CN110601778B - A Cognitive Spectrum Allocation Method for Unidirectional Straight-Road Vehicle Networking Based on Clustering Structure - Google Patents

Abstract

The invention discloses a unidirectional straight road vehicle networking cognitive spectrum allocation mechanism based on a clustering structure, comprising the following steps: judging network load states, and defining the network load states into three types: light load, heavy load, and super-heavy load; When the network load state is overloaded or super-heavy, the cognitive spectrum mechanism is activated. First, the joint level proportional allocation algorithm based on distance and congestion index is used to allocate spectrum between clusters. The level of each cognitive cell and all cognitive cells are used. The ratio of the sum of the total levels to allocate the idle spectrum provided by authorized users; then use the bidding auction spectrum allocation algorithm based on message priority to allocate spectrum within the cluster, and introduce the priority of the message into the bidding function according to the priority of the user's transmission message. The level coefficient solves the problem of lack of spectrum resources in the specific IoV scenario of one-way straight road, and the spectrum revenue and spectrum utilization rate are high, reducing traffic congestion.

Description

A Cognitive Spectrum Allocation Method for Unidirectional Straight-Road Vehicle Networking Based on Clustering Structure

technical field

The invention belongs to the technical field of vehicle networking cognitive spectrum allocation, and relates to a unidirectional straight road vehicle networking cognitive spectrum allocation method based on a clustering structure.

Background technique

The Internet of Vehicles is one of the most popular research fields of intelligent transportation systems. It is a new type of self-organizing network composed of sensing, computing, storage and wireless communication units on the road. The Internet of Vehicles uses advanced information and communication technology to play an important role in solving various traffic problems. Information sharing and timely transmission are the main goals of the Internet of Vehicles. In recent years, with the development of the Internet of Vehicles, while enjoying the convenience of car travel, people have put forward higher requirements for vehicle services. It is hoped that vehicles can improve road safety and increase in-vehicle entertainment. At the same time, with the development of the economy, the number of cars Rapid growth. According to Gartner, 2.5 billion vehicles will be able to use the Internet of Vehicles in 2020. It can be seen that the amount of Internet of Vehicles data will increase exponentially.

Data transmission in the Internet of Vehicles mainly uses wireless communication technology, and wireless spectrum resources are the premise of wireless communication. At present, the spectrum allocation in my country adopts a fixed allocation method. Different application fields have their own fixed communication frequency bands. The dedicated spectrum for Internet of Vehicles communication (FCC allocation). 5.9GHz frequency band) is becoming increasingly tight, and it can no longer meet the demand for Internet of Vehicles communication. Some research institutions use advanced wireless communication technologies such as link adaptation technology, multi-antenna and multi-user detection, orthogonal frequency division multiplexing, and time division multiplexing to solve the problem of lack of spectrum resources. spectrum utilization, but the spectrum resources themselves are limited, and these methods are far from solving the current spectrum scarcity problem. According to the investigation report of the Federal Communications Commission (FCC), the use of existing spectrum resources is very unbalanced, some frequency bands are occupied for a long time, while other frequency bands are idle for a long time, resulting in the shortage of spectrum resources. waste. Therefore, people have proposed different schemes from the perspective of spectrum sharing, such as the unlicensed frequency band Industrial Scientific and Medial (ISM). Although this scheme can realize spectrum sharing, it will increase system interference, and it is only applicable to fixed frequency bands, which is not suitable for wireless communication systems. capacity and flexibility are limited. Ultra-Wide Band (UItra-Wide Band, UWB) technology can realize spectrum sharing with narrow-band communication systems, but it lacks the flexibility of networking. Therefore, people continue to explore and find that Cognitive Radio (CR) technology has the function of intelligent learning, which can learn in the environment through intelligent technology, reliably perceive the surrounding spectrum environment, and detect legitimate authorized users. Change parameters (such as transmission power, carrier frequency and modulation technology, etc.), and flexibly utilize spectrum holes to achieve reliable communication on the premise of ensuring no interference to authorized users. Therefore, in recent years, many scholars have applied CR to the Internet of Vehicles to solve the problem of spectrum scarcity, resulting in the emergence of the Cognitive Internet of Vehicles (CIoV).

Cognitive spectrum allocation technology is one of the key technologies in cognitive car networking. At present, the research on cognitive car networking is in the initial stage. There are very few studies on cognitive spectrum allocation in car networking, but the comparison of spectrum allocation technology in cognitive wireless network Mature. Spectrum allocation in cognitive wireless network can be divided into: based on graph theory, based on spectrum trading, based on auction bidding and based on game theory. The scheme based on graph theory is to model the spectrum allocation problem as a process of graph coloring, in which a simple graph represents a cognitive wireless network, a vertex represents a cognitive user, and a color represents a spectrum. If there is interference between two cognitive users, Then the two points are connected, and the two cognitive users cannot share the same spectrum at this time. Although the method based on graph theory is simple and feasible, it is mostly suitable for static network environment. If the network topology changes and needs to be re-allocated, the time required for spectrum allocation is determined by the number of cognitive users, the number of idle channels and the dynamics of the network. As long as the network topology changes, it is necessary to reconstruct the network graph. If the network changes frequently, this poses a great challenge to the allocation algorithm. The spectrum allocation scheme based on spectrum trading regards spectrum as a commodity. Cognitive users and authorized users conduct spectrum transactions, and authorized users allocate idle spectrum to cognitive users according to different transaction rules to realize spectrum sharing. Cognitive spectrum allocation method based on spectrum transaction Authorized users can obtain certain economic benefits by leasing their idle spectrum, which can motivate more authorized users to share their own spectrum with cognitive users, thereby improving spectrum utilization. The auction-based cognitive spectrum allocation method is the auction of spectrum, which is the same as the auction principle in microeconomics, in which the base station or the central control access point acts as the auctioneer, and the cognitive user acts as the spectrum buyer to bid. The basic principles are: " Fair and open auction, the highest bidder wins", that is, the buyer with the highest bid wins the temporary right to use the idle spectrum. Auction-based cognitive spectrum allocation method, although the system can enable authorized users to obtain economic benefits to promote the enthusiasm of authorized users to share their idle spectrum, but auction most focuses on the fairness of cognitive users, and at the same time achieves efficiency and effectiveness , will complicate the problem. The cognitive spectrum allocation method based on game theory is the game process of the spectrum strategy selection of different cognitive users, in which the behavior set of game theory is the different schemes for cognitive users to select different frequency bands from the idle spectrum, and the total system revenue is all cognitive users. The most important thing in the distribution method based on game theory is whether the Nash equilibrium exists or not. The cognitive spectrum allocation method based on game theory can effectively improve the efficiency and fairness of spectrum allocation in the CR environment. It has good flexibility and scalability, and is suitable for dynamic network environments.

Few of the existing cognitive spectrum allocation use the Internet of Vehicles as the research scenario, and none of the existing cognitive spectrum allocation technologies can achieve reasonable allocation of spectrum resources, resulting in the lack of spectrum resources and the average spectrum allocation algorithm in the event of traffic congestion. Emergency messages cannot be transmitted in time, especially on one-way straight roads, safety messages cannot be transmitted in time, traffic congestion is difficult to guide, and continuous rear-end collisions are prone to occur, which will further aggravate traffic congestion and bring great inconvenience to people's safe travel.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a one-way straight road car networking cognitive spectrum allocation method based on a clustering structure, which solves the problem that the cognitive spectrum allocation technology existing in the prior art cannot achieve reasonable allocation of spectrum resources, resulting in lack of spectrum resources and occurrence of The problem that emergency messages cannot be transmitted in time during traffic jams.

The technical solution adopted in the present invention is a method for allocating a cognitive spectrum for one-way straight road IoV based on a clustering structure, comprising the following steps:

Step 1. According to the total channel capacity that the current network can provide and the minimum rate requirements of various services, determine the network load state, and define the network load state into three types: light load, heavy load, and super heavy load;

Step 2. When the network load state reaches the overload or super-heavy load in Step 1, the cognitive spectrum mechanism is activated. First, a joint level-proportional allocation algorithm based on distance and congestion index is used to allocate spectrum between clusters, and each cognitive cell is used. The proportional relationship between the level of all cognitive cells and the sum of the total levels of all cognitive cells, allocate the idle spectrum provided by authorized users;

Step 3. Then use the bidding auction spectrum allocation algorithm based on message priority to perform the intra-cluster spectrum allocation of each cluster head after the allocation in step 2, and introduce the priority coefficient of the message into the bidding function according to the priority of the user transmission message to realize the spectrum. Allocation optimization.

The feature of the present invention also lies in:

The calculation formula of the network load state in step 1 is:

为当前网络安全的最小速率需求；

为当前网络非安全业务的最小速率需求；R_total为当前网络可以提供的总信道容量。where NLS is the network load state;

The minimum rate requirement for current network security;

is the minimum rate requirement for non-secure services on the current network; R _total is the total channel capacity that the current network can provide.

The value range corresponding to each network load state is shown in Table 1:

Table 1 Network Load Status

In step 2, the specific steps of using the joint level proportional allocation algorithm based on distance and congestion index to allocate spectrum between clusters are as follows:

Step 2.1. Division of cognitive cells and selection of cluster heads

When the network load state reaches the heavy load or super heavy load in step 1, the cognitive spectrum mechanism is activated. First, the vehicle closest to the center of each cognitive cell is selected as the cluster head, and the cognitive nodes in each cell are at the cluster head. After the node is determined, it reports the number of channels it needs to the cluster head node;

Step 2.2, priority determination

The multi-cognitive cell adopts a joint level proportional allocation algorithm based on distance and congestion index. Therefore, it is necessary to obtain the distance Dpc _i between each cluster head node and the main user and the congestion coefficient TPI _i of each cognitive cell, that is, the road section. The priority of the cognitive cell is The formula for calculating the level is:

Tc _i =[a*TPI _i +b*Dpc _i ] (2)

where Tci is the priority of the _i -th cognitive cell; TPI _i is the congestion coefficient of the i-th cognitive cell; Dpc _i is the distance between the i-th cluster head node and the main user, and each node can be obtained through GPS and then obtain the Dpc _i value through the distance between the two points; a and b are constants, a represents the influence of the congestion state of the cell on the priority, and b represents the influence of the distance between the cluster head node and the main user on the priority. The priority values of cognitive cells are shown in Table 2:

Table 2 Cognitive cell priorities

Step 2.3, spectrum allocation

Using the rank proportional allocation algorithm, the spectrum number W _i that each cluster head gets from the main user is:

为认知小区的优先级总和。In the formula, W provides the total number of idle channels shared by the main user, and W can be obtained by judging whether the channel is idle through energy detection; Tc _i is the priority of the i-th cognitive cell;

is the sum of priorities of cognitive cells.

In step 3, the specific steps of using the bidding and auction spectrum allocation algorithm based on message priority to allocate spectrum within the cluster are as follows:

Step 3.1. The cluster head node announces the idle channel vector P=(P ₁ , P ₂ ,...,P _j ) and the channel floor price vector d=(d ₁ ,d ₂ ,...,d _j );

The floor price d _j of each idle channel given by the cluster head node is:

d _j =A+α _j (4)

where A is the channel rental cost; α _j is the intensity of channel j bidding in the auction;

Step 3.2, the cognitive user perceives the idle channel and gives the channel evaluation vector v _i ={v _i1 ,v _i2 ,...,v _ij };

The evaluation function of the cognitive user on the channel is:

v _ij =β _i γ(B _ij ) (5)

where v _ij represents the evaluation of the channel j by cognitive user i; β _i is the priority coefficient of the message transmitted by the cognitive user i; γ(B _ij ) is the spectral utility function obtained by the cognitive user using the channel after rounding. the bid function;

The message priority coefficient β _i reflects the importance of the information. The more important the message is, the larger the value of β _i is, and its value range is 0 < β _i <1 _. 3 shows:

Table 3 Value relationship between message type and β _i

Step 3.3, calculate the difference v _ij -d _j between the channel evaluation and the reserve price to obtain the difference matrix D;

Step 3.4, sort each row of the difference matrix D (that is, the difference between the evaluation of the same channel and the floor price) according to the size;

Step 3.5: Allocate the channel to the cognitive user with the largest difference between the channel estimate and the floor price:

The channel allocation matrix M reflects the situation that each cognitive user is allocated to the channel, M={b _ij |b _ij ∈ {0,1}, bij indicates whether the channel j is allocated to the user i, if bij=1 means that the channel j is allocated to the User i, otherwise b _ij =0; to prevent interference, each channel can only be assigned to one cognitive user, and two cognitive users cannot share a channel, that is, ∑ _i b _ij ≤1, ∑ _j b _ij ≤1;

Step 3.6. Cognitive users assigned to the channel launch bidding, and users who are not assigned to the channel continue to evaluate, and go back to step 3.3;

The new valuation for cognitive users is:

v _ij ′=v _ij +G (7)

where v _ij ′ is the new evaluation of channel j by cognitive user i; v _ij is the initial evaluation of channel j by cognitive user i; G is the subsidy function, and the specific subsidy conditions of the subsidy function G are:

In the formula, △α _i is the number of times that cognitive user i has not been allocated spectrum continuously in spectrum allocation; g is the subsidy factor; a is the subsidy threshold value, which is a constant; When the frequency of spectrum allocation exceeds the subsidy threshold, that is, Δα _i ≥ a, cognitive user i can obtain subsidy;

Step 3.7: Determine whether the channel allocation is over, until the maximum value of the difference between the channel evaluation and the reserve price is ≤ 0, and the allocation is over.

In step 3.2, the bidding function γ(B _ij ) is obtained by rounding the spectral utility function B _ij obtained by the cognitive user, and B _ij is:

where W is the channel bandwidth, in Hz; S is the signal power; N is the noise power.

The subsidy threshold a=3 in step 3.6.

The beneficial effects of the present invention are as follows: the spectrum allocation between clusters of the present invention adopts a joint-level proportional spectrum allocation algorithm based on distance and congestion index, and allocates idle channels of primary users to relatively congested cognitive cells to ensure safety in the congested cells when traffic is congested. Priority transmission of critical messages reduces traffic congestion; intra-cluster spectrum allocation adopts a bidding and auction spectrum allocation algorithm based on message priority, and allocates idle channels allocated to cognitive cells to cognitive users transmitting safety-critical messages, ensuring safety The priority transmission of key messages improves the average benefit and spectrum utilization rate of the cognitive cell; the present invention solves the problem that the cognitive spectrum allocation technology existing in the prior art cannot achieve reasonable allocation of spectrum resources, resulting in the one-way straight road, which is a specific vehicle. In the networking scenario, due to the lack of spectrum resources and the problem that emergency messages cannot be transmitted in time when traffic congestion occurs, the "spectrum hole" sensed by cognitive radio technology is divided into channels with orthogonal equal bandwidths, and the idle channels are preferentially allocated to congested and distant mains. Cognitive cells that are close to users. In the cognitive cell, the spectrum is preferentially allocated to users transmitting safety-critical messages. The number of channels allocated by this mechanism is not much different from the actual channel requirements of cognitive users, and the spectrum income and spectrum utilization rate are high. , to ensure the priority transmission of safety-critical messages in the Internet of Vehicles, reducing traffic congestion.

Description of drawings

Fig. 1 is a kind of structure schematic diagram of inter-cluster allocation in the one-way straight road vehicle networking cognitive spectrum allocation method based on clustering structure of the present invention;

2 is a flowchart of an allocation algorithm based on message priority bidding and auction in a one-way straight road vehicle networking cognitive spectrum allocation method based on clustering structure of the present invention;

Fig. 3 is the flow chart in a kind of one-way straight road vehicle networking cognitive spectrum allocation method based on clustering structure of the present invention;

Fig. 4 is a kind of spectrum number comparison diagram of the one-way straight road car networking cognitive spectrum allocation method based on clustering structure of the present invention and other different allocation algorithms;

FIG. 5 is a comparison diagram of the average income of cognitive cells under different numbers of idle channels for a one-way straight road vehicle networking cognitive spectrum allocation method based on a clustering structure of the present invention;

FIG. 6 is a comparison diagram of fairness between a method for cognitive spectrum allocation based on a clustering structure for one-way straight road IoV according to the present invention under different numbers of idle channels and different allocation algorithms.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The one-way straight-way vehicle networking cognitive spectrum allocation mechanism based on the clustering structure proposed by the present invention assumes that all vehicles on the road are equipped with the CR function, and adopts the dynamic spectrum access technology of overlay, that is, the primary user and the cognitive user use the same frequency spectrum at different times. A segment of spectrum resources, but the primary user has an absolute advantage in using this frequency band. As long as the primary user uses this frequency band, the cognitive user must withdraw unconditionally. It is assumed that the spectrum detection technology can accurately perceive the idle spectrum within a certain range of cognitive users, and can obtain the regularity of the spectrum used by the main user, and the time when the main user uses the spectrum can be predicted; in addition, it is assumed that the traffic congestion index of the road is obtained. Has good real-time and accuracy.

A method for allocating cognitive spectrum for one-way straight road vehicle networking based on a clustering structure of the present invention comprises the following steps:

Step 1. According to the total channel capacity that the current network can provide and the minimum rate requirements of various services, determine the network load state, and define the network load state into three types: light load, heavy load, and super heavy load;

Step 2. When the network load state reaches the overload or super-heavy load in Step 1, the cognitive spectrum mechanism is activated. First, a joint level-proportional allocation algorithm based on distance and congestion index is used to allocate spectrum between clusters, and each cognitive cell is used. The proportional relationship between the level of all cognitive cells and the sum of the total levels of all cognitive cells, allocate the idle spectrum provided by the primary user;

Step 3. Then use the bidding auction spectrum allocation algorithm based on message priority to perform the intra-cluster spectrum allocation of each cluster head after the allocation in step 2, and introduce the priority coefficient of the message into the bidding function according to the priority of the user transmission message to realize the spectrum. Allocation optimization.

The calculation formula of the network load state in step 1 is:

为当前网络安全的最小速率需求；

为当前网络非安全业务的最小速率需求；R_total为当前网络可以提供的总信道容量。where NLS is the network load state;

It is the minimum rate requirement for current network security;

is the minimum rate requirement for non-secure services on the current network; R _total is the total channel capacity that the current network can provide.

The value range corresponding to each network load state is shown in Table 1:

Table 1 Network Load Status

In step 2, the specific steps of using the joint level proportional allocation algorithm based on distance and congestion index to allocate spectrum between clusters are as follows:

Step 2.1. Division of cognitive cells and selection of cluster heads

When the network load state reaches the heavy load or super heavy load in step 1, the cognitive spectrum mechanism is activated. First, the vehicle closest to the center of each cognitive cell is selected as the cluster head. As shown in Figure 1, it is a one-way straight vehicle The inter-cluster allocation diagram of the networked cognitive spectrum allocation mechanism, the length of each cell is L, and the cognitive nodes in each cell report the number of channels they need to the cluster head node after the cluster head node is determined;

Step 2.2, priority determination

The multi-cognitive cell adopts the joint level proportional allocation method based on distance and congestion index. Therefore, it is necessary to obtain the distance Dpc _i between each cluster head node and the main user and the congestion coefficient TPI _i of each cognitive cell, that is, the road section. The priority of the cognitive cell is The formula for calculating the level is:

Tc _i =[a*TPI _i +b*Dpc _i ] (2)

where Tci is the priority of the _i -th cognitive cell; TPI _i is the congestion coefficient of the i-th cognitive cell; Dpc _i is the distance between the i-th cluster head node and the main user, and each node can be obtained through GPS and then obtain the Dpc _i value through the distance between the two points; a and b are constants, a represents the influence of the congestion state of the cell on the priority, and b represents the influence of the distance between the cluster head node and the main user on the priority. The priority values of cognitive cells are shown in Table 2:

Table 2 Cognitive cell priorities

Step 2.3, spectrum allocation

Using the rank proportional allocation algorithm, the spectrum number W _i that each cluster head gets from the main user is:

为认知小区的优先级总和。In the formula, W provides the total number of idle channels shared by the main user; Tc _i is the priority of the i-th cognitive cell;

is the sum of priorities of cognitive cells.

Step 3: Spectrum allocation within a cluster, that is, the spectrum in a single cognitive cell is equivalent to the spectrum allocation problem of a mobile centralized network. The cluster head node is equivalent to the seller, and the cognitive node in the cluster is equivalent to the buyer. The application scenarios include in-vehicle navigation and positioning, cooperative road condition monitoring, traffic control, road congestion analysis, accident warning, mobile office, business promotion, leisure and entertainment, etc., so its information can be roughly divided into safety-critical, traffic efficiency, business entertainment information. Different types of information have different degrees of urgency for spectrum needs. Among them, cognitive users who transmit safety-critical messages have the most urgent need for spectrum, so the final transaction price of spectrum is also higher.

The channel auction idea of closed price is adopted, that is, each cognitive user does not know each other's valuation of a certain channel, nor does he know the reserve price given by the cluster head user for each channel, and cognitive users can only use the channel according to their own benefits. and previous rounds of bidding to adjust the price. The cluster head user allocates the channel according to the evaluation of the channel by each cognitive user. If the cognitive user is not allocated to the channel, the cognitive user re-bids, and the next round of auction is carried out until it is allocated to the channel. If the estimates are lower than the channel reserve price, the auction ends.

As shown in Figure 2, in step 3, the specific steps of using the bidding and auction spectrum allocation algorithm based on message priority to allocate spectrum within the cluster are as follows:

Step 3.1. The cluster head node announces the idle channel vector P=(P ₁ , P ₂ ,...,P _j ) and the channel floor price vector d=(d ₁ ,d ₂ ,...,d _j );

The floor price d _j of each idle channel given by the cluster head node is:

d _j =A+α _j (4)

where A is the channel rental cost; α _j is the intensity of channel j bidding in the auction;

Step 3.2. Cognitive users perceive idle channels and give a channel evaluation vector v _i ={v _i1 ,v _i2 ,...,v _ij }; the evaluation of each cognitive user for the same channel is different, and the evaluation can reflect The benefits obtained by cognitive users using a certain channel, the higher the benefits obtained by cognitive users using a certain channel, the higher the evaluation of the channel; ensuring traffic safety and efficiency is the primary task of the Internet of Vehicles, so in the evaluation The function introduces the priority coefficient of information. The higher the priority of the message, the larger the coefficient, and the higher the channel evaluation. The cognitive users who ensure the transmission of traffic safety and efficiency messages should be assigned to the channel first;

The evaluation function of the cognitive user on the channel is:

v _ij =β _i γ(B _ij ) (5)

where v _ij represents the evaluation of the channel j by cognitive user i; β _i is the priority coefficient of the message transmitted by the cognitive user i; γ(B _ij ) is the spectral utility function obtained by the cognitive user using the channel after rounding. the bid function;

The message priority coefficient β _i reflects the importance of the information. The more important the message is, the larger the value of β _i is, and its value range is 0 < β _i <1 _. 3 shows:

Table 3 Value relationship between message type and β _i

Step 3.3, calculate the difference v _ij -d _j between the channel evaluation and the reserve price to obtain the difference matrix D;

Step 3.4, sort each row of the difference matrix D (that is, the difference between the evaluation of the same channel and the floor price) according to the size;

Step 3.5: Allocate the channel to the cognitive user with the largest difference between the channel estimate and the floor price:

The channel allocation matrix M reflects the situation that each cognitive user is allocated to the channel, M={b _ij |b _ij ∈ {0,1}, bij indicates whether the channel j is allocated to the user i, if bij=1 means that the channel j is allocated to the User i, otherwise b _ij =0; to prevent interference, each channel can only be assigned to one cognitive user, and two cognitive users cannot share a channel, that is, ∑ _i b _ij ≤1, ∑ _j b _ij ≤1;

Step 3.6. Cognitive users assigned to the channel launch bidding, and users who are not assigned to the channel go back to step 3.3 to continue the evaluation;

The new valuation for cognitive users is:

v _ij ′=v _ij +G (7)

where v _ij ′ is the new evaluation of the cognitive user i on the channel j; v _ij is the initial evaluation of the channel j by the cognitive user i; G is the subsidy function, the main purpose of the subsidy function G is to ensure the fairness of the allocation algorithm, Low-efficiency cognitive users can never be allocated channels, improve their competitiveness, and fulfill their communication needs. The specific subsidy conditions of subsidy function G are:

In the formula, △α _i is the number of times that cognitive user i has not been allocated spectrum continuously in spectrum allocation; g is the subsidy factor, which is valued according to the actual situation; a is the subsidy threshold value, which is a constant; When the number of consecutive unallocated spectrums in spectrum allocation exceeds the subsidy threshold, that is, Δα _i ≥ a, cognitive user i can obtain subsidy;

Step 3.7: Determine whether the channel allocation is over, until the maximum value of the difference between the channel evaluation and the reserve price is ≤ 0, the allocation is over, and the spectrum allocation is optimized.

In step 3.2, the bidding function γ(B _ij ) is obtained by rounding the spectral utility function B _ij obtained by the cognitive user, and B _ij is:

where W is the channel bandwidth, in Hz; S is the signal power; N is the noise power.

The subsidy threshold a=3 in step 3.6.

Through the one-way straight-way vehicle networking cognitive spectrum allocation mechanism proposed by the present invention, the idle spectrum in the wireless network in a certain period of time can be effectively utilized, the total spectrum utilization rate can be provided, and the vehicle networking can be guaranteed to be overloaded or overloaded. (ie, traffic congestion), the priority transmission of traffic safety messages reduces the accident rate, and reduces personal and economic losses. Figure 3 shows the specific implementation process of the cognitive spectrum allocation mechanism for the IoV.

Example

In order to verify the correctness and feasibility of the mechanism, MATLAB simulation tool is used to simulate and analyze the inter-cluster and intra-cluster allocation algorithms in the cognitive spectrum allocation mechanism of the one-way straight road vehicle networking proposed by the present invention, assuming the simulation environment of the present invention There is one primary user, three cluster head nodes, and I _i ordinary nodes in the i-th cluster.

Simulation parameters: as shown in Table 4.

Table 4 Simulation parameters

(1) Inter-cluster allocation

As shown in Figure 4, when the number of idle channels is 30, the frequency spectrum obtained by the primary user and the cluster head node in different cognitive cells under the two allocation schemes of the joint level proportional allocation algorithm based on distance and congestion index and the average allocation scheme The quantity is compared with the actual demand. It can be seen from Figure 4 that the joint level proportional allocation algorithm based on distance and congestion index is adopted, and the number of spectrums obtained by cluster head nodes with different priorities is different. The condition of the large demand for spectrum in the traffic-congested cell, while ensuring the phenomenon that the spectrum is wasted due to the long distance. In addition, the channels allocated by the joint-level proportional allocation algorithm based on distance and congestion index are not much different from the actual channel requirements, making full use of spectrum resources; while under the average allocation scheme, the channels obtained by each cell are the same, and the channels allocated by each cell are the same as the actual channel requirements. The actual demand is very different, resulting in a waste of resources.

(2) In-cluster allocation

This stage simulates and verifies the bidding and auction spectrum allocation algorithm based on message priority. In this stage, the average benefit index APro and fairness index Fair are introduced to evaluate the bidding and auction spectrum allocation algorithm based on message priority in the cluster. APro will allocate all the idle spectrum at the cluster head. The ratio of the total benefit obtained after completion to the total number of cognitive nodes in the cluster. The larger APro, the greater the system benefit obtained by the cognitive cell; Fair indicates the fairness when each cognitive node in the cluster is allocated to the spectrum channel , the value range of Fair (0,1), the closer the value is to 1, the better the fairness. The average benefit index APro and the fairness index Fair are defined as follows:

In the formula, APro is the average benefit of the system; I is the number of cognitive users in the cluster; b _ij represents whether to assign channel j to user i, if b _ij =1, it means that channel j is assigned to user i; B _ij is your cognitive The spectral utility function obtained by the user;

In the formula, Fair is the fairness index; β _i is the type of data transmitted by the cognitive user i; b _ij represents whether the channel j is allocated to the user i, if b _ij =1, the channel _j is allocated to the user i; Know the spectral utility function obtained by the user.

In order to verify the effectiveness of the algorithm proposed in the cluster, the simulation and comparison of the bidding auction spectrum allocation algorithm based on message priority and the fast auction allocation algorithm are carried out under different subsidy thresholds.

As shown in Figure 5, the number of idle channels takes different values, the spectrum allocation algorithm of the bidding auction based on message priority and the fast auction allocation algorithm allocate the average benefit of cognitive cells. It can be seen from the figure that with the increase of the number of idle channels, the average benefit in the cognitive cell also increases, but the average benefit of the IoV cognitive spectrum allocation algorithm based on the bidding auction based on message priority is greater than that under the three subsidy thresholds. Fast auction allocation algorithm. The subsidy threshold is different, and the average benefit obtained in the cognitive cell is different. When the threshold is 3, the average benefit is greater than the subsidy threshold, which is 0 and 5; if the subsidy threshold is too low, such as a=0, at this time Too many users with low information level are allocated to the spectrum, which reduces the average benefit of the cognitive cell. When the subsidy threshold is too high, such as a=5, many users cannot be allocated channels, and the channel utilization rate is low, which reduces the cognitive cell. average benefit.

Figure 6 shows the fairness comparison between the spectrum allocation algorithm of the bidding auction based on message priority and the fast auction allocation algorithm when the number of idle channels takes different values. It can be seen from the figure that the number of idle channels takes different values, and the fairness index of the fast auction algorithm is Fair = 1, because the fast auction algorithm uses fairness as the allocation index to ensure the same benefits among each cognitive user, but this allocation algorithm cognitive cell The total benefit and average benefit are both low; the fairness index Fair of the bidding auction spectrum allocation algorithm based on message priority changes with the change of the number of idle channels, and the subsidy threshold is different, the fairness index Fair is also different, when the subsidy threshold a = When it is 0 or 5, the fairness index Fair of the bidding and auction spectrum allocation algorithm based on message priority is relatively low. When the limit a=0 or 5, the fairness index Fair value of the bidding and auction spectrum allocation algorithm based on message priority fluctuates around 1, and the cognitive users are basically balanced.

Claims (2)

1. A unidirectional straight-way internet of vehicles cognitive spectrum allocation method based on a clustering structure is characterized by comprising the following steps:

step 1, judging the network load state according to the total channel capacity which can be provided by the current network and the minimum speed requirement of various services, and defining the network load state into three types: light load, heavy load, and overload;

the calculation formula of the network load state is as follows:

NLS is network load state;

the minimum rate requirement of the current network security service is met;

the minimum rate requirement of the current network non-safety service is met; r_totalTotal channel capacity available for the current network;

the value ranges corresponding to the load states of the networks are shown in table 1:

TABLE 1 network load status

Step 2, when the network load state reaches the heavy load or the overload in the step 1, starting a cognitive spectrum mechanism, firstly, performing inter-cluster spectrum allocation by adopting a joint grade proportion allocation algorithm based on distance and congestion indexes, and allocating an idle spectrum provided by a main user by utilizing the proportion relation between the grade of each cognitive cell and the sum of the total grades of all cognitive cells;

the specific steps of the inter-cluster spectrum allocation by adopting a joint grade proportion allocation algorithm based on the distance and the congestion index are as follows:

step 2.1, cognitive cell division and cluster head selection

When the network load state reaches the heavy load or the overload in the step 1, starting a cognitive spectrum mechanism, firstly, selecting a vehicle closest to the central position of each cognitive cell as a cluster head, and reporting the required channel number to the cluster head node by the cognitive node in each cell after the cluster head node is determined;

step 2.2, determination of priority

The multi-cognitive cell adopts a joint grade proportion distribution algorithm based on distance and congestion index, so that the distance Dpc between each cluster head node and the master user needs to be acquired_iAnd congestion coefficient TPI of each cognitive cell or road section_iThe calculation formula of the priority of the cognitive cell is as follows:

Tc_i＝[a*TPI_i+b*Dpc_i] (2)

tc in the formula_iThe priority of the ith cognitive cell; TPI_iThe congestion coefficient of the ith cognitive cell is the congestion coefficient of the ith cognitive cell; dpc_iThe distance between the ith cluster head node and the master user is obtained; a. b is a constant, a represents the influence of the congestion state of the cell on the priority, b represents the influence of the distance between the cluster head node and the master user on the priority, and the priority value of each cognitive cell is shown in table 2:

table 2 cognitive cell priority

Constant a Constant b Cognitive cell TPI_i Distance Dpc between cluster head and master user_i(km) Cognitive cell priority Tc_i 0.8 0.2 8.2 6 8 0.8 0.2 4.5 5 5 0.8 0.2 2.4 4 3 0.5 0.5 8.2 6 7 0.5 0.5 4.5 5 5 0.5 0.5 2.4 4 3 0.3 0.7 8.2 6 7 0.3 0.7 4.5 5 5 0.3 0.7 2.4 4 4

Step 2.3, Spectrum Allocation

The frequency spectrum W of each cluster head divided from the main user by adopting a grade proportion distribution algorithm_iComprises the following steps:

where W provides shared idle channel for primary userTotal number; tc_iThe priority of the ith cognitive cell;

is the sum of the priorities of the cognitive cells;

step 3, then, a bidding auction frequency spectrum allocation algorithm based on message priority is adopted to perform intra-cluster frequency spectrum allocation of each cluster head after allocation in step 2, and a priority coefficient of the message is introduced into a bidding function according to the priority of the message transmitted by the user to realize optimization of the frequency spectrum allocation;

the specific steps of using the bidding auction spectrum allocation algorithm based on the message priority to allocate the intra-cluster spectrum are as follows:

step 3.1, the cluster head node publishes the idle channel vector P ═ (P)₁,P₂,…,P_j) And the base value vector d ═ of the channel (d ═ d)₁,d₂,…,d_j)；

Wherein the base price d of each idle channel given by the cluster head node_jComprises the following steps:

d_j＝A+α_j (4)

wherein A is the channel rental cost; alpha is alpha_jThe severity of the auction for channel j in the auction;

step 3.2, the cognitive user perceives the idle channel and provides a channel estimation vector v_i＝{v_i1,v_i2,…,v_ij}；

The cognitive user's valuation function for the channel is:

v_ij＝β_iγ(B_ij) (5)

in the formula v_ijRepresenting the evaluation of the cognitive user i on the channel j; beta is a_iTransmitting a message priority coefficient for a cognitive user i; gamma (B)_ij) Obtaining a bid function after rounding a spectrum utility function obtained by using a channel for a cognitive user; the bid function γ (B)_ij) Is a spectrum utility function B obtained by a cognitive user_ijObtained by rounding conversion, B_ijComprises the following steps:

wherein W is the channel bandwidth in Hz; s is signal power; n is noise power;

message priority coefficient beta_iReflecting the importance of the message, the more important the message is_iThe larger the value is, the range of the value is 0<β_iBeta less than or equal to 1, cognitive user transmits different types of information_iThe value correspondence of (a) is shown in table 3:

TABLE 3 message types and beta_iValue relationship of

Step 3.3, calculating the difference v between the channel valuation and the base value_ij-d_jObtaining a difference matrix D;

step 3.4, sorting each row of the difference matrix D, namely the same channel estimation and base price difference according to the size;

and 3.5, allocating the channel to the cognitive user with the maximum difference value between the channel estimation value and the base value:

the channel allocation matrix M reflects the situation that each cognitive user is allocated to a channel, and M ═ b_ij|b_ij∈{0,1}}，b_ijIndicating whether channel j is assigned to user i, if b_ij1 means that channel j is allocated to user i, otherwise b_ij0; to prevent interference, each channel can only be allocated to one cognitive user, and two cognitive users cannot share one channel, namely sigma_ib_ij≤1，∑_jb_ij≤1；

Step 3.6, the cognitive users distributed to the channels quit bidding, the users not distributed to the channels continue to evaluate, and the step 3.3 is returned to;

the new valuation of the cognitive user is as follows:

v_ij′＝v_ij+G (7)

in the formula v_ij' is cognitionNew valuation of channel j by user i; v. of_ijInitial valuation of channel j for cognitive user i; g is a subsidy function, and the concrete subsidy conditions of the subsidy function G are as follows:

in the formula, delta alpha_iThe frequency of continuous undistributed frequency spectrums in frequency spectrum allocation is given to the cognitive user i; g is a subsidy factor; a is a subsidy threshold, which is a constant; wherein, only when the frequency of the cognitive user i not continuously dividing into the frequency spectrum in the frequency spectrum allocation exceeds the subsidy threshold value, namely delta alpha_iWhen the user is more than or equal to a, the cognitive user i can obtain subsidies;

and 3.7, judging whether the channel allocation is finished or not until the maximum value of the difference between the evaluation value and the base value of each channel is less than or equal to 0, and finishing the allocation.

2. The method for allocating cognitive spectrum in one-way direct-drive internet of vehicles according to claim 1, wherein the subsidized threshold value a in step 3.6 is 3.

Images