Low-power-consumption RSU deployment method based on genetic algorithm

Abstract

The invention discloses a low-power-consumption RSU deployment method based on a genetic algorithm, which is used in the technical field of Internet of Vehicles, and is used for reducing power consumption by considering the vehicle distribution condition, adaptively adjusting the transmitting power of RSUs and switching between an active mode and a sleep mode, and specifically comprises the following steps of: firstly, determining a communication service area of the RSUs according to the vehicle distribution; modeling an RSU deployment problem into an RSU communication service area combination problem; then, constructing a digital sequence as a genotype individual of a solution, and evaluating the advantages and disadvantages of the RSU communication service area combination through a fitness function; through crossing, variation and iteration processes in the genetic algorithm, the synergistic effect between RSUs is optimized, and the overall energy efficiency is maximized; according to the method, the stability of a communication link between the RSU and the vehicle can be guaranteed, meanwhile, the total power consumption of the RSU in the system is remarkably reduced, and the effectiveness of the method in the aspect of reducing the total power consumption of the RSU is verified through a simulation experiment.

Description

Technical Field

[0001] The present invention relates to a vehicle networking technology, and in particular to a low-power RSU (Roadside Unit) deployment method based on a genetic algorithm. Background Art

[0002] As the commercialization of 5G technology continues to accelerate, the Internet of Things (IoT) has entered a new stage of development. As one of the applications with huge market potential in the field of IoT, the Internet of Vehicles (IoV), its key component, the Road Side Unit (RSU), plays a vital role in providing network coverage services. RSUs are usually deployed along the road and are responsible for V2I (Vehicle-to-Infrastructure) communication with vehicles. However, the uneven distribution of V2I communication requirements in time and space causes many RSUs to be in a low-load or no-load state for a long time. Despite this, RSUs still need to consume a certain amount of basic power consumption to maintain their normal operation, which invisibly causes a large amount of energy waste. This problem has become the main obstacle to RSU achieving full network coverage of roads. In addition, the current full coverage RSU deployment strategy (FCRD) provides a fixed non-overlapping coverage range, which fails to fully consider the rapid changes in the wireless channel, resulting in the inability to effectively provide communication services in non-hotspot areas, further increasing energy consumption.

[0003] In order to reduce the power consumption of RSU, the industry has proposed a working mechanism that enables RSU to switch between active mode and sleep mode. When a vehicle enters the coverage area of ​​​​the RSU, the RSU switches to active mode and establishes a communication link; when the vehicle leaves, the RSU switches to sleep mode to reduce energy consumption. Although this mode switching strategy reduces energy consumption to a certain extent, in order to reduce the transmission power of the RSU while ensuring the stability of the communication link, it is necessary to design a low-power RSU deployment strategy based on the distribution of vehicles. Such a strategy can more effectively adapt to the spatiotemporal changes in V2I communication needs, thereby optimizing energy consumption in different scenarios.

[0004] At present, some studies have been conducted on reducing power consumption by controlling the RSU to switch between active and sleep modes through the vehicle detection system (VDS) in the IoV system. However, this method has the following limitations: First, when the traffic flow is low, the RSU runs longer in sleep mode, which significantly reduces power consumption; but when the traffic flow is high, its energy saving effect is greatly reduced. Second, the VDS method is mainly designed for a single RSU and cannot adapt to the scenario of multiple RSU deployment. At the same time, the independent node RSU deployment (INRD) scheme determines the communication relationship based on the minimum link attenuation between the vehicle and the adjacent RSU, and defines the coverage of the RSU with the farthest communication boundary, which is used to optimize each RSU individually. Although it performs well in single node optimization, it fails to fully utilize the synergy between RSUs in two-dimensional scenarios and ignores the possibility of helping adjacent RSUs enter sleep mode by providing a larger cellular communication service area (CCSA) through a few RSUs, which may lead to unnecessary redundant power consumption.

[0005] Therefore, it is very necessary to study a low-power RSU deployment strategy that can reduce the power consumption of a single RSU while ensuring the stability of the communication link between the RSU and the vehicle, and is suitable for scenarios with multiple RSUs, and optimizes the synergy between RSUs to maximize the overall energy efficiency. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a low-power RSU deployment method based on genetic algorithm, which can adapt to the temporal and spatial changes of vehicle distribution, adaptively adjust the transmission power required by RSU, optimize the synergy between RSUs, switch the working mode of RSU, and achieve the goal of reducing the total power consumption of RSU while ensuring the stability of communication link.

[0007] The technical solution adopted by the present invention to solve the above technical problems is: a low-power RSU deployment method based on genetic algorithm, characterized by comprising the following steps:

[0008] Step 1: In the 2-D vehicle networking system, an RSU is deployed at each road intersection, and the initial coverage of a single RSU is set as the maximum coverage of the RSU. It is set that there is an overlapping area on the maximum coverage between adjacent RSUs, and the area on the maximum coverage of a single RSU except all overlapping areas is defined as a non-overlapping area;

[0009] A single RSU defines two types of communication service areas based on vehicle distribution, namely the first type of communication service area and the second type of communication service area;

[0010] For the i-th RSU, if there is no vehicle in the non-overlapping area of ​​the i-th RSU's maximum coverage, the first-class communication service area of ​​the i-th RSU is Empty and use Identification; otherwise, the first type of communication service area of ​​the i-th RSU It contains a circular communication service area, whose center is the position of the i-th RSU, and the radius is determined by calculating the communication link attenuation of the signal from the i-th RSU to each vehicle traveling in the non-overlapping area of ​​the maximum coverage range of the i-th RSU, and then taking the distance between the vehicle corresponding to the maximum communication link attenuation and the position of the i-th RSU as the radius. This communication service area uses Identifier, where i=1, 2, ..., N, N represents the total number of RSUs, N>2;

[0011] For the i-th RSU, if there are no vehicles in all overlapping areas on the maximum coverage of the i-th RSU, the second type of communication service area of ​​the i-th RSU is is empty and not marked; otherwise, the second type of communication service area of ​​the i-th RSU Contains Num i There are circular communication service areas, the center of any communication service area is the position of the i-th RSU, and the radius is the distance between the position of the i-th RSU and any vehicle in any overlapping area on the maximum coverage of the i-th RSU. i Each communication service area uses Identification, where Num i ≥1;

[0012] Step 2: Traverse each RSU in order. For the i-th RSU currently traversed, obtain the identification set S corresponding to the i-th RSU according to the first communication service area and the second communication service area of ​​the i-th RSU. i , if the first type of communication service area and the second type of communication service area of ​​the i-th RSU are both empty, then In this case, the i-th RSU can enter sleep mode; if the first type of communication service area of ​​the i-th RSU is empty and the second type of communication service area contains Num i communication service area, then If the first type of communication service area of ​​the i-th RSU contains a communication service area and the second type of communication service area is empty, then If the first type of communication service area of ​​the i-th RSU contains the communication service area and the second type of communication service area contains Num i communication service area, then

[0013] Step 3: Let iter represent the number of rounds and initialize iter = 0; let iter max Indicates the global maximum number of rounds; S 1 ,…,S i ,…,S N Input into the genetic algorithm to create the initial population of the genetic algorithm. The initial population contains M genotype individuals. Each genotype individual contains N genes, and each gene is an identifier, where M>2;

[0014] Step 4: In the iter round, select the genotype individuals with the largest and second largest fitness values ​​from the population in the iter round as the two parent individuals of the iter generation, denoted as r and p1 and r p2 ; Among them, when iter=0, the population in the iter round is the initial population;

[0015] Step 5: In the iter round, generate a random number rand1, rand1∈[0,1]; if rand1<p Cross ,but and r p2 Generate two offspring individuals of the iter generation, denoted as and r c2 , and take these two offspring individuals as the individuals to be corrected, and then execute step 6; if rand1≥p Cross ,but and r p2 Do not generate two offspring individuals of the iter generation, and and r p2 As the individual to be corrected, execute step 6 again; where p Cross represents the exchange probability;

[0016] Step 6: In the iter round, for each gene in each individual to be corrected, generate a random number rand2, rand2∈[0,1], if rand2<p Mut , then modify this gene to any one of the identifiers in the identifier set to which the identifier represented by this gene belongs; if rand2≥p Mut , then this gene is not modified; after all genes in each individual to be corrected are processed, the corrected individual is obtained, and then step 7 is executed; where p Mut represents the probability of mutation;

[0017] Step 7: In the iter round, for each corrected individual, determine whether the fitness value of this corrected individual is greater than The fitness value or r p2 If the fitness value is yes, then add the corrected individual to the population of the iter round as the population of the iter+1 round; otherwise, do not process the corrected individual; when both corrected individuals are added to the population of the iter round or one corrected individual is added to the population of the iter round, execute step 8; when both corrected individuals are not added to the population of the iter round, return to step 5 and continue to execute;

[0018] Step 8: Let iter = iter + 1; then determine whether iter is equal to iter max If yes, execute step 9; otherwise, return to step 5 to continue execution; where "=" in iter = iter + 1 is an assignment symbol;

[0019] Step 9: Find the iter max The genotype individual with the largest fitness value in the population of the round;

[0020] Then, each gene in the genotype individual is traversed in turn. For the currently traversed gene, according to the identifier represented by this gene, the RSU corresponding to the identifier set to which the identifier represented by this gene belongs is used as the current RSU, and the working mode of the current RSU is switched. If the identifier represented by this gene corresponds to a no-communication service area, the working mode of the current RSU is switched to the sleep mode, and the power consumption of the current RSU is the given RSU sleep mode power consumption P sleep ; If the identification represented by this gene corresponds to a communication service area, then the working mode of the current RSU is switched to the active mode, and the power consumption of the current RSU is P t By formula P t =P 0 +P t x Calculate, P 0 represents the minimum non-zero transmit power of RSU, P 0 >P sleep , P t x Indicates the transmission power required by the current RSU when forming the communication service area corresponding to the identity represented by this gene, P r It represents the minimum receiving power of the vehicle under the premise of ensuring service quality, h x Indicates the attenuation of the communication link between the signal from the current RSU to the vehicle corresponding to the communication service area corresponding to the identification represented by this gene;

[0021] After all genes in this genotype individual are traversed, the power consumption of all RSUs is obtained; then the total power consumption of RSUs in the system is calculated, which is the sum of the power consumption of all RSUs.

[0022] In step 3, the process of creating the initial population is as follows: randomly select an identifier from the identifier set corresponding to each RSU; then use the N identifiers randomly selected for the N RSUs as a genotype individual in the genetic algorithm; after performing M random selection operations in total, M genotype individuals are obtained, and the M genotype individuals constitute the initial population in the genetic algorithm.

[0023] In step 4, the fitness value of a genotype individual is the sum of the fitness values ​​of all genes in the genotype individual. The fitness value F of a gene is obtained as follows: if the identifier represented by the gene corresponds to a non-communication service area, then F = 0; if the identifier represented by the gene corresponds to a communication service area, then F is obtained by the formula F = γN a +βN v -ωh calculation, where h represents the attenuation of the communication link between the RSU belonging to the communication service area corresponding to the identification represented by this gene and the vehicle corresponding to the communication service area corresponding to the identification represented by this gene, when there is no overlapping area between the communication service area corresponding to the identification represented by this gene and the maximum coverage range of all adjacent RSUs. a and N v When all are 0, the communication service area corresponding to the identifier represented by this gene overlaps with the maximum coverage of all adjacent RSUs. a represents the number of overlapping areas where vehicles exist, N v N a The number of vehicles in the overlapping area where there are vehicles, γ is N a The weight coefficient, β is N v is the weight coefficient of , and ω is the weight coefficient of h.

[0024] In step 5, and r c2 The generation process is: Two intersection points are randomly selected from the a and cr b ; then Medium cr a The genes represented by b The gene represented by a With cr b All genes between p2 The genes at the same position in the and r c2 .

[0025] In step 9, P r =N 0 ·B·r th , where N 0 represents the power spectral density of Gaussian white noise received by the vehicle, B represents the bandwidth of the cellular downlink channel, r th Indicates the signal-to-noise ratio threshold.

[0026] In step 9, h x =PL x ·ψ 0 ·ψ h , where PL x It represents the path loss between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene. f represents the transmission carrier frequency of the signal, d 0 The far-field reference distance of the antenna on the RSU, d x represents the distance between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene, α represents the path loss index, ψ 0 represents shadow fading, which is a log-normally distributed random variable with a logarithmic mean of 0 and a standard deviation of σ s , ψ h Represents Rayleigh fading, which is a random variable with logarithmic Rayleigh distribution, whose logarithm has a mean of 0 and a standard deviation of σ h .

[0027] Compared with the prior art, the advantages of the present invention are:

[0028] 1) The method of the present invention can adaptively adjust the transmission power required by the RSU according to the different vehicle distributions, thereby solving the energy waste problem of the RSU in a low-load or no-load state for a long time.

[0029] 2) The method of the present invention switches the working mode of the RSU to reduce the overall power consumption of the RSU in the system, while ensuring the stability of the communication link between the RSU and the vehicle, solving the problem of low communication efficiency of the full coverage RSU deployment strategy in non-hotspot areas.

[0030] 3) The method of the present invention effectively conducts a holistic analysis of the RSUs in the system, and optimizes the synergy between RSUs. That is, in the case of small communication link attenuation, a few RSUs can help adjacent RSUs enter sleep mode by providing a larger communication service area, thereby maximizing the overall energy efficiency. This solves the problem of redundant power consumption caused by the failure to fully utilize the synergy between RSUs in the prior art.

[0031] 4) The method of the present invention provides a flexible RSU deployment strategy that can adapt to different traffic flows and vehicle distribution conditions and has good adaptability and application range. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flowchart of the overall implementation process of the method of the present invention;

[0033] Figure 2 is a 2-D vehicle networking system model diagram in the method of the present invention;

[0034] Figure 3 A schematic diagram showing the comparison of the total power consumption of RSUs in the system of the method of the present invention (e-GA) with the theoretical value (Theoretical), the existing deployment method based on the greedy algorithm (Greedy), the independent node deployment method (INRD) and the full coverage deployment method (FCRD) as the number of vehicles increases;

[0035] Figure 4 CDF diagram of total power consumption of RSU in the system using the method of the present invention (e-GA) and the theoretical value (Theoretical);

[0036] Figure 5 The CDF diagram of the total power consumption of RSU in the system is the existing deployment method based on the greedy algorithm (Greedy) and the theoretical value (Theoretical). DETAILED DESCRIPTION

[0037] The present invention is further described in detail below with reference to the accompanying drawings.

[0038] The present invention proposes a low-power RSU deployment method based on genetic algorithm, and its overall implementation flow chart is as follows: Figure 1 As shown, it includes the following steps:

[0039] Step 1: In a 2-D connected vehicle system, Figure 2 As shown, it is assumed that one RSU is deployed at each road intersection, and the initial coverage range of a single RSU is set to be the maximum coverage range of the RSU. It is assumed that there is an overlapping area on the maximum coverage range between adjacent RSUs, and the area on the maximum coverage range of a single RSU except all overlapping areas is defined as a non-overlapping area.

[0040] A single RSU defines two types of communication service areas according to vehicle distribution, namely the first type of communication service area and the second type of communication service area.

[0041] For the i-th RSU, if there is no vehicle in the non-overlapping area of ​​the i-th RSU's maximum coverage, the first-class communication service area of ​​the i-th RSU is Empty and use Identification; otherwise, the first type of communication service area of ​​the i-th RSU It contains a circular communication service area, whose center is the position of the i-th RSU, and the radius is determined by calculating the communication link attenuation of the signal from the i-th RSU to each vehicle traveling in the non-overlapping area of ​​the maximum coverage range of the i-th RSU, and then taking the distance between the vehicle corresponding to the maximum communication link attenuation and the position of the i-th RSU as the radius. This communication service area uses Identifier, where i=1,2,…,N, N represents the total number of RSUs, and N>2.

[0042] For the i-th RSU, if there are no vehicles in all overlapping areas on the maximum coverage of the i-th RSU, the second type of communication service area of ​​the i-th RSU is is empty and not marked; otherwise, the second type of communication service area of ​​the i-th RSU Contains Num i There are circular communication service areas, the center of any communication service area is the position of the i-th RSU, and the radius is the distance between the position of the i-th RSU and any vehicle in any overlapping area on the maximum coverage of the i-th RSU. i Each communication service area uses Identification, where Num i ≥1.

[0043] exist Figure 2 In the example, an RSU (roadside unit) is deployed at each of the four road intersections. The maximum coverage of the RSU is as follows: Figure 2As shown in the middle gray area, there are overlapping areas in the maximum coverage range between adjacent RSUs. There is a vehicle in the non-overlapping area of ​​the maximum coverage range of the upper left RSU, and the communication service area included in the first type of communication service area of ​​this RSU is directly determined by the distance between the vehicle and the location of this RSU; there are no vehicles in all overlapping areas of the maximum coverage range of this RSU, and the second type of communication service area of ​​this RSU is empty. In general, this RSU has only one communication service area. There are no vehicles in the non-overlapping area of ​​the maximum coverage range of the upper right RSU, and the first type of communication service area of ​​this RSU is empty; there is a vehicle in all overlapping areas of the maximum coverage range of this RSU, that is, there is one corresponding communication service area. In general, this RSU has only one communication service area. There are four vehicles in the non-overlapping area of ​​the maximum coverage range of the lower left RSU, and the communication service area included in the first type of communication service area of ​​this RSU is determined by the distance between the vehicle corresponding to the maximum communication link attenuation and the location of this RSU; there is a vehicle in all overlapping areas of the maximum coverage range of this RSU, that is, there is one corresponding communication service area. In general, this RSU has two communication service areas. There are four vehicles in the non-overlapping area of ​​the maximum coverage of the RSU on the lower right. The communication service area included in the first-class communication service area of ​​this RSU is determined by the distance between the vehicle corresponding to the maximum communication link attenuation and the position of this RSU; there are two vehicles in all overlapping areas of the maximum coverage of this RSU, that is, there are two corresponding communication service areas. In general, this RSU has three communication service areas.

[0044] Step 2: Traverse each RSU in order. For the i-th RSU currently traversed, obtain the identification set S corresponding to the i-th RSU according to the first communication service area and the second communication service area of ​​the i-th RSU. i , if the first type of communication service area and the second type of communication service area of ​​the i-th RSU are both empty, then In this case, the i-th RSU can enter sleep mode; if the first type of communication service area of ​​the i-th RSU is empty and the second type of communication service area contains Num i communication service area, then If the first type of communication service area of ​​the i-th RSU contains a communication service area and the second type of communication service area is empty, then If the first type of communication service area of ​​the i-th RSU contains the communication service area and the second type of communication service area contains Num i communication service area, then

[0045] Step 3: Let iter represent the number of rounds and initialize iter = 0; let iter maxIndicates the global maximum number of rounds; S 1 ,…,S i ,…,S N Input into the genetic algorithm to create the initial population of the genetic algorithm. The initial population contains M genotype individuals, each genotype individual contains N genes, and each gene is an identifier, where M>2.

[0046] In this specific embodiment, potential solutions to the RSU deployment problem include the communication service areas of all RSUs, which constitute the phenotype of the genetic algorithm. In the process of creating the initial population, the genotype individuals are formed by randomly selecting identifiers, which increases the diversity of RSU deployment strategies and helps to find a better RSU deployment solution. This method can better adapt to different traffic flows and vehicle distribution conditions, and improves the flexibility and adaptability of RSU deployment compared to the prior art. The process of creating the initial population is as follows: randomly select an identifier from the identifier set corresponding to each RSU; then use the N randomly selected identifiers for the N RSUs as a genotype individual in the genetic algorithm; after performing M random selection operations, M genotype individuals are obtained, and the M genotype individuals constitute the initial population in the genetic algorithm.

[0047] Step 4: In the iter round, select the genotype individuals with the largest and second largest fitness values ​​from the population in the iter round as the two parent individuals of the iter generation, denoted as r and p1 and r p2 ; Among them, when iter=0, the population in the iter round is the initial population.

[0048] In this specific embodiment, in step 4, the fitness value of a genotype individual is the sum of the fitness values ​​of all genes in the genotype individual. The fitness value F of a gene is obtained as follows: if the identifier represented by the gene corresponds to a non-communication service area, then F = 0; if the identifier represented by the gene corresponds to a communication service area, then F is obtained by the formula F = γN a +βN v -ωh calculation, where h represents the attenuation of the communication link between the RSU belonging to the communication service area corresponding to the identification represented by this gene and the vehicle corresponding to the communication service area corresponding to the identification represented by this gene, when there is no overlapping area between the communication service area corresponding to the identification represented by this gene and the maximum coverage range of all adjacent RSUs. a and N v When all are 0, there is an overlap between the communication service area corresponding to the identifier represented by this gene and the maximum coverage of all adjacent RSUs (all or part of them have overlaps). a represents the number of overlapping areas where vehicles exist, Nv N a The number of vehicles in the overlapping area where there are vehicles, γ is N a The weight coefficient, β is N v is the weight coefficient of , ω is the weight coefficient of h, γ+β+ω=1.

[0049] By defining the fitness function, a balance is achieved between the communication link attenuation and the size of the communication service area. In the case of small communication link attenuation, a few RSUs can help adjacent RSUs enter sleep mode by providing a larger communication service area, thereby reducing the total power consumption of RSUs in the 2-D Internet of Vehicles system. Compared with the existing technology, this method more effectively optimizes energy consumption.

[0050] Step 5: In the iter round, generate a random number rand1, rand1∈[0,1]; if rand1<p Cross , then r p1 and r p2 Generate two offspring individuals of the iter generation, denoted as and r c2 , and take these two offspring individuals as the individuals to be corrected, and then execute step 6; if rand1≥p Cross , then r p1 and r p2 Do not generate two offspring individuals of the iter generation, and r p1 and r p2 As the individual to be corrected, execute step 6 again; where p Cross represents the exchange probability.

[0051] In this specific embodiment, and r c2 The generation process is: (or r p2 ) randomly selects two intersection points, namely cr a and cr b ; then (or r p2 ) a The genes represented by b The gene represented by a With cr b All genes between p2 (or r p1 ) are crossed to form and r c2 In the process of generating offspring individuals, the search space of the genetic algorithm is increased through the selection of crossover points and the crossover of genes, which helps to find a better solution. This method improves the search efficiency of the algorithm and the quality of the solution.

[0052] Step 6: In the iter round, for each gene in each individual to be corrected, generate a random number rand2, rand2∈[0,1], if rand2<p Mut , then modify this gene to any one of the identifiers in the identifier set to which the identifier represented by this gene belongs; if rand2≥p Mut , then this gene is not modified; after all genes in each individual to be corrected are processed, the corrected individual is obtained, and then step 7 is executed; where p Mut represents the probability of mutation.

[0053] Step 7: In the iter round, for each corrected individual, determine whether the fitness value of this corrected individual is greater than r p1 The fitness value or r p2 fitness value, if yes, then add this corrected individual to the population of the iter round as the population of the iter+1 round; otherwise, do not process this corrected individual; when both corrected individuals are added to the population of the iter round or one corrected individual is added to the population of the iter round, execute step 8; when both corrected individuals are not added to the population of the iter round, return to step 5 and continue execution.

[0054] Here, the fitness value of the corrected individual, r p1 The fitness value, r p2 The fitness values ​​of are calculated in the same way as the fitness value of a genotype individual in step 4.

[0055] Step 8: Let iter = iter + 1; then determine whether iter is equal to iter max If yes, execute step 9; otherwise, return to step 5 to continue execution; where the "=" in iter=iter+1 is the assignment symbol.

[0056] Step 9: Find the iter max The genotype individual with the largest fitness value in the population of the round.

[0057] Then, each gene in the genotype individual is traversed in turn. For the currently traversed gene, according to the identifier represented by this gene, the RSU corresponding to the identifier set to which the identifier represented by this gene belongs is used as the current RSU, and the working mode of the current RSU is switched. If the identifier represented by this gene corresponds to a no-communication service area, the working mode of the current RSU is switched to the sleep mode, and the power consumption of the current RSU is the given RSU sleep mode power consumption P sleep; If the identification represented by this gene corresponds to a communication service area, then the working mode of the current RSU is switched to the active mode, and the power consumption of the current RSU is P t By formula P t =P 0 +P t x Calculate, P 0 represents the minimum non-zero transmit power of RSU, P 0 >P sleep , P t x Indicates the transmission power required by the current RSU when forming the communication service area corresponding to the identity represented by this gene, P r It represents the minimum receiving power of the vehicle under the premise of ensuring service quality, h x Indicates the attenuation of the communication link from the current RSU to the vehicle corresponding to the communication service area corresponding to the identification represented by this gene.

[0058] After all genes in this genotype individual are traversed, the power consumption of all RSUs is obtained; then the total power consumption of RSUs in the system is calculated, which is the sum of the power consumption of all RSUs.

[0059] Here, the fitness value is calculated in the same way as the fitness value of an individual of a genotype in step 4.

[0060] In this specific embodiment, P r =N 0 ·B·r th , where N 0 represents the power spectral density of Gaussian white noise received by the vehicle, B represents the bandwidth of the cellular downlink channel, r th In order to maintain the stability of the communication link between the current RSU and the vehicle corresponding to the communication service area corresponding to the identification represented by this gene, the signal-to-noise ratio (SNR) of the current RSU must be greater than the signal-to-noise ratio threshold r th ,Right now Among them, SNR represents the signal-to-noise ratio of the current RSU, P' r Indicates the received power of the vehicle corresponding to the communication service area corresponding to the identification represented by this gene.

[0061] In this specific embodiment, h x =PL x ·ψ 0 ·ψ h , where PL x Indicates the path loss between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene, PL xIt depends on the location of the vehicle corresponding to the communication service area corresponding to the identification represented by this gene. f represents the transmission carrier frequency of the signal, d 0 The far-field reference distance of the antenna on the RSU, d x represents the distance between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene, α represents the path loss index, ψ 0 represents shadow fading, which is a log-normally distributed random variable with a logarithmic mean of 0 and a standard deviation of σ s , ψ h Represents Rayleigh fading, which is a random variable with logarithmic Rayleigh distribution, whose logarithm has a mean of 0 and a standard deviation of σ h .

[0062] By considering factors such as path loss, shadow fading, and Rayleigh fading, the required transmit power of the RSU is calculated, so that the RSU can adjust the transmit power according to the actual communication environment, ensuring communication quality while reducing power consumption. Compared with the existing technology, this method more accurately reflects the actual communication environment and improves the reliability and energy efficiency of communication.

[0063] In order to further verify the feasibility and effectiveness of the method of the present invention, a simulation experiment was carried out on the method of the present invention.

[0064] In the simulation experiment, the simulation parameters are set as listed in Table 1.

[0065] Table 1 Simulation parameter settings

[0066]

[0067] In the simulation experiment, the existing methods involved in the performance comparison include the existing deployment method based on the greedy algorithm (Greedy), the independent node deployment method (INRD) and the full coverage deployment method (FCRD). Among them, the Greedy method is recorded in the literature Yang, H.; Jia, Z.; Xie, G. Delay-Bounded and Cost-Limited RSU Deployment in Urban Vehicular Ad Hoc Networks. Sensors 2018, 18, 2764.

[0068] Figure 3 The total power consumption of RSU in the system under different numbers of vehicles is demonstrated by the proposed method (e-GA) and the existing deployment method based on greedy algorithm (Greedy), independent node deployment method (INRD) and full coverage deployment method (FCRD). Figure 3It can be seen that the gap between the total power consumption of RSU in the system of the FCRD method and the theoretical power consumption (Theoretical) shows a trend of gradual reduction, because the increase in the number of communication service areas leads to a decrease in the number of RSUs that can enter sleep mode. In addition, compared with the total power consumption of RSU in the system of the INRD method that considers each RSU independently, considering the RSU in the system as a whole can significantly reduce the total power consumption. In particular, when the number of vehicles increases, the total power consumption of RSU in the system of the INRD method exceeds that of the FCRD method, reflecting that the increase in redundant coverage caused by the randomness of communication link attenuation leads to increased power consumption. The method of the present invention (e-GA) shows a total power consumption level close to the theoretical power consumption (Theoretical), and when the number of vehicles is small, it is close to the theoretical power consumption. As the number of vehicles increases, the gap between the remaining theoretical values ​​is also significantly smaller than that of other algorithms. At the same time, by comparing the method of the present invention (e-GA) with the INRD method, it can be found that the method of the present invention (e-GA) has a certain inhibitory effect on the aggravation of the random attenuation of the communication link as the number of vehicles increases. It can be concluded that the method of the present invention (e-GA) is superior to other algorithms and is applicable to situations with different numbers of vehicles, indicating that the optimal solution can be approached with lower complexity without adding additional redundancy while ensuring the stability of the communication link.

[0069] Figure 4 The CDF diagram of the total power consumption of RSU in the system and the theoretical power consumption (Theoretical) of the method of the present invention (e-GA) is given. Figure 5 The CDF graph of the total power consumption of RSU in the system and the theoretical power consumption (Theoretical) of the existing deployment method based on the greedy algorithm (Greedy) is given. Figure 4 and Figure 5 It can be seen that compared with the Greedy method, the deviation between the total power consumption of RSU in the system of the method of the present invention and the theoretical power consumption is smaller.

Claims

1. A low-power RSU deployment method based on genetic algorithm, characterized in that The following steps are involved: Step 1: In the 2-D vehicle networking system, an RSU is deployed at each road intersection, and the initial coverage of a single RSU is set as the maximum coverage of the RSU. It is set that there is an overlapping area on the maximum coverage between adjacent RSUs, and the area on the maximum coverage of a single RSU except all overlapping areas is defined as a non-overlapping area; A single RSU defines two types of communication service areas based on vehicle distribution, namely the first type of communication service area and the second type of communication service area; For the i-th RSU, if there is no vehicle in the non-overlapping area of ​​the i-th RSU's maximum coverage, the first-class communication service area of ​​the i-th RSU is Empty and use Identification; otherwise, the first type of communication service area of ​​the i-th RSU It contains a circular communication service area, whose center is the position of the i-th RSU, and the radius is determined by calculating the communication link attenuation of the signal from the i-th RSU to each vehicle traveling in the non-overlapping area of ​​the maximum coverage range of the i-th RSU, and then taking the distance between the vehicle corresponding to the maximum communication link attenuation and the position of the i-th RSU as the radius. This communication service area uses Identifier, where i=1, 2, ..., N, N represents the total number of RSUs, N>2; For the i-th RSU, if there are no vehicles in all overlapping areas on the maximum coverage of the i-th RSU, the second type of communication service area of ​​the i-th RSU is is empty and not marked; otherwise, the second type of communication service area of ​​the i-th RSU Contains Num i There are circular communication service areas, the center of any communication service area is the position of the i-th RSU, and the radius is the distance between the position of the i-th RSU and any vehicle in any overlapping area on the maximum coverage of the i-th RSU. i Each communication service area uses Identification, where Num i ≥1; Step 2: Traverse each RSU in order. For the i-th RSU currently traversed, obtain the identification set S corresponding to the i-th RSU according to the first communication service area and the second communication service area of ​​the i-th RSU. i , if the first type of communication service area and the second type of communication service area of ​​the i-th RSU are both empty, then In this case, the i-th RSU can enter sleep mode; if the first type of communication service area of ​​the i-th RSU is empty and the second type of communication service area contains Num i communication service area, then If the first type of communication service area of ​​the i-th RSU contains a communication service area and the second type of communication service area is empty, then If the first type of communication service area of ​​the i-th RSU contains the communication service area and the second type of communication service area contains Num i communication service area, then Step 3: Let iter represent the number of rounds and initialize iter = 0; let iter max Indicates the global maximum number of rounds; S 1 ,…,S i ,…,S N Input into the genetic algorithm to create the initial population of the genetic algorithm. The initial population contains M genotype individuals. Each genotype individual contains N genes, and each gene is an identifier, where M>2; Step 4: In the iter round, select the genotype individuals with the largest and second largest fitness values ​​from the population in the iter round as the two parent individuals of the iter generation, denoted as and r p2 ; Among them, when iter=0, the population in the iter round is the initial population; Step 5: In the iter round, generate a random number rand1, rand1∈[0,1]; if rand1<p Cross ,but and r p2 Generate two offspring individuals of the iter generation, denoted as and r c2 , and take these two offspring individuals as the individuals to be corrected, and then execute step 6; if rand1≥p Cross ,but and r p2 Do not generate two offspring individuals of the iter generation, and and r p2 As the individual to be corrected, execute step 6 again; where p Cross represents the exchange probability; Step 6: In the iter round, for each gene in each individual to be corrected, generate a random number rand2, rand2∈[0,1], if rand2<p Mut , then modify this gene to any one of the identifiers in the identifier set to which the identifier represented by this gene belongs; if rand2≥p Mut , then this gene is not modified; after all genes in each individual to be corrected are processed, the corrected individual is obtained, and then step 7 is executed; where p Mut represents the probability of mutation; Step 7: In the iter round, for each corrected individual, determine whether the fitness value of this corrected individual is greater than The fitness value or r p2 If the fitness value is yes, then add the corrected individual to the population of the iter round as the population of the iter+1 round; otherwise, do not process the corrected individual; when both corrected individuals are added to the population of the iter round or one corrected individual is added to the population of the iter round, execute step 8; when both corrected individuals are not added to the population of the iter round, return to step 5 and continue to execute; Step 8: Let iter = iter + 1; then determine whether iter is equal to iter max If yes, execute step 9; otherwise, return to step 5 to continue execution; where "=" in iter=iter+1 is an assignment symbol; Step 9: Find the iter max The genotype individual with the largest fitness value in the population of the round; Then, each gene in the genotype individual is traversed in turn. For the currently traversed gene, according to the identifier represented by this gene, the RSU corresponding to the identifier set to which the identifier represented by this gene belongs is used as the current RSU, and the working mode of the current RSU is switched. If the identifier represented by this gene corresponds to a no-communication service area, the working mode of the current RSU is switched to the sleep mode, and the power consumption of the current RSU is the given RSU sleep mode power consumption P sleep ; If the identification represented by this gene corresponds to a communication service area, then the working mode of the current RSU is switched to the active mode, and the power consumption of the current RSU is P t By formula P t =P0+P t x Calculation, P0 represents the minimum non-zero transmit power of RSU, P0>P sleep , P t x Indicates the transmission power required by the current RSU when forming the communication service area corresponding to the identity represented by this gene, P r It represents the minimum receiving power of the vehicle under the premise of ensuring service quality, h x Indicates the attenuation of the communication link between the signal from the current RSU to the vehicle corresponding to the communication service area corresponding to the identification represented by this gene; After all genes in this genotype individual are traversed, the power consumption of all RSUs is obtained; then the total power consumption of RSUs in the system is calculated, which is the sum of the power consumption of all RSUs.

2. According to claim 1, a low-power RSU deployment method based on genetic algorithm is characterized in that In step 3, the process of creating the initial population is as follows: randomly select an identifier from the identifier set corresponding to each RSU; then use the N identifiers randomly selected for the N RSUs as a genotype individual in the genetic algorithm; after performing M random selection operations in total, M genotype individuals are obtained, and the M genotype individuals constitute the initial population in the genetic algorithm.

3. According to claim 1, a low-power RSU deployment method based on genetic algorithm is characterized in that In step 4, the fitness value of a genotype individual is the sum of the fitness values ​​of all genes in the genotype individual. The fitness value F of a gene is obtained as follows: if the identifier represented by the gene corresponds to a non-communication service area, then F = 0; if the identifier represented by the gene corresponds to a communication service area, then F is obtained by the formula F = γN a +βN v -ωh calculation, where h represents the attenuation of the communication link between the RSU belonging to the communication service area corresponding to the identification represented by this gene and the vehicle corresponding to the communication service area corresponding to the identification represented by this gene, when there is no overlapping area between the communication service area corresponding to the identification represented by this gene and the maximum coverage range of all adjacent RSUs. a and N v When all are 0, the communication service area corresponding to the identifier represented by this gene overlaps with the maximum coverage of all adjacent RSUs. a represents the number of overlapping areas where vehicles exist, N v N a The number of vehicles in the overlapping area where there are vehicles, γ is N a The weight coefficient, β is N v is the weight coefficient of , and ω is the weight coefficient of h.

4. According to claim 1, a low-power RSU deployment method based on genetic algorithm is characterized in that In step 5, r c1 and r c2 The generation process is: p1 Two intersection points are randomly selected from the a and cr b ; Then r p1 Medium cr a The genes represented by b The gene represented by a With cr b All genes between p2 The genes at the same position in the and r c2 .

5. According to claim 1, a low-power RSU deployment method based on genetic algorithm is characterized in that In step 9, P r =N0·B·r th , where N0 represents the power spectrum density of Gaussian white noise received by the vehicle, B represents the bandwidth of the cellular downlink channel, and r th Indicates the signal-to-noise ratio threshold.

6. A low-power RSU deployment method based on genetic algorithm according to claim 1 or 5, characterized in that In step 9, h x =PL x ·ψ0·ψ h , where PL x It represents the path loss between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene. f represents the transmission carrier frequency of the signal, d0 represents the far-field reference distance of the antenna on the RSU, and d x represents the distance between the current RSU and the vehicle corresponding to the communication service area corresponding to the identifier represented by this gene, α represents the path loss index, ψ0 represents shadow fading, which is a random variable with a log-normal distribution, and its logarithm has a mean of 0 and a standard deviation of σ s , ψ h Represents Rayleigh fading, which is a random variable with logarithmic Rayleigh distribution, whose logarithm has a mean of 0 and a standard deviation of σ h .

Images