoff counter. The backoff counter randomly selects a value between 0 and CW for initialization, where CW is noted as the initial contention window size. If a channel busy is detected, the channel freezes and the vehicle may transmit data until the backoff counter is decremented to zero. Further, if channel busy is detected, the back-off counter is decremented to 1. If the transmission fails, the contention window will double if the contention window does not reach the maximum contention window size, and the DSRC vehicle will retransmit the packet and repeat the process.

Vehicles can communicate with each other through three modes of base station relay, D2D or DSRC modes. In cellular mode, the vehicle transmits data in two hops through a transmitter-base station-receiver mode. In D2D mode, two vehicles can communicate directly through the multiplexed cellular uplink by effectively managing interference. In DSRC mode, the vehicle uses CSMA/CA mechanism of ieee802.11p protocol to transfer data in one hop, and access network mode selection has a great influence on network performance. The communication mode of each vehicle is reasonably selected, so that the frequency spectrum resources can be effectively multiplexed without generating excessive interference, and the performance of the whole network is improved.

The proportion of relay communication by the base station in the potential single hop communication vehicle is set to a, the proportion of D2D communication is set to b, and the proportion of DSRC mode communication is set to c. The proportion of communication by base station relay in the whole network is p _C ＝p _c0 +ap _v0 The proportion of D2D communication is p _D ＝bp _v0 The proportion of DSRC mode communication is cp _v0 。

From the channel reverse power control model and the distance probability distribution described by the above movement model, the transmit power in the cellular mode and the D2D mode can be deduced.

Probability density function f of distance through vehicle D2D communication mode _D (x) Derived, according to the channel reverse power control model,

probability density function of D2D communication mode vehicle>

Is>

wherein ,

likewise, the transmit power for a vehicle communicating via the cellular relay mode may be determined from the channel reverse power control model:

probability density function of vehicle in cellular relay communication mode>

The method comprises the following steps:

in the network, since the vehicles in the cellular mode and the D2D mode repeatedly use the same channel, interference is not only from the cellular uplink in the cellular mode but also from the D2D link that repeatedly uses the same channel. Thus, herein, interference of other cellular uplinks and D2D links covered by the same base station is considered. The interference caused by other vehicles not covered by the base station is regarded as noise.

Interference to the ith vehicle:

wherein ,P_D Representing the transmit power, P, of a D2D transmission of a transmission vehicle i _C Representing the transmit power of the cellular uplink transmission. d, d _ji Is the distance from the interfering transmitter to the transmitting receiver. For D2D communication mode, D _ji Refers to the distance from the interfering sending vehicle j to the receiving vehicle i; d for base station relay mode _ji Refers to the distance d from the interfering transmitting vehicle j to the base station _C 。

And

is an indication function +.>

Indicating the selection of D2D mode, ">

Indicating selection of base station forwarding mode, if vehicle j has a transmission request and selects mode X +.>

Otherwise->

wherein />

According to the given total interference I _i Then the vehicles in the heterogeneous Internet of vehicles system can be deducedSINR of a vehicle for D2D/uplink transmission can be expressed as

wherein n₀ Is the noise power.

For vehicles communicating through the cellular network, the SINR interruption probability of the vehicle can be obtained through the transmitting power, and then the transmission rate and the time delay of the vehicle can be obtained.

The SINR outage probability for a vehicle communicating with the cellular network is:

wherein ,

represents SINR outage threshold, ++>

And->

The laplace transform of the random variable PDF when the D2D mode and the cellular mode are selected.

The laplace transform of the interference to the D2D transmission vehicle for other D2D vehicles is:

wherein

The laplace transform of the interference of the vehicle performing the base station transfer communication to the D2D transmission vehicle is:

wherein ,

the same holds true for the laplace transform of the D2D transmitting vehicle's interference with the vehicle that is in base station forwarding communication:

the laplace transform of the interference of the other vehicles performing the base station transfer communication to the vehicles performing the base station transfer communication is:

the average transmission rate of the vehicle is:

the time delay is

Under the CSMA/CA mechanism of IEEE802.11p protocol, considering the performance analysis of DSRC mode, N vehicles are considered in the carrier sensing range, N may be less than or equal to ncp, considering that N is the number of vehicles selecting DSRC communication mode _v0 Is the largest integer of (a). Mainly, the problem of transmission between the vehicle i and the vehicle j is studied, and data are mutually transmitted between the vehicles through a CSMA/CA mechanism. Consider each DSRC transmission vehicleThe average transmission probability is denoted by τ, τ=2/(cw+1) calculated. P is p _s Representing the probability of any DSRC mode vehicle within the carrier sense range to successfully transmit data in one time slot, denoted as p _s ＝nτ(1-τ) ^n-1 。p _t Representing the probability that at least one transmission occurs within a slot time, p _t ＝1-(1-τ) ⁿ . Respectively T _e 、T _s and T_c To represent the channel idle duration, the average time that channel busy was detected due to successful transmission, and the average time that channel busy was detected by the VUE during collision. In the IEEE802.11p standard, T _e =δ, i.e. one time gap. Successful frame transmission T _s And the transmission time T of collision _c Can be expressed as:

wherein E [ P ]]Is the average size of the data packet. H=phy _hdr +MAC _hdr Is the header length of the packet, SIFS is the short inter-frame time, DIFS is the distributed inter-frame time, and ACK is the time to receive point to send point acknowledgements.

Therefore, the average latency of vehicle communication through DSRC is t _DS ＝(1-p _t )T _e +p _s T _s +p _c T _c. wherein p_s Representing the probability of any DSRC mode vehicle within the carrier sense range to successfully transmit data in one time slot, denoted as p _s ＝nτ(1-τ) ^n-1 。p _t Representing the probability that at least one transmission occurs within a slot time, p _t ＝1-(1-τ) ⁿ . And τ is the average transmission probability of the DSRC transmission vehicle.

Average rate of vehicle communication through DSRC is

Note that there is no interference in DSRC mode.

And (3) carrying out performance analysis on the system through the information of the vehicle collected by the data plane, so as to obtain the performance index of the vehicle: transmission rate V, interference I, delay T and energy loss P. Where V is a positive variable and I, T and P are negative variables, and P can be represented by the vehicle transmit power. Since these performance metrics differ in unit, they need to be normalized.

Is provided with

For the actual value of the ith performance indicator, < +.>

and />

The maximum expected value and the minimum expected value of the ith performance index are respectively.

For the forward-direction variable, the reference value,

normalized variable +.>

For negative variables +.>

Normalized variable +.>

Then X is _i After normalization, the attribute is->

The priority for different performance metrics is also different, as the traffic of the vehicle service is different, and the required quality of service requirements are also different. For example, the communication of the emergency information for the vehicle state belongs to the instant service, and the requirement on the time delay is relatively high, so that the priority of the time delay is increased. The priority of different performance indexes can be represented by setting the weight of each performance index, and in order to make the weight setting more reasonable, the weight of each performance index can be determined by a analytic hierarchy process, or can be determined by other methods.

Firstly, establishing a decision matrix, and setting a scale value p _ij Which represents the relative importance of the two properties, taking 1-9, the larger the number the more important i is with respect to j, so the decision matrix p= (P) _ij ) _n×n . In the decision matrix of the present invention,

j ε {1,2,3, …, n }, have:

p _ij >0，

p _ii ＝1

so decision matrix

After the decision matrix is obtained, whether the given scale value is reasonable or not needs to be judged through consistency check. It can be judged by the consistency ratio CR.

Calculating to obtain the maximum characteristic value of P as lambda, and obtaining the consistency proportion CI as

Wherein CI is the consistency ratio,

n is the order of the matrix, the number of performance indicators, which defines 4, i.e. n=4. Whether the decision matrix is reasonable or not needs to be checked by the consistency ratio CR only if CR<0.1 is that the consistency of the decision matrix is acceptable. RI is an average uniformity index.

Order of	1	2	3	4	5	6	7	8	9
										RI	0.00	0.00	0.58	0.90	1.12	1.24	1.32	1.41	1.46

According to the decision matrix P and the maximum eigenvalue lambda of the decision matrix P, a normalized eigenvector corresponding to lambda of P can be calculated

Each vehicle in the coverage area of the three base stations is set as a game player, any one communication mode can be selected, all vehicles in the coverage area of the three base stations are groups, and the selection of the communication mode by each game player in the evolutionary game is a strategy. And the benefit function of each player is determined by the performance metrics together.

The benefit function of evolutionary game is defined as the benefit function, and the benefit function of the vehicle i selecting the j-th mode for communication is defined as

wherein ,

is normalized performance index vector +>

Normalized transmission rate, interference, transmission delay and energy loss P, < >, respectively>

Is the normalized weight vector.

The vehicle i selects j mode can be noted as x _ij ,

j=0, 1,2. j=0 represents communication in the DSRC mode for the vehicle, j=1 represents communication in the cellular relay mode for the vehicle, and j=2 represents communication in the D2D mode in which the cellular channels are multiplexed. The benefit function of communication mode j is

So the average benefit function value of all users in the network is

Defining the difference between the benefit function value of any player i in the kth round of game period and the average benefit function value of all users in the network as

The specific flow of evolutionary game is shown in FIG. 2, and the benefit function w of each vehicle is calculated first _ij Differences from the average benefit function of the previous round

a. If it is

The benefit function value of the strategy j is better than the last round of game confounding strategy (average benefit function), and the player i selects the current strategy;

b. if it is

Then the current strategy is shown to be inferior to the last round of game promiscuous strategy, the gambler i randomly selects one unselected strategy j, and then the previous steps are repeated until +.>

If the ith player has selected all strategies, selecting the strategy with the largest benefit function;

c. this is repeated for each player.

And updating the strategies of all users in each round of game until the strategies of all users are unchanged, so that evolutionary equilibrium is achieved.

To make this embodiment more intuitive to expose software definitionsThe heterogeneous Internet of vehicles access management and optimization method selects a decision matrix with highest transmission rate priority to determine the weight of each network performance index, namely the decision matrix is

Based on this decision matrix, a benefit function is determined. Figures 3 and 4 show the benefits of different numbers of vehicles and the average speed per vehicle for three options of evolutionary game, random selection, single selection base station forwarding mode. It can be seen that the software-defined heterogeneous internet of vehicles access management and optimization method can effectively manage vehicles, and the benefit and the vehicle speed are far greater than those of other two schemes. />

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims

1. A software-defined heterogeneous Internet of vehicles access management and optimization method is characterized in that: the method comprises the following steps:

step 3.1: obtaining probability density function f of distance between vehicles according to distribution condition of vehicles _V (x) Between vehicle and base stationProbability density function f of distance of (2) _C (x) I.e. a movement model of the vehicle;

I.e. transmit power +.>

wherein ρ₀ Represents the received power threshold, h is the channel gain, X represents the communication mode selected by the vehicle, d _X Representing the communication distance of the vehicle, gamma _X Represents a path loss index;

interference to the ith vehicle:

and />

Is an indication function +.>

Indicating the selection of D2D mode, ">

Indicating selection of a base station forwarding mode;

SINR of the vehicle for D2D uplink transmission is expressed as:

wherein n₀ Is the noise power;

wherein ,

representing the probability of occurrence of (·), +.>

Represents SINR outage threshold, ++>

And->

the average transmission rate of the vehicle is:

where B represents the bandwidth of the cellular network,

the delay is:

step 3.5: a vehicle communicating in DSRC mode obtains its average transmission rate and delay according to the characteristics of IEEE802.11p network, and uses T for the vehicle performing DSRC transmission _e 、T _s and T_c To indicate the duration of idle, the average time that channel busy was detected due to successful transmission, and the average time that channel busy was detected by the vehicle during a collision, where T _e =δ, i.e. a time gap, T _s and T_c The method comprises the following steps:

therefore, the average latency of a vehicle communicating through DSRC is:

t _DS ＝(1-p _t )T _e +p _s T _s +p _c T _c ，

The average rate of vehicle communication through DSRC is:

in the mode, a user adopts a competitive communication mode to carry out data transmission, the DSRC mode is mainly based on an IEEE802.11p protocol, adopts a CSMA/CA mechanism, does not use a channel of a cellular network, and uses an independent channel, so that the channel state can be detected until the channel is idle before data is transmitted under the mechanism, and then the data is transmitted, so that collision is avoided, and the mechanism avoids interference;

the optimization method of the step 4 comprises the following steps:

wherein ,

for normalized performance index vector, +.>

The normalized weight vector;

Setting j=0;

a. If it is

b. if it is

c. repeating this step for each gambler;

2. The method for managing and optimizing access to heterogeneous internet of vehicles defined by software according to claim 1, wherein the method comprises the following steps: in step 1, the communication network architecture includes a data plane and a control plane, the data plane includes a local database for collecting data of the vehicle, transmitting the data to the control plane and receiving a management policy issued by the control plane, the control plane includes a global database and an SDN controller, calculating a vehicle communication performance index according to the collected vehicle information, and determining a network access policy.

3. The method for managing and optimizing access to heterogeneous internet of vehicles defined by software according to claim 1, wherein the method comprises the following steps: heterogeneous internet of vehicles access technologies include cellular network and DSRC access technologies, where cellular network technologies include two D2D modes of relaying and multiplexing cellular channels through base stations.

4. The method for managing and optimizing access to heterogeneous internet of vehicles defined by software according to claim 1, wherein the method comprises the following steps: the data plane in step 2 includes a vehicle moving scene, the collected vehicle information includes state information such as speed, acceleration, position, etc. of the vehicle, and the updated policy information includes an access network control policy, that is, a manner of controlling the vehicle on the data plane to access the network.

5. The method for managing and optimizing access to heterogeneous internet of vehicles defined by software according to claim 1, wherein the method comprises the following steps: the vehicle network performance indexes in the step 3 include throughput, transmission rate, interference and energy loss.