CN112020115A - Network access method and system based on intelligent internet - Google Patents
Network access method and system based on intelligent internet Download PDFInfo
- Publication number
- CN112020115A CN112020115A CN202010802476.2A CN202010802476A CN112020115A CN 112020115 A CN112020115 A CN 112020115A CN 202010802476 A CN202010802476 A CN 202010802476A CN 112020115 A CN112020115 A CN 112020115A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- state
- rsu
- calculating
- rss
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000001914 filtration Methods 0.000 claims abstract description 16
- 238000013135 deep learning Methods 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims description 52
- 230000002787 reinforcement Effects 0.000 claims description 50
- 230000009471 action Effects 0.000 claims description 48
- 230000007704 transition Effects 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 20
- 230000003044 adaptive effect Effects 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000006855 networking Effects 0.000 claims description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 7
- 239000000126 substance Substances 0.000 description 6
- 230000008901 benefit Effects 0.000 description 4
- 238000006073 displacement reaction Methods 0.000 description 4
- 238000005562 fading Methods 0.000 description 4
- 230000007774 longterm Effects 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W36/00—Hand-off or reselection arrangements
- H04W36/24—Reselection being triggered by specific parameters
- H04W36/30—Reselection being triggered by specific parameters by measured or perceived connection quality data
-
- 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
- H04W4/023—Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
-
- 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]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W48/00—Access restriction; Network selection; Access point selection
- H04W48/18—Selecting a network or a communication service
Abstract
The invention discloses a network access method and a system based on intelligent internet connection, wherein the switching method comprises the following steps: calculating the predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and a self-adaptive filtering algorithm; broadcasting the predicted motion state of the next moment to the RSU in the preset range around the vehicle, and calculating to obtain the predicted RSS of the vehicle at the next moment according to the position of the vehicle in the predicted motion state of the next moment and the position of the RSU; and predicting RSS according to the next moment, and performing switching decision by utilizing deep learning to obtain the optimal RSU switching of the vehicle. The embodiment of the invention provides a network access method and a network access system based on intelligent internet connection, which can effectively reduce the time delay of switching RSUs when a vehicle is in a high-speed motion state and are beneficial to improving the network access efficiency of the intelligent internet connection.
Description
Technical Field
The invention relates to the technical field of network communication, in particular to a network access method and a network access system based on intelligent internet.
Background
At present, the automobile industry has formed a consensus, and the information-based and intelligent internet of vehicles will be an important direction for future development. In the network access process of the intelligent network connection, the switching of the vehicle networking RSU (road side unit) is an extremely important ring, the existing network access method of the intelligent network connection generally realizes the switching of the RSU based on the optimization problem of Received Signal Strength (RSS), the method is simple, direct and easy to use, however, the existing network access method of the intelligent network connection only considers the received signal strength and cannot well adapt to the high-speed motion state of a vehicle, so that the time delay of switching the RSU is large and even the RSU is disconnected when the vehicle runs at a high speed, and the network access efficiency of the intelligent network connection is low.
Disclosure of Invention
The invention provides a network access method and a network access system based on intelligent networking, which aim to solve the technical problem that the prior art cannot be well adapted to the high-speed motion state of a vehicle, so that the switching time delay of an RSU (remote subscriber Unit) is longer or even the line is disconnected when the vehicle is in a high-speed running state, and the network access efficiency of the intelligent networking is lower.
The first embodiment of the invention provides a network access method based on intelligent internet connection, which comprises the following steps:
calculating a predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and a self-adaptive filtering algorithm;
broadcasting the next-time predicted motion state to the RSU in a preset range around the vehicle, and calculating to obtain the next-time predicted RSS of the vehicle according to the position of the vehicle in the next-time predicted motion state and the position of the RSU;
and predicting RSS according to the next moment, and performing switching decision by utilizing deep learning to obtain the optimal RSU switching of the vehicle.
Further, the calculating the predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm specifically includes:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and a state transition matrix from the current moment to the next moment;
and calculating a self-adaptive gain, and calculating to obtain a next-time predicted motion state of the vehicle according to the self-adaptive gain and the next-time motion state of the vehicle.
Further, the broadcasting the next-time predicted motion state to the RSU in the preset range around the vehicle, and calculating the next-time predicted RSS of the vehicle according to the vehicle position of the next-time predicted motion state and the position of the RSU specifically include:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
Further, the step of predicting RSS according to the next moment and performing a switching decision by using deep learning to obtain the optimal RSU switching of the vehicle specifically includes:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
and obtaining a decision set based on all the feedback values, calculating an optimal solution of total expected profit feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle.
Further, the obtaining of a decision set based on all the feedback values, calculating an optimal solution of total expected revenue feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle specifically includes:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set;
defining a reinforcement learning function on state-actions from the total expected revenue feedback;
and updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
A second embodiment of the present invention provides a network access system based on an intelligent internet vehicle, including:
the first calculation module is used for calculating a predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and a self-adaptive filtering algorithm;
the second calculation module is used for broadcasting the next-time predicted motion state to the RSU in the preset range around the vehicle, and calculating the next-time predicted RSS of the vehicle according to the position of the vehicle in the next-time predicted motion state and the position of the RSU;
and the switching decision module is used for predicting RSS according to the next moment, and performing switching decision by utilizing deep learning to obtain the optimal RSU switching of the vehicle.
Further, the first calculating module is specifically configured to:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and a state transition matrix from the current moment to the next moment;
and calculating a self-adaptive gain, and calculating to obtain a next-time predicted motion state of the vehicle according to the self-adaptive gain and the next-time motion state of the vehicle.
Further, the second calculation module is specifically configured to:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
Further, the handover decision module is specifically configured to:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
and obtaining a decision set based on all the feedback values, calculating an optimal solution of total expected profit feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle.
Further, the obtaining of a decision set based on all the feedback values, calculating an optimal solution of total expected revenue feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle specifically includes:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set;
defining a reinforcement learning function on state-actions from the total expected revenue feedback;
and updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
The embodiment of the invention provides a network access method and a network access system based on intelligent internet connection, which can effectively reduce the time delay of switching RSUs when a vehicle is in a high-speed motion state and avoid disconnection, and are beneficial to improving the network access efficiency of the intelligent internet connection.
Drawings
Fig. 1 is a schematic flow chart of a network access method based on an intelligent internet vehicle according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a handover decision according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a network access system based on an intelligent internet vehicle according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the description of the present application, it is to be understood that the terms "first", "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present application, "a plurality" means two or more unless otherwise specified.
In the description of the present application, it is to be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present application can be understood in a specific case by those of ordinary skill in the art.
Referring to fig. 1, an embodiment of the present invention provides a network access method based on an intelligent internet vehicle as shown in fig. 1, including:
s1, calculating the predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm;
the embodiment of the invention adopts the self-adaptive filtering to measure the RSS and the motion state information of the vehicle in real time.
S2, broadcasting the next-time predicted motion state to the RSU in the preset range around the vehicle, and calculating to obtain the next-time predicted RSS of the vehicle according to the position of the vehicle in the next-time predicted motion state and the position of the RSU;
and S3, forecasting RSS according to the next moment, and performing switching decision by using deep learning to obtain the optimal RSU switching of the vehicle.
The embodiment of the invention predicts the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm, and broadcasts the prediction result to the RSU in the preset range around the vehicle, so that the RSU can calculate the predicted RSS of the vehicle at the next moment according to the motion state of the vehicle at the next moment, and when switching the RSU, switching can be carried out in advance according to the predicted RSS at the next moment, and the switching time delay can be effectively reduced, thereby improving the network access efficiency of the intelligent internet.
Furthermore, after the RSS of the vehicle is predicted at the next moment through calculation, a deep reinforcement learning algorithm is adopted to link the current switching decision of the vehicle with the long-term benefit, and the switching decision is more accurate through continuous reinforcement learning, so that the switching efficiency of the vehicle RSU can be improved, and the network access efficiency of the intelligent networked vehicle can be further improved.
As a specific implementation manner of the embodiment of the present invention, the method for calculating the predicted motion state of the vehicle at the next time according to the current time state of the vehicle and the adaptive filtering algorithm specifically includes:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the state transition matrix from the current moment to the next moment;
in the embodiment of the invention, the vehicle motion state m at the next momentt+1And measured values z of the vehiclet+1Are respectively:
mt+1=Φmt+wt,wt~N(0,Q) (1)
zt+1=Hmt+vt,vt~N(0,R) (2)
wherein m ist([xtytvx,tvy,t]T) Representing the vehicle motion at time t, at mtIn, xtIs the displacement of the vehicle in the horizontal direction at time t, ytIs the displacement of the vehicle in the vertical direction at time t, vx,tIs the speed in the horizontal direction of the vehicle at time t, vy,tIs the speed of the vehicle in the vertical direction at time t, phi is the state transition matrix from the current time t to the next time t +1, zt+1Is a measured value of the vehicle, H is a measurement matrix linking the measured value and the state value, and a random number wtFor parameters related to process noise, the random number vtFor parameters related to measurement noise, wtHas a covariance matrix of Q, vtThe covariance matrix of (2) is R.
And calculating the self-adaptive gain, and calculating the predicted motion state of the vehicle at the next moment according to the self-adaptive gain and the motion state of the vehicle at the next moment.
In the embodiment of the invention, the state transition matrix and w are calculated according to the state transition matrix from the current time t to the next time t +1tThe covariance matrix of (A) is calculated for Q to obtain the vehicle error covariance prediction at the next moment
Wherein the content of the first and second substances,the vehicle motion state prediction for the next time t + 1.
According to the next momentVehicle error covariance predictionCalculating to obtain self-adaptive gain, and calculating to obtain the predicted motion state m of the vehicle at the next moment according to the self-adaptive gain and the motion state of the vehicle at the next momentt+1。
Wherein, Kt+1The adaptive gain at time t + 1.
The embodiment of the invention calculates the predicted motion state of the vehicle at the next moment according to the adaptive gain and the motion state of the vehicle at the next moment, calculates the adaptive gain at the next moment by using the vehicle error covariance prediction at the next moment, and combines the measured value z of the vehiclet+1The predicted motion state of the vehicle at the next moment is calculated, and the influence of the error of the measured value in the actual motion of the vehicle on the predicted value is considered, so that the calculation of the predicted motion state at the next moment is more accurate, and the reliability and the accuracy of the switching of the RSU of the vehicle are improved.
As a specific implementation manner of the embodiment of the present invention, the next-time predicted motion state is broadcasted to the RSU within the preset range around the vehicle, and the next-time predicted RSS of the vehicle is calculated according to the vehicle position of the next-time predicted motion state and the position of the RSU, specifically:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
Distance d 'between vehicle position and RSU position at next moment't+1The expression is as follows:
according to distance d't+1Average signal received strength and inter-link shadowing fading, calculating the predicted RSS at the next time:
X'RSU=ZRSU+μRSU-10ηlogd't+1 (8)
wherein, X'RSUPredicted RSS for the vehicle; zRSUIs the shadow fading between links; mu.sRSUIs the average signal received strength; η is the propagation constant.
In the embodiment of the invention, the next moment predicted RSS of the vehicle is calculated, and when the vehicle is in a high-speed motion state, the next moment predicted RSS is utilized to advance the switching of the RSU, so that the time delay of the switching of the RSU can be effectively reduced, and the success rate of the switching of the RSU of the vehicle is improved.
As a specific implementation manner of the embodiment of the present invention, RSS is predicted according to the next moment, and a switching decision is made by deep learning to obtain the optimal RSU switching of the vehicle, which specifically includes:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
as a specific implementation, the RSS and the velocity v are divided according to size and the size grade is represented by specific numerical values, specifically: rssi∈{1,2,3,4,5},viE { -4, -3, -2, -1,0,1,2,3,4}, it should be noted that when the speed v is negative, it indicates that the vehicle is moving toward the direction of the principle base station. The vehicle state may be expressed as S { { id { (id) }1,rss1,v1},…,{id4,rss4,v4}}. The action is denoted by α, the network number corresponding to the action is {1,2, …, N }, which represents the handover decision, and the action space can be denoted as a ═ a ∈ {1,2, …, N } }.
Based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
based on the current vehicle state stBelongs to S, learns the agent and performs an action atE.g. a to perform the switching of RSUs. With probability of state transition Ps,s'(at) The vehicle is enabled to enter the next state, and a feedback value r for executing the action is obtainedt。
And obtaining a decision set based on all the feedback values, and calculating an optimal solution of total expected profit feedback according to the decision set to obtain the optimal RSU switching of the vehicle.
Based on the technical scheme, an optimal decision set pi is finally obtained through repeated learning processes in a learning agent*(s)∈A。
As a specific implementation manner of the embodiment of the present invention, a decision set is obtained based on all feedback values, an optimal solution of total expected revenue feedback is calculated according to the decision set, and optimal RSU switching of a vehicle is obtained, specifically:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set;
as a specific implementation mode, the decision set is pi(s) epsilon A, and the total expected income feedback Vπ(s) is:
wherein, Vπ(s) total expected revenue feedback; gamma raytE [0,1) is a revenue factor; e is the desired operator; r(s)t,x(st) ) is a feedback function for the state action.
Defining a reinforcement learning function on the state-action according to the total expected revenue feedback;
converting formula (9) to be based on optimal strategy pi*(s) form of Bayesian equation for E A:
wherein, V*(s) feeding back an optimal solution for the total expected revenue of the current system; gamma raytIs a yield factor and gammat∈[0,1); Ps,s'(a) A state transition probability of transitioning to s' for performing the action α at state s; r (s, a) is a feedback function corresponding to the state action; v*(s ') feeding back an optimal solution for the total expected revenue for the system state s'.
Defining a reinforcement learning function for the state-action pairs (s, α):
under the action of the optimal strategy, continuously defining the optimal solution of the reinforcement learning function in the current state:
wherein the content of the first and second substances,performing reinforcement learning function optimal solution for the current state;the reinforcement learning function may be updated based on each execution of the action and the environmental feedback.
And updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
In the embodiment of the present invention, the reinforcement learning function is calculated as:
wherein the content of the first and second substances,for the updated reinforcement learning function;the original reinforcement learning function is adopted; gamma is a revenue factor; ρ is a learning rate andt (s, a) represents a certain state- -the number of times an action pair is accessed. Mqt(s ', a') is a reinforcement learning function that performs action a 'at state s'.
When T (s, a) goes to infinity, ρ goes to 0, at which timeConverge toThrough repeated learning and decision-making processes, the learner obtains an optimal decision set, and obtains a maximum reinforcement learning function as an optimal solution for the feedback of the optimal total expected income.
Please refer to fig. 2, which is a flowchart illustrating a handover decision according to an embodiment of the present invention.
The embodiment of the invention has the following beneficial effects:
the embodiment of the invention predicts the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm, and broadcasts the prediction result to the RSU in the preset range around the vehicle, so that the RSU can calculate the predicted RSS of the vehicle at the next moment according to the motion state of the vehicle at the next moment, and when switching the RSU, switching can be carried out in advance according to the predicted RSS at the next moment, and the switching time delay can be effectively reduced, thereby improving the network access efficiency of the intelligent internet.
Furthermore, after the RSS of the vehicle is predicted at the next moment through calculation, a deep reinforcement learning algorithm is adopted to link the current switching decision of the vehicle with the long-term benefit, and the switching decision is more accurate through continuous reinforcement learning, so that the switching efficiency of the vehicle RSU can be improved, and the network access efficiency of the intelligent networked vehicle can be further improved.
Referring to fig. 3, a second embodiment of the present invention provides a network access system based on an intelligent internet vehicle, including:
the first calculation module 10 is used for calculating a predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and an adaptive filtering algorithm;
the embodiment of the invention adopts the self-adaptive filtering to measure the RSS and the motion state information of the vehicle in real time.
The second calculating module 20 is configured to broadcast the predicted motion state at the next time to the RSU within the preset range around the vehicle, and calculate the predicted RSS at the next time of the vehicle according to the vehicle position of the predicted motion state at the next time and the position of the RSU;
and the switching decision module 30 is used for predicting the RSS at the next moment and performing switching decision by using deep learning to obtain the optimal RSU switching of the vehicle.
The embodiment of the invention predicts the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm, and broadcasts the prediction result to the RSU in the preset range around the vehicle, so that the RSU can calculate the predicted RSS of the vehicle at the next moment according to the motion state of the vehicle at the next moment, and when switching the RSU, switching can be carried out in advance according to the predicted RSS at the next moment, and the switching time delay can be effectively reduced, thereby improving the network access efficiency of the intelligent internet.
Furthermore, after the RSS of the vehicle is predicted at the next moment through calculation, a deep reinforcement learning algorithm is adopted to link the current switching decision of the vehicle with the long-term benefit, and the switching decision is more accurate through continuous reinforcement learning, so that the switching efficiency of the vehicle RSU can be improved, and the network access efficiency of the intelligent networked vehicle can be further improved.
As a specific implementation manner of the embodiment of the present invention, the first calculating module 10 is specifically configured to:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the state transition matrix from the current moment to the next moment;
in the embodiment of the invention, the vehicle motion state m at the next momentt+1And measured values z of the vehiclet+1Are respectively:
mt+1=Φmt+wt,wt~N(0,Q) (1)
zt+1=Hmt+vt,vt~N(0,R) (2)
wherein m ist([xtytvx,tvy,t]T) Representing the vehicle motion at time t, at mtIn, xtIs the displacement of the vehicle in the horizontal direction at time t, ytIs the displacement of the vehicle in the vertical direction at time t, vx,tIs the speed in the horizontal direction of the vehicle at time t, vy,tIs the speed of the vehicle in the vertical direction at time t, phi is the state transition matrix from the current time t to the next time t +1, zt+1Is a measured value of the vehicle, H is a measurement matrix linking the measured value and the state value, and a random number wtFor parameters related to process noise, the random number vtFor parameters related to measurement noise, wtHas a covariance matrix of Q, vtThe covariance matrix of (2) is R.
And calculating the self-adaptive gain, and calculating the predicted motion state of the vehicle at the next moment according to the self-adaptive gain and the motion state of the vehicle at the next moment.
In the embodiment of the invention, the state transition matrix and w are calculated according to the state transition matrix from the current time t to the next time t +1tThe covariance matrix of (A) is calculated for Q to obtain the vehicle error covariance prediction at the next moment
Wherein the content of the first and second substances,the vehicle motion state prediction for the next time t + 1.
Vehicle error covariance prediction from next time instantCalculating to obtain self-adaptive gain, and calculating to obtain the predicted motion state m of the vehicle at the next moment according to the self-adaptive gain and the motion state of the vehicle at the next momentt+1。
Wherein, Kt+1The adaptive gain at time t + 1.
The embodiment of the invention calculates the predicted motion state of the vehicle at the next moment according to the adaptive gain and the motion state of the vehicle at the next moment, calculates the adaptive gain at the next moment by using the vehicle error covariance prediction at the next moment, and combines the measured value z of the vehiclet+1The predicted motion state of the vehicle at the next moment is calculated, and the influence of the error of the measured value in the actual motion of the vehicle on the predicted value is considered, so that the calculation of the predicted motion state at the next moment is more accurate, and the reliability and the accuracy of the switching of the RSU of the vehicle are improved.
As a specific implementation manner of the embodiment of the present invention, the second calculating module 20 is specifically configured to:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
Distance d 'between vehicle position and RSU position at next moment't+1The expression is as follows:
according to distance d't+1Average signal received strength and inter-link shadowing fading, calculating the predicted RSS at the next time:
X'RSU=ZRSU+μRSU-10ηlogd't+1 (8)
wherein, X'RSUPredicted RSS for the vehicle; zRSUIs the shadow fading between links; mu.sRSUIs the average signal received strength; η is the propagation constant.
In the embodiment of the invention, the next moment predicted RSS of the vehicle is calculated, and when the vehicle is in a high-speed motion state, the next moment predicted RSS is utilized to advance the switching of the RSU, so that the time delay of the switching of the RSU can be effectively reduced, and the success rate of the switching of the RSU of the vehicle is improved.
As a specific implementation manner of the embodiment of the present invention, the switching decision module 30 is specifically configured to:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
as a specific implementation, the RSS and the velocity v are divided according to size and the size grade is represented by specific numerical values, specifically: rssi∈{1,2,3,4,5},viE { -4, -3, -2, -1,0,1,2,3,4}, it should be noted that when the speed v is negative, it indicates that the vehicle is moving toward the direction of the principle base station. The vehicle state may be expressed as S { { id { (id) }1,rss1,v1},…,{id4,rss4,v4}}. The action is denoted by α, the network number corresponding to the action is {1,2, …, N }, which represents the handover decision, and the action space may be denoted as a ═ N{a|a∈{1,2,…,N}}。
Based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
based on the current vehicle state stBelongs to S, learns the agent and performs an action atE.g. a to perform the switching of RSUs. With probability of state transition Ps,s'(at) The vehicle is enabled to enter the next state, and a feedback value r for executing the action is obtainedt。
And obtaining a decision set based on all the feedback values, and calculating an optimal solution of total expected profit feedback according to the decision set to obtain the optimal RSU switching of the vehicle.
Based on the technical scheme, an optimal decision set pi is finally obtained through repeated learning processes in a learning agent*(s)∈A。
As a specific implementation manner of the embodiment of the present invention, a decision set is obtained based on all feedback values, an optimal solution of total expected revenue feedback is calculated according to the decision set, and optimal RSU switching of a vehicle is obtained, specifically:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set; as a specific implementation mode, the decision set is pi(s) epsilon A, and the total expected income feedback Vπ(s) is:
wherein, Vπ(s) total expected revenue feedback; gamma raytE [0,1) is a revenue factor; e is the desired operator; r(s)t,x(st) ) is a feedback function for the state action.
Defining a reinforcement learning function on the state-action according to the total expected revenue feedback;
converting formula (9) to be based on optimal strategy pi*(s) form of Bayesian equation for E A:
wherein, V*(s) feeding back an optimal solution for the total expected revenue of the current system; gamma raytIs a yield factor and gammat∈[0,1); Ps,s'(a) A state transition probability of transitioning to s' for performing the action α at state s; r (s, a) is a feedback function corresponding to the state action; v*(s ') feeding back an optimal solution for the total expected revenue for the system state s'.
Defining a reinforcement learning function for the state-action pairs (s, α):
under the action of the optimal strategy, continuously defining the optimal solution of the reinforcement learning function in the current state:
wherein the content of the first and second substances,performing reinforcement learning function optimal solution for the current state;the reinforcement learning function may be updated based on each execution of the action and the environmental feedback.
And updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
In the embodiment of the present invention, the reinforcement learning function is calculated as:
wherein the content of the first and second substances,for the updated reinforcement learning function;the original reinforcement learning function is adopted; gamma is a revenue factor; ρ is a learning rate andt (s, a) represents a certain state- -the number of times an action pair is accessed. Mqt(s ', a') is a reinforcement learning function that performs action a 'at state s'.
When T (s, a) goes to infinity, ρ goes to 0, at which timeConverge toThrough repeated learning and decision-making processes, the learner obtains an optimal decision set, and obtains a maximum reinforcement learning function as an optimal solution for the feedback of the optimal total expected income.
Please refer to fig. 2, which is a flowchart illustrating a handover decision according to an embodiment of the present invention.
The embodiment of the invention has the following beneficial effects:
the embodiment of the invention predicts the motion state of the vehicle at the next moment according to the current moment state of the vehicle and the adaptive filtering algorithm, and broadcasts the prediction result to the RSU in the preset range around the vehicle, so that the RSU can calculate the predicted RSS of the vehicle at the next moment according to the motion state of the vehicle at the next moment, and when switching the RSU, switching can be carried out in advance according to the predicted RSS at the next moment, and the switching time delay can be effectively reduced, thereby improving the network access efficiency of the intelligent internet.
Furthermore, after the RSS of the vehicle is predicted at the next moment through calculation, a deep reinforcement learning algorithm is adopted to link the current switching decision of the vehicle with the long-term benefit, and the switching decision is more accurate through continuous reinforcement learning, so that the switching efficiency of the vehicle RSU can be improved, and the network access efficiency of the intelligent networked vehicle can be further improved.
The foregoing is a preferred embodiment of the present invention, and it should be noted that it would be apparent to those skilled in the art that various modifications and enhancements can be made without departing from the principles of the invention, and such modifications and enhancements are also considered to be within the scope of the invention.
Claims (10)
1. A network access method based on intelligent Internet connection is characterized by comprising the following steps:
calculating a predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and a self-adaptive filtering algorithm;
broadcasting the next-time predicted motion state to the RSU in a preset range around the vehicle, and calculating to obtain the next-time predicted RSS of the vehicle according to the position of the vehicle in the next-time predicted motion state and the position of the RSU;
and predicting RSS according to the next moment, and performing switching decision by utilizing deep learning to obtain the optimal RSU switching of the vehicle.
2. The network access method based on the intelligent internet vehicle as claimed in claim 1, wherein the step of calculating the predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and an adaptive filtering algorithm specifically comprises the steps of:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and a state transition matrix from the current moment to the next moment;
and calculating a self-adaptive gain, and calculating to obtain a next-time predicted motion state of the vehicle according to the self-adaptive gain and the next-time motion state of the vehicle.
3. The network access method based on the intelligent internet vehicle as claimed in claim 1, wherein the next-time predicted motion state is broadcasted to the RSU within a preset range around the vehicle, and the next-time predicted RSS of the vehicle is calculated according to the vehicle position of the next-time predicted motion state and the position of the RSU, specifically:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
4. The network access method based on the intelligent internet vehicle as claimed in claim 1, wherein the RSS is predicted according to the next moment, and a switching decision is made by deep learning to obtain the optimal RSU switching of the vehicle, specifically:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
and obtaining a decision set based on all the feedback values, calculating an optimal solution of total expected profit feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle.
5. The network access method based on the intelligent internet vehicle as claimed in claim 4, wherein a decision set is obtained based on all the feedback values, an optimal solution of total expected revenue feedback is calculated according to the decision set, and optimal RSU switching of the vehicle is obtained, specifically:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set;
defining a reinforcement learning function on state-actions from the total expected revenue feedback;
and updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
6. The utility model provides a network access system based on intelligent networking car which characterized in that includes:
the first calculation module is used for calculating a predicted motion state of the vehicle at the next moment according to the current moment state of the vehicle and a self-adaptive filtering algorithm;
the second calculation module is used for broadcasting the next-time predicted motion state to the RSU in the preset range around the vehicle, and calculating the next-time predicted RSS of the vehicle according to the position of the vehicle in the next-time predicted motion state and the position of the RSU;
and the switching decision module is used for predicting RSS according to the next moment, and performing switching decision by utilizing deep learning to obtain the optimal RSU switching of the vehicle.
7. The network access system based on the intelligent internet vehicle as claimed in claim 6, wherein the first computing module is specifically configured to:
calculating the motion state of the vehicle at the next moment according to the current moment state of the vehicle and a state transition matrix from the current moment to the next moment;
and calculating a self-adaptive gain, and calculating to obtain a next-time predicted motion state of the vehicle according to the self-adaptive gain and the next-time motion state of the vehicle.
8. The network access system based on the intelligent internet vehicle as claimed in claim 6, wherein the second computing module is specifically configured to:
and extracting the vehicle position at the next moment in the predicted motion state at the next moment, calculating the distance between the vehicle position at the next moment and the RSU position in the preset range around the vehicle, and calculating to obtain the predicted RSS (received signal strength) of the vehicle at the next moment according to the distance.
9. The network access system based on the intelligent internet protocol as claimed in claim 6, wherein the switching decision module is specifically configured to:
setting a vehicle state set, wherein the vehicle state of the vehicle state set comprises RSS of a current service network and RSS of a neighboring network, a network ID corresponding to each RSS and the speed of a vehicle relative to a corresponding base station;
based on the current vehicle state, learning an agent and selecting to execute an action to switch the RSU, after executing the action, enabling the vehicle to enter the next state according to the state transition probability, and obtaining a feedback value of the executed action;
and obtaining a decision set based on all the feedback values, calculating an optimal solution of total expected profit feedback according to the decision set, and obtaining the optimal RSU switching of the vehicle.
10. The network access system based on the intelligent internet vehicle as claimed in claim 9, wherein a decision set is obtained based on all the feedback values, an optimal solution of total expected revenue feedback is calculated according to the decision set, and optimal RSU switching of the vehicle is obtained, specifically:
calculating total expected income feedback according to the income factors, the expected operation method and the feedback function corresponding to the state action on the basis of the decision set;
defining a reinforcement learning function on state-actions from the total expected revenue feedback;
and updating the reinforcement learning function according to the reinforcement learning function, the income factor and the learning rate, and when the learning rate tends to 0, obtaining the maximum reinforcement learning function as the optimal solution of the total expected income feedback to realize the optimal RSU switching of the vehicle.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010802476.2A CN112020115A (en) | 2020-08-11 | 2020-08-11 | Network access method and system based on intelligent internet |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010802476.2A CN112020115A (en) | 2020-08-11 | 2020-08-11 | Network access method and system based on intelligent internet |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112020115A true CN112020115A (en) | 2020-12-01 |
Family
ID=73504235
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010802476.2A Pending CN112020115A (en) | 2020-08-11 | 2020-08-11 | Network access method and system based on intelligent internet |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112020115A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104320789A (en) * | 2014-11-05 | 2015-01-28 | 安徽大学 | Internet of vehicles RSU optimal allocation method based on game theory |
CN106879034A (en) * | 2015-12-14 | 2017-06-20 | 电信科学技术研究院 | A kind of method for switching network and device of vehicle node |
CN107613532A (en) * | 2017-08-31 | 2018-01-19 | 重庆邮电大学 | Fuzzy vertical switching method based on residence time prediction in vehicle heterogeneous network |
US20180184350A1 (en) * | 2016-12-22 | 2018-06-28 | Microsoft Technology Licensing, Llc | Handover in 5g microcellular |
CN110446184A (en) * | 2019-07-29 | 2019-11-12 | 华南理工大学 | A kind of car networking method for routing of multi-mode switching |
CN111083634A (en) * | 2019-12-16 | 2020-04-28 | 重庆邮电大学 | CDN and MEC-based vehicle networking mobility management method |
CN111314983A (en) * | 2020-03-31 | 2020-06-19 | 四川九强通信科技有限公司 | Cross-domain seamless switching method based on heaven-earth integrated network |
-
2020
- 2020-08-11 CN CN202010802476.2A patent/CN112020115A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104320789A (en) * | 2014-11-05 | 2015-01-28 | 安徽大学 | Internet of vehicles RSU optimal allocation method based on game theory |
CN106879034A (en) * | 2015-12-14 | 2017-06-20 | 电信科学技术研究院 | A kind of method for switching network and device of vehicle node |
US20180184350A1 (en) * | 2016-12-22 | 2018-06-28 | Microsoft Technology Licensing, Llc | Handover in 5g microcellular |
CN107613532A (en) * | 2017-08-31 | 2018-01-19 | 重庆邮电大学 | Fuzzy vertical switching method based on residence time prediction in vehicle heterogeneous network |
CN110446184A (en) * | 2019-07-29 | 2019-11-12 | 华南理工大学 | A kind of car networking method for routing of multi-mode switching |
CN111083634A (en) * | 2019-12-16 | 2020-04-28 | 重庆邮电大学 | CDN and MEC-based vehicle networking mobility management method |
CN111314983A (en) * | 2020-03-31 | 2020-06-19 | 四川九强通信科技有限公司 | Cross-domain seamless switching method based on heaven-earth integrated network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105791143A (en) | Adaptive rate control method based on mobility and DSRC/WAVE network relevance feedback | |
CN114389678B (en) | Multi-beam satellite resource allocation method based on decision performance evaluation | |
CN111314983A (en) | Cross-domain seamless switching method based on heaven-earth integrated network | |
CN103957525A (en) | Malicious node detection method based on clustering trust evaluation in internet of vehicles | |
CN102547833A (en) | Heterogeneous network pre-switching method based on mobile prediction | |
CN113904948B (en) | 5G network bandwidth prediction system and method based on cross-layer multidimensional parameters | |
CN111935303A (en) | Task unloading method based on intention perception in air-ground integrated Internet of vehicles | |
CN115209426B (en) | Dynamic deployment method for digital twin servers in edge car networking | |
CN109859515B (en) | GPS positioning data compensation method in public transport system and electronic equipment | |
CN113487860A (en) | Intelligent traffic signal control method | |
CN114650567B (en) | Unmanned aerial vehicle auxiliary V2I network task unloading method | |
CN110445825B (en) | Super-dense network small station code cooperation caching method based on reinforcement learning | |
CN111343608A (en) | Internet of vehicles reinforcement learning routing method based on position information | |
Xu et al. | Link reliability-based adaptive routing for multilevel vehicular networks | |
CN115529278A (en) | Multi-agent reinforcement learning-based automatic data center network ECN regulation and control method | |
CN115412134A (en) | Off-line reinforcement learning-based user-centered non-cellular large-scale MIMO power distribution method | |
CN112020115A (en) | Network access method and system based on intelligent internet | |
CN113923737A (en) | Intelligent handover switching method for LTE-M system | |
Yacheur et al. | DRL-based RAT Selection in a Hybrid Vehicular Communication Network | |
CN116782296A (en) | Digital twinning-based internet-of-vehicles edge computing and unloading multi-objective decision method | |
CN113950113A (en) | Hidden Markov-based Internet of vehicles switching decision algorithm | |
CN110087280B (en) | Vehicle density estimation method based on beacon message | |
CN108327718A (en) | A kind of Vehicle Adaptive Cruising Control Systems and its control method | |
CN109041009B (en) | Internet of vehicles uplink power distribution method and device | |
Tong et al. | Adaptive on-ramp merging strategy under imperfect communication performance |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201201 |
|
RJ01 | Rejection of invention patent application after publication |