CN115314944A - Internet of vehicles cooperative caching method based on mobile vehicle social relation perception - Google Patents
Internet of vehicles cooperative caching method based on mobile vehicle social relation perception Download PDFInfo
- Publication number
- CN115314944A CN115314944A CN202210955357.XA CN202210955357A CN115314944A CN 115314944 A CN115314944 A CN 115314944A CN 202210955357 A CN202210955357 A CN 202210955357A CN 115314944 A CN115314944 A CN 115314944A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- content
- vehicles
- buffer
- indicating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/08—Load balancing or load distribution
- H04W28/09—Management thereof
- H04W28/0925—Management thereof using policies
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
- G08G1/012—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from other sources than vehicle or roadside beacons, e.g. mobile networks
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/056—Detecting movement of traffic to be counted or controlled with provision for distinguishing direction of travel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/02—Services making use of location information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses an Internet of vehicles cooperative caching method based on social relation perception of mobile vehicles, which is characterized by predicting a moving track of each vehicle in an Internet of vehicles, determining a roadside unit to which the vehicle belongs at the next moment to obtain a vehicle set in the coverage range of each roadside unit, calculating the request probability prediction of the vehicle in the coverage range of the roadside unit for each content to obtain a content set requested by each vehicle at the next moment, calculating the social relation strength of each vehicle, screening out cached vehicles, calculating the delay of each vehicle for obtaining the content from the cached vehicles and the roadside units, and finally constructing a cooperative caching optimization problem and solving to obtain a cooperative caching scheme under each roadside unit. The method and the system predict and obtain the roadside unit and the vehicle request content of the next moment, select the cache vehicle according to the social relation of the vehicles, and accordingly formulate a correct cooperative cache strategy, improve the cache hit rate and reduce the content acquisition delay.
Description
Technical Field
The invention belongs to the technical field of Internet of vehicles, and particularly relates to an Internet of vehicles collaborative caching method based on mobile vehicle social relation perception.
Background
Along with the development of vehicle network, vehicle-mounted applications such as vehicle-mounted intelligent terminal, automatic driving, vehicle entertainment and safety are constantly emerging, positive influence is brought for the development of vehicle roads, traffic efficiency is improved, and user's driving experience is promoted. However, the advent of these intelligent applications may require significant computational, storage, and communication resources, causing severe traffic load on the core network. The mobile edge cache technology pre-caches popular contents in roadside units and partial vehicles, and when a requesting user requests the contents, the contents are acquired in a wireless transmission mode without a remote core network, so that the delay of content acquisition is reduced, the content cache hit rate is improved, and the experience quality of the user is improved. However, the performance of the cache is poor considering the space limitation of the edge device, the randomness of the movement of the vehicle, the time-varying nature of the requested content and the complexity of the social relationship.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a vehicle networking cooperative caching method based on social relation perception of a mobile vehicle.
In order to achieve the purpose, the Internet of vehicles cooperative caching method based on the mobile vehicle social relation perception comprises the following steps:
s1: set RSU = { S) of roadside units in Internet of vehicles recording 1 ,…,S n ,…,S N In which S is n N represents the nth roadside unit, N =1,2, …, N represents the number of roadside units;
the requested content set in the Internet of vehicles is X = { X = 1 ,…,x q ,…,x Q },x q Represents the Q content, Q =1,2, …, Q, Q represents the quantity of the content which can be requested in the Internet of vehicles, and the content x is recorded q Has a size of d q ;
S2: for each vehicle in the internet of vehicles, predicting the moving track of the vehicle to obtain the predicted position of the next moment, and determining the roadside unit to which the vehicle belongs at the next moment; obtaining each roadside unit S at the next moment n Set of vehicles V in the coverage area n ={v n,1 ,…,v n,m ,…,v n,M },v n,m Indicating roadside units S n The M-th covered vehicle, M =1,2, …, M n ,M n Indicating roadside units S n Number of vehicles within coverage;
s3: firstly, counting the request times of each content in a preset historical time period, sequencing the contents according to the request times from large to small, and recording the content x q After sorting, the sequence number is i q (ii) a Calculating the request probability of each content in a preset historical time period, taking the content with the request probability exceeding a preset threshold value as popular content, and summing the request probabilities of the popular content to be used as a distribution slope omega of the popular content; calculating the content popularity pop of the q-th content at the next moment by adopting the following formula q :
For roadside units S n Set of vehicles V in the coverage area n Obtaining each vehicle v n,m In the preset calendarRequested content set L within a history period n,m Then, the interest similarity sim between the vehicles is calculated pairwise by adopting the following calculation formula m,m′ :
Wherein phi is m,m′ Indicating a vehicle v n,m And a vehicle v n,m′ Content set L n,m 、L n,m′ M' =1,2, …, M n And | l represents the quantity of contents in the set, M q Indicating that content x has been requested within a predetermined historical period of time q The number of vehicles of (c);
the vehicle v is calculated by adopting the following formula n,m Interest similarity SIM m :
Similarity of interest s m Normalization is carried out to obtain interest similarity SIM 'after normalization' m :
Wherein, the SIM max 、SIM min Respectively representing a set of vehicles V n SIM (subscriber identity module) for interest similarity of all vehicles m Maximum and minimum values of;
the vehicle v is calculated by adopting the following formula n,m Interest probability in for each content n,m :
For content popularity pop q And a vehicle v n,m Interest probability in of n,m,q Carrying out weighted summation to obtain the vehicle v at the next moment n,m For content x q Request probability p of n,m,q :
p n,m,q =α×pop q +β×in n,m
Wherein, alpha and beta respectively represent a preset content popularity weighted value and a vehicle personal interest preference weighted value;
setting the requested content of the vehicle to obey the poisson distribution according to the vehicle v n,m For each content x q Request probability p of n,m,q Predicting a vehicle v n,m Content collections requested at the next time
S4: for roadside units S n Set of vehicles V n Each vehicle v in n,m And counting the number of vehicles which can be associated in unit time during the driving process as a communication contact rate f n,m Fusing interest similarity sim between vehicles m,m′ Calculating the social relationship strength z between the vehicles pairwise by adopting the following formula m,m′ :
z m,m′ =f n,m ×sim m,m′
The vehicle v is calculated by adopting the following formula n,m Social relationship strength Z m :
Will be roadside unit S n Sequencing all vehicles from large to small according to the social relationship strength, selecting the front K vehicles as cache vehicles, determining the value of K according to actual needs, and obtaining a cache vehicle set CV n ={cv n,1 ,…,cv n,k ,…,cv n,K },cv n,k Indicating roadside units S n The kth buffer vehicle from bottom, K =1,2, …, K;
s5: calculating the time delay of different content acquisition modes, wherein the specific method comprises the following steps:
vehicle v n,m Buffering vehicle cv through V2V communication n,k ObtainingContent x q Time delay T of n,m,k,q Comprises the following steps:
wherein, rate n,m,k Indicating a vehicle v n,m And buffer the vehicle cv n,k The communication transmission rate between;
vehicle v n,m From roadside units S by V2R communication n Obtaining content x q Time delay T of n,m,K+1,q Comprises the following steps:
wherein, rate n,m,K+1 Indicating a vehicle v n,m And roadside unit S n The communication transmission rate therebetween;
s6: solving the following cooperative cache optimization problem to obtain each roadside unit S n The following cooperative caching scheme:
C1:a q,k′ ∈{0,1},k′=1,2,…,K+1
wherein, A = (a) q,1 ,…,a q,k ,…,a q,K+1 ) Represents a buffer decision vector, a q,k′ To relieveStore decision identification, K' =1,2, …, K +1,a q,k =1 denotes converting content x q Buffer to buffer vehicle cv n,k ,a q,k =0 means that the content x is not deleted q Buffer to buffer vehicle cv n,k ,a q,K+1 =1 denotes converting content x q Buffer to roadside unit S n ,a q,K+1 =0 indicates that the content x is not written q Buffer to roadside unit S n ,Indicating buffer vehicle cv n,k The buffer capacity of the buffer memory (c),indicating roadside units S n The buffer capacity of (2).
The invention discloses an Internet of vehicles cooperative caching method based on social relation perception of mobile vehicles, which is characterized in that a moving track of each vehicle in an Internet of vehicles is predicted, a roadside unit to which the vehicle belongs at the next moment is determined, a vehicle set in the coverage range of each roadside unit is obtained, the request probability prediction of the vehicle in the coverage range of the roadside unit for each content is calculated, a content set requested by each vehicle at the next moment is obtained, the social relation strength of each vehicle is calculated, a cache vehicle is screened out, the delay of each vehicle for obtaining the content from the cache vehicle and the roadside unit is calculated, and finally, a cooperative caching optimization problem is established and a cooperative caching scheme under each roadside unit is solved.
According to the method, the roadside unit to which the next moment belongs is obtained according to the vehicle movement track prediction, and the request content is predicted based on the content popularity and the vehicle user interest, so that the cache vehicle is selected according to the social relation of the vehicle, and a correct collaborative cache strategy is formulated, thereby improving the cache hit rate and reducing the content acquisition delay.
Drawings
FIG. 1 is a flowchart of an embodiment of a mobile vehicle social relationship awareness-based collaborative caching method for Internet of vehicles according to the present invention;
FIG. 2 is a diagram illustrating the effect of predicting the movement trajectory based on the LSTM network in the present embodiment;
FIG. 3 is a flow chart of reinforcement learning;
FIG. 4 is a graph comparing hit rates for different numbers of vehicles according to the present invention and the comparison method;
FIG. 5 is a comparison of time delays for different vehicle data for the present embodiment of the invention and comparison method;
FIG. 6 is a comparison graph of cache hit rates of the present invention and the comparison method under different total cache capacities in this embodiment;
fig. 7 is a comparison diagram of the time delays of the present invention and the comparison method under different total buffer capacities in this embodiment.
Detailed Description
Specific embodiments of the present invention are described below in conjunction with the accompanying drawings so that those skilled in the art can better understand the present invention. It is to be expressly noted that in the following description, a detailed description of known functions and designs will be omitted when it may obscure the subject matter of the present invention.
Examples
FIG. 1 is a flowchart of an embodiment of a method for collaborative caching in Internet of vehicles based on social relationship awareness of mobile vehicles according to the present invention. As shown in FIG. 1, the method for caching the vehicle networking collaboration based on the social relationship awareness of the mobile vehicle comprises the following specific steps:
s101: acquiring the parameters of the Internet of vehicles:
set RSU = { S) of roadside units in Internet of vehicles recording 1 ,…,S n ,…,S N In which S is n Denotes the nth curb unit, N =1,2, …, N denotes the number of curb units. Generally, the roadside units should be distributed as evenly as possible on both sides of the road, and the coverage areas of different roadside units do not overlap, so as to provide better service for vehicles.
The requested content set in the Internet of vehicles is X = { X = { X = } 1 ,…,x q ,…,x Q },x q Represents the Q content, Q =1,2, …, Q represents the quantity of content which can be requested in the internet of vehicles, and the content x is recorded q Has a size of d q 。
S102: predicting the moving track of the vehicle:
and predicting the moving track of each vehicle in the internet of vehicles to obtain the predicted position of the next moment, so as to determine the roadside unit to which the vehicle belongs at the next moment. Obtaining each roadside unit S at the next moment n Set of vehicles V in the coverage area n ={v n,1 ,…,v n,m ,…,v n,M },v n,m Indicating roadside units S n The M-th covered vehicle, M =1,2, …, M n ,M n Indicating roadside units S n Number of vehicles in the coverage area.
The specific method for predicting the movement track can be selected according to actual needs. The invention carries out the prediction of the moving track based on the LSTM (Long Short-Term Memory) network, and the specific method comprises the following steps:
and sampling historical position vectors of the vehicles according to the historical movement tracks of the vehicles, inputting the historical position vectors into a pre-trained LSTM network, outputting the probability that the vehicles are positioned at each accessible position, and taking the accessible position corresponding to the maximum probability value as the predicted position of the vehicle.
The LSTM network is an improvement of a recurrent neural network, is often used for processing time series data, and can effectively memorize relevant information of a vehicle historical track. Fig. 2 is a diagram illustrating the effect of predicting the movement trajectory based on the LSTM network in the present embodiment. As shown in fig. 2, the predicted position and the real position obtained based on the LSTM network are almost consistent, which shows that the LSTM has a very accurate effect of predicting the vehicle trajectory.
S103: predicting vehicle request content:
in order to make caching decisions of contents better, the probability that the vehicle requests each content at the next moment needs to be determined first, and the vehicle requests the content based on the request probability is predicted so as to determine which contents need to be cached. The present invention integrates the popularity of content and the personal interests of the vehicle user to determine the probability that the vehicle will request the content.
For the content, counting the request times of each content in a preset historical time period, and according to the request times, counting the request times from large to smallSorting the contents, recording the contents x q After sorting, the serial number is i q . Calculating the request probability (namely the ratio of the content request times to the total request times) of each content in a preset historical time period, taking the content with the request probability exceeding a preset threshold value as popular content, and summing the request probabilities of the popular content as the distribution slope omega of the popular content. And finally, calculating the probability that the q content is requested by all vehicles at the next moment by adopting the following formula, namely the content popularity pop q :
The content needs of a vehicle are not only influenced by the content itself, but are also closely related to the personal preferences of the vehicle user. The vehicle preference may be expressed as a probability that a single vehicle requests content at the present time, and thus a future content request probability may be predicted from the personal preference of the vehicle content request history information. For roadside units S n Set of vehicles V in the coverage area n Obtaining each vehicle v n,m Requested content set L within a preset history period n,m Then, the interest similarity sim between the vehicles is calculated pairwise by adopting the following calculation formula m,m′ :
Wherein phi is m,m′ Indicating a vehicle v n,m And a vehicle v n,m′ Content set L n,m 、L n,m′ M' =1,2, …, M, the intersection of (i.e. the set of content requested by both vehicles) n And | l represents the quantity of contents in the set, M q Indicating that content x has been requested within a preset historical period of time q The number of vehicles (a) of (b),for representing content x q The weight of (c).
Is calculated by the following formulaCalculating a vehicle v n,m Interest similarity SIM m :
Similarity of interest s m Normalization is carried out to obtain interest similarity SIM 'after normalization' m :
Wherein, the SIM max 、SIM min Respectively representing a set of vehicles V n The maximum value and the minimum value of the interest similarity of all vehicles.
The vehicle v is calculated by adopting the following formula n,m Interest probability in for each content n,m :
For content popularity pop q And a vehicle v n,m Interest probability in of n,m,q Carrying out weighted summation to obtain the vehicle v at the next moment n,m For content x q Request probability p of n,m,q :
p n,m,q =α×pop q +β×in n,m
Wherein, alpha and beta respectively represent a preset content popularity weighted value and a vehicle personal interest preference weighted value;
setting the content of the vehicle request to comply with a poisson distribution according to the vehicle v n,m For each content x q Request probability p of n,m,q Predicting a vehicle v n,m Content collections requested at the next time
S104: selecting a cached vehicle based on the social relationship:
to roadside unit S n Set of vehicles V n And selecting the vehicles capable of being cached according to the social relationship among the vehicles. Vehicles are said to be socially related when the demand between them is consistent, enabling the establishment of communication links for data transfer. Thus, the social relationship has two features: first, there is a similarity of interest between vehicles; second is the contact rate between vehicles. For roadside units S n Set of vehicles V n Each vehicle v in n,m And counting the number of vehicles which can be associated in unit time during the driving process as a communication contact rate f n,m Fusing interest similarity sim between vehicles m,m′ Calculating the social relationship strength z between the vehicles pairwise by adopting the following formula m,m′ :
z m,m′ =f n,m ×sim m,m x
The vehicle v is calculated by adopting the following formula n,m Social relationship strength Z m :
Will be roadside unit S n Sequencing all vehicles from large to small according to the social relationship strength, selecting the front K vehicles as cache vehicles, determining the value of K according to actual needs, and obtaining a cache vehicle set CV n ={cv n,1 ,…,cv n,k ,…,cv n,K },cv n,k Representing roadside units S n The K-th buffer vehicle from below, K =1,2, …, K.
S105: calculating content acquisition delay:
content acquisition latency is an important performance indicator in cache systems, which may reflect the quality of communication between the requesting user and the cached content. When the delay in obtaining content is large, the content requested by the vehicle needs to be obtained from a remote content provider over a backhaul link and cannot be obtained from a caching vehicle or roadside unit over wireless transmission, resulting in high-delay transmission. Therefore, it is necessary to calculate the time delays of different content acquisition modes and provide a basis for subsequent caching decisions, and the specific method is as follows:
vehicle v n,m Buffering vehicles cv through V2V communication n,k Obtaining content x q Time delay T of n,m,k,q Comprises the following steps:
wherein, rate n,m,k Indicating a vehicle v n,m And buffer the vehicle cv n,k The communication transmission rate between the two can be calculated by adopting the following formula:
where B denotes the transmission bandwidth, σ 2 Representing additive white Gaussian noise, P n,k Indicating buffer vehicle cv n,k Transmit power of h n,m,k Indicating a vehicle v n,m And buffer the vehicle cv n,k Distance between, h n,m″,k Indicating a vehicle v n,m″ And buffer vehicle cv n,k Distance of (e) between m″≠m P n,k h n,m″,k Indicating other vehicles v n,m″ For vehicle v n,m And buffer the vehicle cv n,k Interference of communication between M =1,2, …, M ″ n And m ≠ m.
Vehicle v n,m From roadside units S through V2R communication n Obtaining content x q Time delay T of n,m,K+1,q Comprises the following steps:
wherein, rate n,m,K+1 Indicating a vehicle v n,m And roadside unit S n The communication transmission rate between the two can be calculated by adopting the following formula:
wherein, P n Representing roadside units S n Transmit power of h n,m,K+1 Indicating a vehicle v n,m And roadside unit S n The channel gain in between. Since each roadside unit uses a different frequency band, interference of other roadside units may not be considered.
S106: and (3) collaborative caching decision:
the method aims to minimize content acquisition delay when performing cooperative caching decision, namely, a requesting user acquires the requested content from the network edge equipment as much as possible instead of a remote content provider, so that the load of a backhaul link is reduced, and the content acquisition delay is reduced. Based on the above analysis, the optimization problem of collaborative caching is presented as follows:
C1:a q,k′ ∈{0,1},k′=1,2,…,K+1
wherein, A = (a) q,1 ,…,a q,k ,…,a q,K+1 ) Represents a buffer decision vector, a q,k′ For caching decision identification, K' =1,2, …, K +1,a q,k =1 denotes converting content x q Buffer to buffer vehicle cv n,k ,a q,k =0 means that the content x is not deleted q Buffer-to-buffer vehicle cv n,k ,a q,K+1 =1 denotes that the content x q Buffer to roadside unit S n ,a q,K+1 =0 indicates that the content x is not written q Buffer to roadside unit S n ,Indicating buffer vehicle cv n,k The buffer capacity of the buffer memory (c),indicating roadside units S n The buffer capacity of (2).
The constraint conditions C1 and C2 are limits on the value of the cache decision identifier, and C1 is a value that can cache or not cache contents in the cache vehicle and the roadside unit, and is represented by 0 or 1. To make reasonable use of storage capacity and reduce cache redundancy, C2 is used to ensure that each content can only be cached in one cache vehicle within the roadside unit coverage. C3 and C4 are to ensure that the content cached in the respective caching vehicles and roadside units does not exceed their storage capacity.
And solving the optimization problem to obtain a cache decision vector so as to determine a cooperative caching scheme.
In this embodiment, a reinforcement learning method is used to solve the optimization problem of the cooperative cache, where the set state s is (K +1,K × C) CV +C RSU ) Setting action a to buffer decision vector a = (a) q,1 ,…,a q,k ,…,a q,K+1 ) The reward function R is the delay corresponding to the buffered decision vector, i.e. the delay
Fig. 3 is a flowchart of reinforcement learning. As shown in fig. 3, the specific process of reinforcement learning is as follows:
1) Initializing the current network Q (s, a; theta) and target networkTheta andrepresenting parameters of the current network and the target network, respectively.
2) Set state s to (K +1,K × C CV +C RSU )。
3) Selecting a random action a by probability, or selecting the action with the largest current function value by greedy algorithmAnd executed.
4) And calculating the content acquisition reward R according to the cache decision vector represented by the action a, and acquiring the next state s'.
5) This experience (s, a, R, s') is stored in an experience replay unit from which small batches of samples are drawn for training.
8) Updating next state s' ← s
9) Training is stopped until the network parameter theta converges.
The principles and specific processes of Reinforcement Learning can be found in the literature "Zhu H, cao Y, wang W, et al. Deep recovery Learning for Mobile Edge Learning: review, new Features, and Open esses [ J ]. IEEE Network,2018,32 (6): 50-57.
In order to better illustrate the technical effects of the invention, the invention is experimentally verified by using specific examples. In the experimental verification, two existing Caching strategies are adopted as comparison methods, namely a DAC algorithm and an RC algorithm, wherein the DAC algorithm only considers the social relationship of the vehicle and does not predict the vehicle track, the specific contents are shown in the document 'Zhuo X, li Q, cao G, et al, social-Based collaborative Caching in DTNs' A Contact Duration Approach [ C ].2011IEEE origin International Conference on Mobile Ad-Hoc and Sensor Systems,2011, and the RC algorithm is a random Caching algorithm, namely the roadside unit and the vehicle adopt the same probability random Caching content.
FIG. 4 is a graph comparing hit rates of the present invention and the comparison method for different numbers of vehicles in the present embodiment. As shown in fig. 4, the DAC algorithm does not consider the mobility of the vehicle, and at this time, the cache result obtained by the static network topology cannot adapt to the actual dynamic network topology, resulting in the degradation of the cache performance. The RC algorithm does not take into account social relationships and the likelihood of sharing content between vehicles is reduced, thereby reducing network performance. The present invention combines vehicle mobility and sociability and therefore has better performance than the two comparison methods. When the number of vehicles is 50, the cache hit rate of the method is improved by 14.4% and 24.9% compared with the RC algorithm and the DAC algorithm respectively.
Fig. 5 is a comparison graph of time delays of the present invention and the comparison method under different vehicle data in the present embodiment. As shown in fig. 5, the content buffering delay gradually increases as the number of vehicles increases, but the present invention has a greater advantage in terms of delay. When the social relationship between the vehicles is not considered, the possibility of sharing content between the two vehicles may be reduced, resulting in performance degradation. When the vehicle mobility is not considered, the trajectory information of the vehicle cannot be predicted, and thus the request content of the vehicle and the vehicle cache node cannot be determined, thereby increasing the content acquisition delay.
FIG. 6 is a comparison graph of cache hit rates of the present invention and the comparison method under different total cache capacities in this embodiment. As shown in fig. 6, the total cache capacity has an effect on the cache hit rates of the three methods, and the cache hit rate increases as the total cache capacity increases. This is because the larger buffer capacity means that more content can be buffered so that the buffer vehicle and the roadside unit can satisfy more requests from the vehicle user. The method predicts the social relationship between the track of the vehicle and the vehicle user, selects the vehicle as the cache more accurately, has obvious advantages compared with other strategies, and obtains better cache hit rate. When the total cache capacity is 100MB, the hit rate of the proposed strategy is improved by 18.1% and 26% compared with RC and DAC respectively.
Fig. 7 is a comparison diagram of the time delays of the present invention and the comparison method under different total buffer capacities in this embodiment. As shown in fig. 7, the total content retrieval latency of the three methods decreases as the buffer capacity increases, because as the total buffer capacity increases, each vehicle can retrieve more content from the roadside units and buffer vehicles, thus reducing the total retrieval latency. As can be seen from fig. 7, the overall delay of the present invention is lower than that of the RC method and the DAC method, because the present invention predicts the position of the vehicle at the next time, and on the basis of this, selects the cached vehicle in consideration of the interest similarity and the contact rate of the vehicle, increasing the probability that the vehicle will obtain the content from the cached vehicle, and thus having a lower content obtaining delay, the content obtaining delay of the present invention is reduced by 24.2% and 27.3% compared with the RC and the DAC, respectively, when the total cache capacity is 100 Mb.
In summary, the performance of the invention in terms of content hit rate and time delay is superior to that of the two comparison methods.
Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.
Claims (5)
1. A car networking cooperative caching method based on mobile vehicle social relation perception is characterized by comprising the following steps:
s1: set RSU = { S) of roadside units in Internet of vehicles recording 1 ,…,S n ,…,S N In which S is n N represents the nth roadside unit, N =1,2, …, N represents the number of roadside units;
the requested content set in the Internet of vehicles is X = { X = 1 ,…,x q ,…,x Q },x q Represents the Q content, Q =1,2, …, Q represents the quantity of content which can be requested in the internet of vehicles, and the content x is recorded q Has a size of d q ;
S2: for each vehicle in the internet of vehicles, predicting the moving track of the vehicle to obtain the predicted position of the next moment, and determining the roadside unit to which the vehicle belongs at the next moment; obtaining each roadside unit S at the next moment n Set of vehicles V in the coverage area n ={v n,1 ,…,v n,m ,…,v n,M },v n,m Indicating roadside units S n The M-th covered vehicle, M =1,2, …, M n ,M n Representing roadside units S n Number of vehicles within coverage;
s3: firstly, counting the number of times of requests of each content in a preset historical time period, sequencing the contents according to the number of times of the requests from large to small, and recording the content x q After sorting, the serial number is i q (ii) a Calculating the request probability of each content in a preset historical time period, taking the content with the request probability exceeding a preset threshold value as popular content, and summing the request probabilities of the popular content to be used as a distribution slope omega of the popular content; calculating the content popularity pop of the q content at the next moment by adopting the following formula q :
For roadside units S n Set of vehicles V in the coverage area n Obtaining each vehicle v n,m Requested content set L within a preset historical period of time n,m Then, the interest similarity sim between the vehicles is calculated pairwise by adopting the following calculation formula m,m′ :
Wherein phi is m,m′ Indicating a vehicle v n,m And a vehicle v n,m′ Content set L n,m 、L n,m′ M' =1,2, …, M n And | l represents the quantity of contents in the set, M q Indicating a preset historical period of timeRequested-in content x q The number of vehicles of (c);
the vehicle v is calculated by adopting the following formula n,m Interest similarity SIM m :
Similarity of interest s m Normalization is carried out to obtain interest similarity SIM 'after normalization' m :
Wherein, the SIM max 、SIM min Respectively representing a set of vehicles V n SIM (subscriber identity module) for interest similarity of all vehicles m Maximum and minimum values of;
the vehicle v is calculated by adopting the following formula n,m Interest probability in for each content n,m :
For content popularity pop q And a vehicle v n,m Interest probability in of n,m,q Carrying out weighted summation to obtain the vehicle v at the next moment n,m For content x q Request probability p of n,m,q :
p n,m,q =α×pop q +β×in n,m
Wherein, alpha and beta respectively represent a preset content popularity weighted value and a vehicle personal interest preference weighted value;
setting the requested content of the vehicle to obey the poisson distribution according to the vehicle v n,m For each content x q Request probability p of n,m,q Predicting a vehicle v n,m Content collections requested at the next time
S4: for roadside units S n Set of vehicles V n Each vehicle v in n,m And counting the number of vehicles which can be associated in unit time during the driving process as the communication contact rate f n,m Fusing interest similarity sim between vehicles m,m′ Calculating the social relationship strength z between the vehicles pairwise by adopting the following formula m,m′ :
z m,m′ =f n,m ×sim m,m′
The vehicle v is calculated by adopting the following formula n,m Strength of social relationship Z m :
Will be roadside unit S n Sequencing all vehicles from large to small according to the social relationship strength, selecting the front K vehicles as cache vehicles, determining the value of K according to actual needs, and obtaining a cache vehicle set CV n ={cv n,1 ,…,cv n,k ,…,cv n,K },cv n,k Indicating roadside units S n The kth buffer vehicle from bottom, K =1,2, …, K;
s5: calculating the time delay of different content acquisition modes, wherein the specific method comprises the following steps:
vehicle v n,m Buffering vehicle cv through V2V communication n,k Obtaining content x q Time delay T of n,m,k,q Comprises the following steps:
wherein, rate n,m,k Indicating a vehicle v n,m And buffer vehicle cv n,k The communication transmission rate between;
vehicle v n,m From roadside units S through V2R communication n Obtaining content x q Time delay T of n,m,K+1,q Comprises the following steps:
wherein, rate n,m,K+1 Indicating vehicle v n,m And roadside unit S n The communication transmission rate between;
s6: solving the following cooperative cache optimization problem to obtain each roadside unit S n The following cooperative caching scheme:
C1:a q,k′ ∈{0,1},k′=1,2,…,K+1
wherein, A = (a) q,1 ,…,a q,k ,…,a q,K+1 ) Represents a buffer decision vector, a q,k′ For caching decision identification, K' =1,2, …, K +1,a q,k =1 denotes converting content x q Buffer to buffer vehicle cv n,k ,a q,k =0 means that the content x is not deleted q Buffer-to-buffer vehicle cv n,k ,a q,K+1 =1 denotes converting content x q Buffer to roadside unit S n ,a q,K+1 =0 indicates that the content x is not written q Buffer to roadside unit S n ,Indicating buffer vehicle cv n,k The buffer capacity of the buffer memory (c),indicating roadside units S n The buffer capacity of (2).
2. The vehicle networking cooperative caching method according to claim 1, wherein the vehicle movement track prediction in the step S2 is realized based on an LSTM network, and the specific method is as follows:
and sampling historical position vectors of the vehicle according to the historical movement track of the vehicle, inputting the historical position vectors into a pre-trained LSTM network, outputting the probability that the vehicle is positioned at each accessible position, and taking the accessible position corresponding to the maximum probability as the predicted position of the vehicle.
3. The Internet of vehicles cooperative caching method according to claim 1, wherein in the step S5, the vehicles v n,m And buffer the vehicle cv n,k Rate of communication therebetween n,m,k The calculation formula of (a) is as follows:
where B denotes the transmission bandwidth, σ 2 Representing additive white Gaussian noise, P n,k Representing buffer vehicle cv n,k Transmit power of h n,m,k Indicating a vehicle v n,m And buffer the vehicle cv n,k Distance between, h n,m″,k Indicating vehicle v n,m″ And buffer the vehicle cv n,k The distance between them.
4. The Internet of vehicles cooperative caching method according to claim 1, wherein in the step S5, the vehicles v n,m And roadside unit S n Rate of communication therebetween n,m,K+1 The calculation formula of (a) is as follows:
wherein, P n Indicating roadside units S n Transmit power of h n,m,K+1 Indicating a vehicle v n,m And roadside unit S n The channel gain between.
5. The method for collaborative caching in the internet of vehicles according to claim 1, wherein the optimization problem of collaborative caching in step S6 is solved by using a reinforcement learning method, wherein a set state S is (K +1,K xc) CV +C RSU ) Setting action a to buffer decision vector a = (a) q,1 ,…,a q,k ,…,a q,K+1 ) The reward function R is the delay corresponding to the buffered decision vector, i.e. the delay
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210955357.XA CN115314944A (en) | 2022-08-10 | 2022-08-10 | Internet of vehicles cooperative caching method based on mobile vehicle social relation perception |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210955357.XA CN115314944A (en) | 2022-08-10 | 2022-08-10 | Internet of vehicles cooperative caching method based on mobile vehicle social relation perception |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115314944A true CN115314944A (en) | 2022-11-08 |
Family
ID=83861013
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210955357.XA Pending CN115314944A (en) | 2022-08-10 | 2022-08-10 | Internet of vehicles cooperative caching method based on mobile vehicle social relation perception |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115314944A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116761152A (en) * | 2023-08-14 | 2023-09-15 | 合肥工业大学 | Roadside unit edge cache placement and content delivery method |
CN116866995A (en) * | 2023-09-05 | 2023-10-10 | 江西师范大学 | Content layering edge cache strategy based on mobile vehicle |
-
2022
- 2022-08-10 CN CN202210955357.XA patent/CN115314944A/en active Pending
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116761152A (en) * | 2023-08-14 | 2023-09-15 | 合肥工业大学 | Roadside unit edge cache placement and content delivery method |
CN116761152B (en) * | 2023-08-14 | 2023-11-03 | 合肥工业大学 | Roadside unit edge cache placement and content delivery method |
CN116866995A (en) * | 2023-09-05 | 2023-10-10 | 江西师范大学 | Content layering edge cache strategy based on mobile vehicle |
CN116866995B (en) * | 2023-09-05 | 2023-12-01 | 江西师范大学 | Content layering edge cache strategy based on mobile vehicle |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhong et al. | A deep reinforcement learning-based framework for content caching | |
CN115314944A (en) | Internet of vehicles cooperative caching method based on mobile vehicle social relation perception | |
Wu et al. | Mobility-aware cooperative caching in vehicular edge computing based on asynchronous federated and deep reinforcement learning | |
CN107592656B (en) | Caching method based on base station clustering | |
CN109218747A (en) | Video traffic classification caching method in super-intensive heterogeneous network based on user mobility | |
CN113283177B (en) | Mobile perception caching method based on asynchronous federated learning | |
CN109787699B (en) | Wireless sensor network routing link state prediction method based on mixed depth model | |
CN113158544B (en) | Edge pre-caching strategy based on federal learning under vehicle-mounted content center network | |
CN115297170A (en) | Cooperative edge caching method based on asynchronous federation and deep reinforcement learning | |
CN112565377A (en) | Content grading optimization caching method for user service experience in Internet of vehicles | |
CN115633380B (en) | Multi-edge service cache scheduling method and system considering dynamic topology | |
CN114374949B (en) | Information freshness optimization-based power control mechanism in Internet of vehicles | |
CN113012013A (en) | Cooperative edge caching method based on deep reinforcement learning in Internet of vehicles | |
CN116321307A (en) | Bidirectional cache placement method based on deep reinforcement learning in non-cellular network | |
CN114995990A (en) | Method and device for unloading computing tasks, electronic equipment and computer storage medium | |
CN113114762B (en) | Data caching method and system | |
CN111465057B (en) | Edge caching method and device based on reinforcement learning and electronic equipment | |
CN110784881B (en) | Method, equipment and medium for actively caching multi-level edge nodes of Internet of things terminal | |
CN114979145B (en) | Content distribution method integrating sensing, communication and caching in Internet of vehicles | |
CN116249162A (en) | Collaborative caching method based on deep reinforcement learning in vehicle-mounted edge network | |
CN116320000A (en) | Collaborative caching method, collaborative caching device, electronic equipment and storage medium | |
Hoiles et al. | Risk-averse caching policies for YouTube content in femtocell networks using density forecasting | |
Li et al. | A proactive mobile edge cache policy based on the prediction by partial matching | |
Bai et al. | A new cooperative cache optimization algorithm for Internet of vehicles based on edge cloud network | |
CN116634396A (en) | Graph attention-based vehicle networking multi-agent edge computing content caching decision method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20230531 Address after: No.263, Kaiyuan Avenue, Luolong District, Luoyang City, Henan Province, 471000 Applicant after: HENAN University OF SCIENCE AND TECHNOLOGY Applicant after: Southwest University of Science and Technology Address before: 621000, No. 59, Qinglong Avenue, Fucheng District, Sichuan, Mianyang Applicant before: Southwest University of Science and Technology |