CN116095770A - Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network - Google Patents

Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network Download PDF

Info

Publication number
CN116095770A
CN116095770A CN202211573834.2A CN202211573834A CN116095770A CN 116095770 A CN116095770 A CN 116095770A CN 202211573834 A CN202211573834 A CN 202211573834A CN 116095770 A CN116095770 A CN 116095770A
Authority
CN
China
Prior art keywords
network
terminal
switching
vehicle
handover
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
Application number
CN202211573834.2A
Other languages
Chinese (zh)
Inventor
吴利平
钟世林
马彬
陈鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202211573834.2A priority Critical patent/CN116095770A/en
Publication of CN116095770A publication Critical patent/CN116095770A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/0005—Control or signalling for completing the hand-off
    • H04W36/0083—Determination of parameters used for hand-off, e.g. generation or modification of neighbour cell lists
    • H04W36/00837—Determination of triggering parameters for hand-off
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/14—Reselecting a network or an air interface
    • 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
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/24—Reselection being triggered by specific parameters
    • H04W36/32—Reselection being triggered by specific parameters by location or mobility data, e.g. speed data
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a method for judging the cross-region cooperation self-adaptive switching in ultra-dense heterogeneous wireless network, belonging to the mobile communication field, which comprises the following steps: first, according to a Kalman filter position prediction model with improved historical track information of a vehicle, the positions of the switching terminal at two moments are predicted. Second, an alternative network set for handover and hopping is generated in advance according to the predicted location. Then, by defining a jump factor and adopting a multi-attribute decision algorithm for correcting the interval number of Jaccard similarity, a self-adaptive switching decision scheme of cross-region cooperation is provided to generate an optimal switching strategy for the terminal. Finally, experimental simulation shows that the algorithm can reduce the switching times of the vehicle-mounted terminal, reduce the switching failure rate and improve the transmission efficiency of the network.

Description

Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network
Technical Field
The invention belongs to a vertical switching method in an ultra-dense heterogeneous wireless network, and belongs to the field of mobile communication. In particular to a cross-region cooperation self-adaptive switching judgment method.
Background
With the evolution of 5G technology, deployment of small cell base stations on infrastructures of telegraph poles, street lamps, buses and the like in urban areas becomes possible in the future, dense deployment of small cell networks can improve spectrum efficiency and network access capacity, and conditions are created for explosive growth of data transmission in the Internet of vehicles. However, in the vehicular ad hoc heterogeneous wireless network, due to the high dynamic movement of the vehicle and the miniaturization of the cell structure, the vehicular terminal is also faced with the embarrassment of continuously switching between networks, which tends to increase signaling overhead and risk of link disconnection, thereby affecting user experience. Therefore, how to reduce the switching times as much as possible while guaranteeing the service quality of the vehicle-mounted terminal through a vertical switching algorithm is a hot problem in research in the field aiming at the frequent switching problem caused by the continuous crossing of the 5G microcellular network, the WiFi network and the like by the vehicle-mounted terminal with high dynamic property.
Currently, there are several documents that have studied the problem of frequent handover in heterogeneous wireless networks, and all achieve certain results. Such as: the literature [ Palas M.R., islam R., roy P., et al multi-criteria handover mobility management in 5G cellular network[J ]. Computer Communications,2021,174 (8): 81-91] proposes a multi-attribute vertical switching algorithm based on movement trend quantization, and the problem of excessive switching in a common multi-attribute decision algorithm is alleviated by predicting a target area of a terminal by considering movement trend quantization parameters of the terminal. The literature [ Yang Mingji, wu, fan Huafeng ] heterogeneous internet of vehicles vertical handover algorithm based on motion trend prediction [ J ] microelectronics and computer, 2018,35 (4): 119-123,129 ] calculates the duration of access of a vehicle-mounted terminal to a base station by predicting the motion trend of a vehicle, divides the terminal into a narrow mobile node and a wide mobile node according to the duration, and then performs corresponding handover strategies respectively, thereby reducing handover delay. Document [ Tokuyama k., kimura t., miyoshi n.data rate and handoff rate analysis for user mobility in cellular networks [ C ]//2018IEEE Wireless Communications and Networking Conference (WCNC) & Barcelona, spaina: IEEE Press 2018:1-6 ] proposes a time-based jump switching algorithm, which reduces the switching rate of a terminal by setting a jump time threshold to control the switching frequency of a mobile user. The literature [ Al-Naffouri, tareq Y., elSawy, et Al, velocity-aware handover management in two-tier cellular networks [ J ]. IEEE Transactions on Wireless Communications,2017,16 (3): 1851-1867 ] proposes a switching scheme based on velocity sensing, and a velocity sensing model is built according to random geometric theory, so that base stations needing to be skipped on a terminal motion track are determined, and the switching failure rate is reduced. The literature [ Costa A, pacheco L, D Ros rio, et al, skip-based handover algorithm for video distribution over ultra-dense VANET [ J ]. Computer Networks,2020,176:1-12 ] proposes a multi-attribute-based jump switching algorithm by designing a jump mechanism combining mobility prediction, quality of service and quality of experience perception, thereby improving switching reliability and relieving ping-pong switching.
Although the above-mentioned document can alleviate frequent switching to a certain extent, in the vehicular ad hoc heterogeneous wireless network environment, if the movement trend is only predicted, and accurate analysis is not performed on specific position changes of the vehicle-mounted terminal, the frequent switching problem is more serious. Furthermore, the jump handover algorithms mentioned in these documents, although reducing the number of handovers of the terminal to some extent, fail to decide on the target network to which the terminal can access after jumping. In view of the above, a vertical handover algorithm based on location prediction and cross-region cooperation (Location Prediction and Cross Region Cooperation, LPCRC) is proposed herein, which first introduces an improved kalman filter (Improve Kalman Filter, IKF) model to predict the location of the terminal at the next two moments, generating in advance an alternative network set for handover and hopping; then, a self-adaptive cross-region cooperation switching judgment scheme is provided by defining a jump factor and adopting an interval number multi-attribute decision algorithm for correcting Jaccard similarity, and an optimal switching strategy is generated for the terminal.
Disclosure of Invention
The present invention is directed to solving the above problems of the prior art. A method for determining handover in a heterogeneous ultra-dense wireless network is provided. The technical scheme of the invention is as follows:
A method for judging the cross-region cooperation self-adaptive switching in ultra-dense heterogeneous wireless network includes the following steps:
101. a switching triggering step: the signal intensity and the network bandwidth of the vehicle-mounted terminal in the current network are collected periodically, a switching trigger factor alpha is calculated, and if the terminal i is in the network j, the switching trigger factor alpha is calculated ij (t) =1, triggering a handover, otherwise, not triggering;
102. mobility prediction step: the method comprises the steps that motion data of a vehicle are collected through positioning equipment on the vehicle and recorded in a historical track database, and when the vehicle-mounted terminal triggers switching, the position of the terminal at the next two moments is predicted through an IKF model by utilizing historical track information of the vehicle;
103. a switching decision step: firstly, a candidate network set CNS_1 for switching and a cooperative network set CNS_2 for jumping are generated in advance according to a predicted position, then a network topology and a terminal motion state are adopted to define a jumping factor, the scores of each network in the CNS_1 and the CNS_2 are calculated through a multi-attribute decision algorithm for correcting the interval number of Jaccard similarity, and finally, an optimal switching strategy is generated for the terminal according to the jumping factor and a network scoring result.
Further, the step 101 periodically collects the signal intensity and the network bandwidth of the vehicle-mounted terminal in the current network, and calculates the switching trigger factor α, which specifically includes:
201. Received signal strength: at time t, the received signal strength of the terminal i accessing the network j is expressed as:
RSS ij (t)＝P j -ηlgdis ij (t)+μ (0,σ) (1)
wherein P is j Representing the wireless signal transmit power of network j, η representing the path loss factor, dis ij (t) represents the distance from the terminal i to the network j at t moments, mu (0,σ) To obey a gaussian random variable with a mean of 0 and a variance of σ;
202. network bandwidth: the network bandwidth obtained by the terminal i accessing the network j at the t-th moment can be expressed as:
Figure BDA0003988496880000031
wherein u represents the number of vehicle-mounted terminals accessed into the network j at the t-th moment, and B r Representing bandwidth per resource block, prb ij (t) represents the number of resource blocks allocated to the network j by the in-vehicle terminal i, PRB j The maximum number of available resource blocks of the network j is represented, so the handover trigger α of the vehicle-mounted terminal i in the original access network j at the t-th moment can be represented as:
Figure BDA0003988496880000041
wherein when the RSS is lower than the set threshold value RSS th And hysteresis margin (Hysteresis Margin, HM) or bandwidth below the minimum bandwidth requirement required by terminal i to operate the service
Figure BDA0003988496880000042
When alpha is ij (t) =1, requiring a trigger switch, otherwise, α ij (t) =0, not triggered.
Further, the step 102 collects motion data of the vehicle through a positioning device on the vehicle and records the motion data to a historical track database, and when the vehicle-mounted terminal triggers switching, the position of the terminal at the next two moments is predicted through an IKF model by utilizing the historical track information of the vehicle, and the method specifically includes:
301. Theoretical prediction: let the state of the vehicle at time t be S (t) = [ la (t), lo (t)] T Where la (t) is latitude data at time t and lo (t) is longitude data; if the optimal estimation state at time t-1 is S' (t-1), estimating the prediction state at time t according to the theoretical model is:
Figure BDA0003988496880000043
wherein F represents a state transition matrix for describing how the state at the previous time transitions to the next state，
Figure BDA0003988496880000044
Representing the prediction noise, the error is represented as a covariance matrix:
Figure BDA0003988496880000045
by Q p And (3) representing noise of the prediction model, and deriving a transmission process of an adjacent moment error covariance matrix by using the formula (4) to the formula (5) is as follows:
P′(t)＝FP(t-1)F T +Q p (6)
302. the state observed by the GPS positioning equipment at the moment t is marked as Z (t), the observation matrix is marked as S (t), and the observation noise is marked as Q g The transformation process of the predicted state to the observed state of the vehicle at time t can be expressed as:
Z(t)＝HS(t)+Q g (7)
303. and (5) updating the state: the predicted state and the observed state at the time t are obtained in the formula (4) and the formula (7), and the predicted value is corrected through the observed value, so that the corrected optimal estimated state is obtained as follows:
Figure BDA0003988496880000051
wherein the method comprises the steps of
Figure BDA0003988496880000052
For the residual error of the actual observation value and the expected observation value, K (t) is the Kalman gain at the time t, and the calculation process is as follows:
K(t)＝P′(t)H T (HP′(t) T H T +Q g ) -1 (9)
The effect of the Kalman gain is to balance the predicted state covariance P with the observed state covariance Q g To determine the ratio of the effect of the predictive model and the observation model in the prediction process; after obtaining the Kalman gainFor the next prediction, the noise covariance matrix P (t) of the best estimation state needs to be updated, where e is the identity matrix;
P(t)＝(Ε-K(t)H)P′(t) (10)
304. prediction model improvement: the Kalman filter is improved by introducing an attenuation memory filtering method, and an improved Kalman filter position prediction model is obtained.
Further, the step 103 generates a candidate network set cns_1 for handover and a cooperative network set cns_2 for hopping, specifically including:
assuming that the vehicle-mounted terminal i triggers switching at the t-th moment, the position of the vehicle-mounted terminal at the t+1th moment can be predicted by utilizing the historical motion trail of the vehicle-mounted terminal at the previous t-th moment according to the IKF model, when a switching request of a user arrives, the background discovers networks in all connection ranges according to the predicted position, and the network obtained at the position is used as a candidate network set after triggering switching and is marked as CNS_1; similarly, according to the historical motion track of the previous t+1 time, the position of the t+2 time can be predicted, the network obtained by the position is used as a cooperation network set for jumping switching after switching triggering, and is denoted as CNS_2, and CNS_1 and CNS_2 are collectively called as an alternative network set.
Further, the handover decision parameter specifically includes: when the vehicle-mounted terminal triggers switching, a new network is decided to be accessed to the terminal in the alternative network sets CNS_1 and CNS_2; in the motion process of the vehicle-mounted terminal, the data transmission rate, the network delay and the packet loss rate are key indexes for measuring the performance of the access network, and the 3 parameters are used for evaluating the network performance; because of the influence of network topology and terminal motion state, accessing new target network may cause frequent switching of terminal, defining jump factor, marking the network which is easy to cause frequent switching in the alternative network set as the network which needs to be skipped.
Further, the calculation formulas of the data transmission rate, the network delay and the packet loss rate specifically include:
data transmission rate: according to shannon formula, the data transmission speed of the terminal accessing the network can be knownThe rate, bandwidth and signal-to-noise ratio parameters are related, and at the t-th moment, the terminal i accesses the data transmission rate e obtained by the network j ij (t) can be expressed as:
e ij (t)＝B ij (t)×log 2 (1+SNR ij (t)) (11)
wherein B is ij (t) bandwidth resources allocated to terminal i by network j at time t, SNR ij (t) represents a signal-to-noise ratio, the value of which approximates the ratio of RSS to interference noise I in the network;
(2) Network delay: setting the relation between the time delay and the network load as an exponential function relation; if the attribute time delay of the network j is d' j The time delay d of the terminal i accessing the network j at the t-th moment ij (t) can be expressed as:
Figure BDA0003988496880000061
(3) Packet loss rate: the packet loss rate refers to the ratio of the number of lost data packets to the number of all transmitted data packets in a certain time, and it is assumed that the number of data packets transmitted by the network j at the first t moments is ψ total The number of data packets for which acknowledgement is received is ψ ack The packet loss rate gamma of the terminal i accessing the network j at the t moment ij (t) can be expressed as:
Figure BDA0003988496880000062
(4) Jump factor: a hopping factor delta is defined to mark networks in the set of alternative networks that need to be skipped. After triggering the handover at the t-th moment, the terminal i hops by a factor delta at the j-th alternative network ij (t) can be expressed as:
Figure BDA0003988496880000063
in the method, in the process of the invention,
Figure BDA0003988496880000064
representing the coverage area, cl, of the candidate network j j Representing the chord length of the trajectory of the candidate network j, τ ij Indicating the residence time of terminal i in network j.
Further, the specific steps of the interval number multi-attribute decision algorithm for correcting the Jaccard similarity in the step 103 are as follows:
(1) Constructing an interval number decision matrix: assuming that N networks are concentrated in the network to be evaluated, the number of the network attributes participating in the evaluation is M, acquiring the interval numbers of the M attributes of the N networks before decision, determining the interval number of each network attribute by the maximum and minimum values in multiple data sampling, and respectively acquiring the maximum and minimum values obtained by multiple sampling the kth attribute of the network j as follows
Figure BDA0003988496880000071
And->
Figure BDA0003988496880000072
The number of intervals of the kth attribute of network j can be expressed as + ->
Figure BDA0003988496880000073
The interval number decision matrix to be decided can be expressed as: />
Figure BDA0003988496880000074
(2) Normalized attribute interval number: matrix of logarithmic interval numbers
Figure BDA0003988496880000075
Normalizing to obtain normalized matrix +.>
Figure BDA0003988496880000076
Wherein->
Figure BDA0003988496880000077
Formulas (15) and (16) are normalized procedures for benefit-type and cost-type network parameters, respectively:
Figure BDA0003988496880000078
Figure BDA0003988496880000079
(3) Determining interval type ideal scheme: for better measurement of the difference between networks, it is assumed that the interval type ideal scheme of each network attribute is Θ= [ Θ ] 1 ,Θ 2 ,...,Θ M ]Wherein Θ is k Can be expressed as:
Figure BDA00039884968800000710
(4) Calculate the revised Jaccard similarity: the Jaccard similarity is used for describing similarity and difference between sets, the greater the Jaccard similarity is, the higher the similarity of the sets is, and since the number of intervals can also be regarded as a set of numbers, the Jaccard similarity of the normalized attribute value with respect to the ideal solution Θ can be expressed as:
Figure BDA00039884968800000711
since the similarity between two sections cannot be compared with the similarity between another section when the middle points of the sections are the same, the right end point of the section can be added into the calculation process of Jaccard to correct the sections, and the corrected Jaccard similarity can be expressed as:
Figure BDA0003988496880000081
Thus, the modified Jaccard similarity for each network attribute value in the normalized decision matrix corresponding to an ideal solution can be represented by the matrix as ζ= (ζ) jk ) NM ；
(5) Determining the optimal weight of each network attribute: the weight of each network attribute in the judgment process is determined according to the minimum sum of the deviation, and the corresponding optimization model is as follows:
Figure BDA0003988496880000082
(6) Calculating comprehensive similarity: after the weight of each network attribute is obtained, the comprehensive similarity theta of the networks j in the network set to be evaluated can be obtained through weighted summation j ；
Figure BDA0003988496880000083
Further, the generating the optimal switching strategy specifically includes:
after the vehicle-mounted terminal triggers the switching, an IKF position prediction model is adopted to generate a candidate network set CNS_1 and a cooperative network set CNS_2 of the terminal in advance, and the comprehensive similarity score of all networks in the two network sets can be calculated through an MJS-INMADM algorithm and recorded as theta 1 And theta 2 By combining the jump factor delta of each network in the alternative network set, an optimal switching strategy can be generated for the terminal, and the generation process is as follows:
selecting network O with highest comprehensive similarity from CNS_1 1 Wherein
Figure BDA0003988496880000084
If network O 1 Jump factor of->
Figure BDA0003988496880000085
The optimal strategy is to switch to network O directly 1 The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, network O with highest comprehensive similarity and jump factor value of 1 is obtained from CNS_2 2 Wherein->
Figure BDA0003988496880000086
The optimal strategy is to switch to network O directly 2 。
The invention has the advantages and beneficial effects as follows:
1. the invention aims at the frequent switching scene caused by the continuous crossing of the 5G microcell, wiFi and other networks by the vehicle-mounted terminal with high dynamic property. A method for determining handover in a heterogeneous ultra-dense wireless network is provided.
2. In the conventional kalman filter model, since the new observed value is suppressed from old data for the correction of the predicted result in the next step, when the vehicle speed is greatly changed, a great error may exist between the predicted result and the actual position. Therefore, an improved kalman filter model is introduced in step 102 to predict the position of the vehicle terminal after triggering the switching, and the dependency on the old value is reduced by strengthening the weight of the new observed value, so that the prediction error is reduced, and the candidate network set is conveniently generated in advance.
3. Since access to a new target network may cause frequent handovers of terminals, due to network topology and terminal motion status, a hopping factor is defined in step 103 by cell area, trace chord length and residence time, which is used to mark networks in the candidate network set that are prone to frequent handovers.
4. In an actual network environment, network parameters can fluctuate within a certain interval range, and a traditional multi-attribute decision algorithm generally adopts a determined network attribute value to judge an alternative network, so that a certain error is brought to a decision result. In addition, the invention obtains the relevant attribute of the time network by predicting the position information of the terminal at the future time. In this case, if the network parameters are still considered as a certain value for making the handover decision, a larger error will be necessary. Based on this, in step 103, a multi-attribute decision algorithm of interval number for correcting Jaccard similarity is adopted to calculate the scores of each network in the alternative network set, and an adaptive cross-region cooperation switching decision scheme is provided in combination with the jump factor. The scheme can adaptively generate the optimal switching strategy for the switching terminal, effectively reduces the switching times and reduces the switching failure rate.
Drawings
FIG. 1 is a diagram of a heterogeneous wireless network simulation scenario in a urban core area in accordance with a preferred embodiment of the present invention;
FIG. 2 is a flow chart of an adaptive cross-zone cooperative handoff decision scheme for mitigating frequent handoffs;
FIG. 3 is a diagram depicting the motion state of a vehicle in a Kalman filter;
FIG. 4 is a graph of a modified Kalman filter position prediction model;
FIG. 5 is a comparison of predicted position data for different prediction models;
FIG. 6 is a comparison of the proportion of the prediction results of different prediction models under different error distances;
FIG. 7 is an alternative network set topology block diagram of a vehicle terminal;
FIG. 8 is a flowchart of a multi-attribute interval number decision algorithm based on modified Jaccard similarity;
FIG. 9 is a graph showing the comparison of the accumulated switching times of different methods;
FIG. 10 is a table tennis hand-off count comparison of different methods;
FIG. 11 is a graph showing the comparison of handover failure rates for different methods;
FIG. 12 is a graph of overall throughput versus network for different approaches;
FIG. 13 is a graph of algorithm time overhead versus different methods;
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and specifically described below with reference to the drawings in the embodiments of the present invention. The described embodiments are only a few embodiments of the present invention.
The technical scheme for solving the technical problems is as follows:
the method comprehensively considers the problem of network congestion caused by short-time aggregated movement of large-scale vehicle-mounted terminals aiming at urban traffic peak periods in the ultra-dense heterogeneous wireless network introduced into the vehicle self-organizing network, can effectively relieve network congestion, balances load among networks and improves user experience.
The self-adaptive cross-region cooperation switching judging method provided by the invention comprises the following steps:
a method for judging handover in a cross-region cooperation self-adaptive manner in an ultra-dense heterogeneous wireless network aims at frequent handover problems caused by continuous crossing of a 5G micro-cell, wiFi and other networks by a high-dynamic vehicle-mounted terminal in a city core region, and comprises the following steps:
101. and (3) switching triggering: the signal intensity and the network bandwidth of the vehicle-mounted terminal in the current network are collected periodically, a switching trigger factor alpha is calculated, and if alpha is calculated ij (t) =1, triggering a handover, otherwise, not triggering;
102. mobility prediction: the method comprises the steps that motion data of a vehicle are collected through positioning equipment on the vehicle and recorded in a historical track database, and when the vehicle-mounted terminal triggers switching, the position of the terminal at the next two moments is predicted through an IKF model by utilizing historical track information of the vehicle;
103. switching decision: firstly, a candidate network set CNS_1 for switching and a cooperative network set CNS_2 for jumping are generated in advance according to a predicted position, then a network topology and a terminal motion state are adopted to define a jumping factor, the scores of each network in the CNS_1 and the CNS_2 are calculated through a multi-attribute decision algorithm for correcting the interval number of Jaccard similarity, and finally, an optimal switching strategy is generated for the terminal according to the jumping factor and a network scoring result.
Further, according to the switching trigger described in step 101, the present invention proposes to periodically collect the signal intensity and the network bandwidth of the vehicle-mounted terminal in the current network, and calculate the switching trigger factor α. The relevant definition and specific steps of the switching trigger decision are as follows:
received signal strength: the received signal strength is a basic index for the terminal to evaluate the network and reflects the channel quality of the network. Since there is a path loss in the signal during transmission, at the t-th moment, the received signal strength of the terminal i accessing the network j can be expressed as:
RSS ij (t)＝P j -ηlgdis ij (t)+μ (0,σ) (1)
wherein P is j Representing the wireless signal transmit power of network j, η representing the path loss factor, dis ij (t) represents the distance from the terminal i to the network j at t moments, mu (0,σ) To obey the mean value 0, the variance is sigmaIs a gaussian random variable of (c).
Network bandwidth: it is assumed that the bandwidth resources of each network are divided into several physical resource blocks, the bandwidth of each resource block being. When the number of the vehicle-mounted terminals accessed in the network j is lower than the rated number, each terminal can obtain the fixed number of the resource blocks, and when the number exceeds the rated number, each terminal equally divides all the resource blocks. Thus, the network bandwidth obtained by the terminal i accessing the network j at the t-th moment can be expressed as:
Figure BDA0003988496880000111
Wherein u represents the number of vehicle-mounted terminals accessed into the network j at the t-th moment, and prb ij (t) represents the number of resource blocks allocated to the network j by the in-vehicle terminal i, PRB j Representing the maximum number of resource blocks that can be provided by network j. Therefore, the switching trigger factor α of the vehicle-mounted terminal i in the original access network j at the t-th moment can be expressed as:
Figure BDA0003988496880000121
further, in the mobility prediction according to step 102, the method is characterized in that, the positioning device on the vehicle collects motion data of the vehicle and records the motion data in the history track database, and when the vehicle-mounted terminal triggers the switching, the position of the terminal at the next two moments is predicted by using the history track information of the vehicle through the IKF model, and specifically includes:
theoretical prediction: let the state of the vehicle at time t be S (t) = [ la (t), lo (t)] T Where la (t) is latitude data at time t and lo (t) is longitude data. If the optimal estimation state at time t-1 is S' (t-1), estimating the prediction state at time t according to the theoretical model is:
Figure BDA0003988496880000122
wherein F represents a stateThe transition matrix is mainly used for describing how the state at the last moment is transited to the next state.
Figure BDA0003988496880000123
Representing the prediction noise, wherein a plurality of uncertainty factors exist in the motion process of the vehicle, so that a gap exists between a prediction state and a best estimation state obtained by a theoretical model, and a covariance matrix is used for representing errors caused by the uncertainty as follows:
Figure BDA0003988496880000124
In addition, since the prediction model itself may also have errors, if Q is used p The noise representing the prediction model is derived by the following transmission process that the adjacent moment error covariance matrix can be derived by the formula (4) and the formula (5):
P′(t)＝FP(t-1)F T +Q p (6)
GPS measurement: because each vehicle is provided with GPS equipment, the motion information in the motion process of the vehicle can be acquired and recorded. If the state observed by the GPS positioning device at the time t is marked as an observation matrix and the observation noise is marked as an observation noise, the transformation process from the predicted state to the observed state of the vehicle at the time t can be expressed as follows:
Z(t)＝HS(t)+Q g (7)
and (5) updating the state: the predicted state and the observed state at the time t are obtained in the formula (4) and the formula (7), respectively, and the predicted value needs to be corrected through the observed value due to errors in the prediction process, so that the corrected optimal estimated state is obtained as follows:
Figure BDA0003988496880000131
wherein the method comprises the steps of
Figure BDA0003988496880000132
For actual observations and pre-observationsResidual error of the period observation value, K (t) is Kalman gain at the moment t, and the calculation process is as follows:
K(t)＝P′(t)H T (HP′(t) T H T +Q g ) -1 (9)
the effect of the Kalman gain is to balance the predicted state covariance P with the observed state covariance Q g To determine the proportion of the predictive and observer models that are active in the prediction process. After obtaining the kalman gain, the noise covariance matrix P (t) of the best estimated state needs to be updated for the next prediction, where e is the identity matrix.
P(t)＝(Ε-K(t)H)P′(t) (10)
Prediction model improvement: in the kalman filter model, since the correction of the new observed value for the next predicted result is suppressed by the old data, there is a possibility that the predicted result has a large error from the actual position when the vehicle speed is greatly changed. Therefore, researchers have proposed methods such as particle filtering and attenuation memory filtering to reduce errors in the prediction process. The attenuation memory filtering method is widely applied to the improvement of the Kalman filter due to the characteristics of simple calculation, excellent performance and the like. The method is that the noise covariance matrix P (t-1) in the formula (6) is multiplied by an attenuation factor eta with a value larger than 1 0 To strengthen the new observation weights and thereby alleviate the dependency on the old values. Therefore, the invention introduces an attenuation memory filtering method to improve the Kalman filter, and obtains an improved Kalman filter position prediction model.
Further, the handover decision according to step 103 is characterized in that, firstly, a candidate network set cns_1 for handover and a cooperative network set cns_2 for hopping are generated in advance according to the predicted location, then, a hopping factor is defined by using a network topology and a terminal motion state, the score of each network in cns_1 and cns_2 is calculated by modifying a multi-attribute decision algorithm of the number of intervals of Jaccard similarity, and finally, an optimal handover strategy is generated for the terminal according to the hopping factor and the network scoring result. The relevant parameter definition and algorithm process in the self-adaptive cross-region cooperation switching decision are as follows:
Generating an alternative network set: assuming that the vehicle-mounted terminal i triggers switching at the t-th moment, the position of the vehicle-mounted terminal at the t+1th moment can be predicted by utilizing the historical motion trail of the vehicle-mounted terminal at the previous t-th moment according to the IKF model, when a switching request of a user arrives, the background discovers networks in all connection ranges according to the predicted position, and the network obtained at the position is used as a candidate network set after triggering switching and is marked as CNS_1. Similarly, according to the historical motion track of the previous t+1 time, the position of the t+2 time can be predicted, and the network obtained at the position is used as a cooperation network set for jumping switching after switching triggering and is marked as CNS_2. The present invention refers to cns_1 and cns_2 collectively as an alternative network set.
Switching judgment parameters: when the vehicle terminal triggers the handover, a new network needs to be decided for the terminal to access in the alternative network sets CNS_1 and CNS_2. In the motion process of the vehicle-mounted terminal, the data transmission rate, the network delay and the packet loss rate are key indexes for measuring the performance of the access network, so that the invention evaluates the network performance by using the 3 parameters. The invention defines the jump factor, marks the network which is easy to cause frequent switching in the alternative network set as the network which needs to be skipped because the access to the new target network is possibly caused by the influence of the network topology and the motion state of the terminal. The definition of the respective network parameters and hopping factors is given below:
(1) Data transmission rate: according to shannon's formula, the data transmission rate of the terminal accessing the network is related to parameters such as bandwidth, signal-to-noise ratio, etc. Thus, at time t, the data transmission rate e obtained by the terminal i accessing the network j ij (t) can be expressed as:
e ij (t)＝B ij (t)×log 2 (1+SNR ij (t)) (11)
wherein B is ij (t) bandwidth resources allocated to terminal i by network j at time t, SNR ij (t) represents a signal-to-noise ratio, the value of which approximates the ratio of RSS to interference noise I in the network;
(2) Network delay: network latency is generally load dependent, adding to a networkThe same load amount may produce a more severe delay, so the relationship between delay and network load may be set as an exponential function. If the attribute time delay of the network j is d' j The time delay d of the terminal i accessing the network j at the t-th moment ij (t) can be expressed as:
Figure BDA0003988496880000151
(3) Packet loss rate: the packet loss rate refers to the ratio of the number of lost data packets to the number of all transmitted data packets in a certain time, and it is assumed that the number of data packets transmitted by the network j at the first t moments is ψ total The number of data packets for which acknowledgement is received is ψ ack The packet loss rate gamma of the terminal i accessing the network j at the t moment ij (t) can be expressed as:
Figure BDA0003988496880000152
(4) Jump factor: because of the diversity of network topologies and the high motion characteristics of the terminals, the coverage of each network, the track chord length formed by the vehicle running and the network, and the residence time of the vehicle are all quite different, and these factors easily cause frequent switching of the terminals. In order to avoid the terminal accessing to the network which is easy to cause frequent switching, the invention comprehensively considers the factors, and defines the hopping factor delta to mark the network which needs to be skipped in the alternative network set. After triggering the handover at the t-th moment, the terminal i hops by a factor delta at the j-th alternative network ij (t) can be expressed as:
Figure BDA0003988496880000153
in the method, in the process of the invention,
Figure BDA0003988496880000154
representing the coverage area, cl, of the candidate network j j Representing the chord length of the trajectory of the candidate network j, τ ij Indicating that terminal i is in the networkResidence time in j.
Interval number multi-attribute decision algorithm for correcting Jaccard similarity: in an actual network environment, network parameters fluctuate within a certain interval, and a traditional multi-attribute decision algorithm generally adopts a determined network attribute value to judge an alternative network, so that a certain error is brought to a decision result. In addition, the invention obtains the relevant attribute of the time network by predicting the position information of the terminal at the future time. In this case, if the network parameters are still considered as a certain value for making the handover decision, a larger error will be necessary. Based on the above, the invention calculates the comprehensive scores of the networks in the alternative network set by adopting an interval number multi-attribute decision algorithm (MJS-INMADM) for correcting the Jaccard similarity. The algorithm can evaluate the performance of the network by utilizing the interval range of the network attribute under the condition that the attribute weight is unknown, and improves the accuracy of the switching judgment. The specific process is as follows:
(1) Constructing an interval number decision matrix: assuming that N networks are concentrated in the network to be evaluated, the number of the network attributes participating in the evaluation is M, and acquiring the interval numbers of the M attributes of the N networks before decision. The present chapter determines the number of intervals of each network attribute by the maximum and minimum values in multiple data samples, for example, the maximum and minimum values obtained by sampling the kth attribute of the network j multiple times are respectively
Figure BDA0003988496880000161
And->
Figure BDA0003988496880000162
The number of intervals of the kth attribute of network j can be expressed as + ->
Figure BDA0003988496880000163
The interval number decision matrix to be decided can be expressed as: />
Figure BDA0003988496880000164
(2) Normalized attribute interval number: matrix of logarithmic interval numbers
Figure BDA0003988496880000165
Normalizing to obtain normalized matrix +.>
Figure BDA0003988496880000166
Wherein->
Figure BDA0003988496880000167
Formulas (15) and (16) are normalized procedures for benefit-type and cost-type network parameters, respectively:
Figure BDA0003988496880000168
Figure BDA0003988496880000169
(3) Determining interval type ideal scheme: for better measurement of the difference between networks, it is assumed that the interval type ideal scheme of each network attribute is Θ= [ Θ ] 1 ,Θ 2 ,...,Θ M ]Wherein Θ is k Can be expressed as:
Figure BDA00039884968800001610
(4) Calculate the revised Jaccard similarity: the Jaccard similarity is used for describing similarity and difference between sets, the greater the Jaccard similarity is, the higher the similarity of the sets is, and since the number of intervals can also be regarded as a set of numbers, the Jaccard similarity of the normalized attribute value with respect to the ideal solution Θ can be expressed as:
Figure BDA0003988496880000171
since the similarity between two interval numbers cannot be compared by the size of Jaccard similarity when the interval midpoints of them are the same. Therefore, the right endpoint of the interval number can be added to the calculation process of Jaccard, and the corrected Jaccard similarity can be expressed as:
Figure BDA0003988496880000172
Thus, the modified Jaccard similarity for each network attribute value in the normalized decision matrix corresponding to an ideal solution can be represented by the matrix as ζ= (ζ) jk ) NM 。
(5) Determining the optimal weight of each network attribute: as the closer the network properties are to the ideal solution, the better, i.e. the smaller the deviation of the network from the ideal solution. Therefore, the weight of each network attribute in the judgment process can be determined according to the minimum sum of the deviations, and the corresponding optimization model is as follows:
Figure BDA0003988496880000173
(6) Calculating comprehensive similarity: after the weight of each network attribute is obtained, the comprehensive similarity theta of the networks j in the network set to be evaluated can be obtained through weighted summation j 。
Figure BDA0003988496880000174
Generating an optimal switching strategy: after the vehicle-mounted terminal triggers the switching, an IKF position prediction model is adopted to generate a candidate network set CNS_1 and a cooperative network set CNS_2 of the terminal in advance, and the comprehensive similarity score of all networks in the two network sets can be calculated through an MJS-INMADM algorithm and recorded as theta 1 And theta 2 . By combining the hopping factors delta of the networks in the alternative network set, an optimal switching strategy can be generated for the terminal. The process of generating the optimal switching strategy is as follows: selecting network O with highest comprehensive similarity from CNS_1 1 Wherein
Figure BDA0003988496880000181
If network O 1 Jump factor of->
Figure BDA0003988496880000182
The optimal strategy is to switch to network O directly 1 The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, network O with highest comprehensive similarity and jump factor value of 1 is obtained from CNS_2 2 Wherein
Figure BDA0003988496880000183
The optimal strategy is to switch to network O directly 2 。
Based on the above analysis, the present invention devised the algorithm flow chart shown in fig. 2.
In order to verify the invention, a simulation experiment is carried out on an MATLAB platform, and the following simulation scene is set: a network consisting of three access technologies of 5G, WLAN and a self-organizing network is taken as an ultra-dense self-organizing heterogeneous network model, and a simulation scene is built on a MATLAB platform for simulation analysis. Assume that 50 5G macro base stations, 200 5G micro base stations, 150 wireless local area networks and a plurality of vehicle self-organizing networks are deployed in a scene, and a heterogeneous wireless network simulation scene of a city core area is shown in fig. 1.
In the simulation process, it is assumed that the arrival of the vehicle within the entire network coverage obeys a poisson distribution with an arrival rate of 1.ltoreq.λ.ltoreq.10. To further highlight the superiority of the present invention, the proposed method (Location Prediction and Cross Region Cooperation, LPCRC) was analyzed in comparison with the multi-attribute decision-based skip switching method (MD-SHO) in literature [ Costa, pacheco L, D Rosa rio, et al, skip-based handover algorithm for video distribution over ultra-dense VANET [ J ]. Computer Networks,2020,176:1-12 ], literature [ Kim J., cho J., jeong J., et al, fuzzy logic based handoff scheme for heterogeneous vehicular mobile networks [ C ]// International Conference on High Performance Computing & formulation, bologna, italy: IEEE,2014:863-870 ], fuzzy logic-based vertical switching method (FL-VHO).
Fig. 5 shows the actual running track of the vehicle, the KF prediction model and the IKF prediction model, and it can be seen from the figure that both the prediction models can be well matched with the actual track. However, by further analyzing the prediction results, the proportion of the prediction results of the two prediction models under different error distances shown in fig. 6 can be obtained. In 1700 sets of data, the KF model predicted data point to real track point distance error is higher than 5m,10m,20m,30m, 3.17%,2.05%,1.23%,1%, and the IKF position prediction model is 2.29%,1.52%,1%,0.82%, respectively. Therefore, the IKF prediction model can more accurately make a position prediction of the vehicle.
Fig. 9 is a graph showing the change of the cumulative switching times of the 3 algorithms with the increase of the moving distance of the vehicle-mounted terminal. As can be seen from the figure, the cumulative switching times of several algorithms gradually increase as the moving distance of the vehicle-mounted terminal increases. The cumulative number of switches for 3 algorithms is only a few times when the distance of movement is less than 3000 meters, but the cumulative number of switches for FL-VHO algorithms is always much higher than for MD-SHO and LPCRC-VHO algorithms when the distance exceeds 3000 meters. This is because the latter two algorithms reduce the number of switching times of the vehicle-mounted terminal by introducing a skip switching mechanism. In addition, since the LPCRC-VHO algorithm introduces the Ad Hoc network, when the vehicle-mounted terminal triggers handover, the vehicle-mounted terminal may access to the Ad Hoc network capable of maintaining persistent connection. Therefore, as the vehicle-mounted terminal continues to move, the accumulated switching times of the LPCRC-VHO algorithm are slightly lower than those of the MD-SHO algorithm.
Fig. 10 shows ping-pong switching times of 3 algorithms at different speeds, where the ping-pong switching times can more directly reflect the influence of frequent switching in the moving process of the vehicle-mounted terminal. As can be seen from the graph, as the vehicle speed increases, the number of ping-pong handovers generated by each algorithm increases, but the ping-pong handovers of the LPCRC-VHO and MD-SHO algorithms are significantly lower than the FL-VHO algorithms. The method is characterized in that the change of the speed enables the vehicle-mounted terminal to trigger more switching times, and the MD-SHO and LPCRC-VHO algorithms can avoid frequent initiation of switching of the vehicle-mounted terminal to a certain extent by introducing a jump switching mechanism, so that the switching times are reduced. In addition, the ping-pong switching times of the LPCRC-VHO algorithm are always the lowest, and the ping-pong switching caused by the fact that the vehicle-mounted terminal accesses to a network with shorter residence time is avoided by defining the jump factor.
Fig. 11 is a graph showing the change of the handover failure rate of 3 algorithms with the increase of the number of vehicle terminals. As can be seen from the graph, as the number of vehicle terminals increases, the switching failure rate of the 3 algorithms gradually increases, wherein the switching failure rate of the FL-VHO algorithm is always higher than that of the MD-SHO and LPCRC-VHO algorithms, and the main reason for this phenomenon is that: the MD-SHO and LPCC-VHO algorithms adopt a jump mechanism, so that the switching times of the vehicle-mounted terminal in the network are reduced, meanwhile, the residence time of the terminal in the network is inspected, and the terminal is prevented from accessing to the network which is easy to frequently switch. In addition, because the LPCC-VHO algorithm can predict the future position of the terminal through the IKF position model after triggering the switching, the target access network can reserve network resources for the switching terminal in advance, and the access success rate of the terminal can be improved.
Fig. 12 is a graph showing the change of the total throughput of the network of 3 algorithms with the increase of the number of the vehicle-mounted terminals. As can be seen from the figure, the overall throughput of the network increases as the number of terminals increases, and the overall throughput of the 3 algorithms increases rapidly when the number of terminals is less than 500. When the number of terminals is higher than 500, the trend of the total throughput rise gradually becomes smooth due to the limited total resources of the network. But when the number of terminals is the same, the overall throughput of the LPCRC-VHO algorithm is always higher than the other 2 algorithms. This is because the LPCRC-VHO algorithm introduces an Ad Hoc network, increasing network capacity. In addition, when switching judgment is carried out, the LPCRC-VHO algorithm increases the network selection range of the vehicle-mounted terminal by generating a candidate network set and a cooperative network set; and the introduction of the jump switching mechanism can also avoid the vehicle-mounted access to the network with too high load to a certain extent, thereby improving the throughput of the network.
Fig. 13 shows the time overhead of 3 algorithms as a function of the number of experiments. It can be seen from the figure that the time overhead of the FL-VHO algorithm is significantly higher than that of the MD-SHO and LPCRC-VHO algorithms. This is because the rule base of the FL-VHO algorithm is large and it takes a long time to make a fuzzy logic inference decision on the optimal network when there are many decision parameters to consider. In addition, the algorithm time cost of the chapter is slightly higher than that of the MD-SHO algorithm because the chapter needs to evaluate the networks in both the candidate network set and the cooperative network set when the optimal switching strategy is generated.
The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
The above examples should be understood as illustrative only and not limiting the scope of the invention. Various changes and modifications to the present invention may be made by one skilled in the art after reading the teachings herein, and such equivalent changes and modifications are intended to fall within the scope of the invention as defined in the appended claims.

Claims (8)

1. A method for judging the cross-region cooperation self-adaptive switching in ultra-dense heterogeneous wireless network is characterized by comprising the following steps:
101. a switching triggering step: the signal intensity and the network bandwidth of the vehicle-mounted terminal in the current network are collected periodically, a switching trigger factor alpha is calculated, and if the terminal i is in the network j, the switching trigger factor alpha is calculated ij (t) =1, triggering a handover, otherwise, not triggering;
102. mobility prediction step: the method comprises the steps that motion data of a vehicle are collected through positioning equipment on the vehicle and recorded in a historical track database, and after the vehicle-mounted terminal triggers switching, the position of the terminal at the next two moments is predicted through an improved Kalman filter model by utilizing historical track information of the vehicle;
103. a switching decision step: firstly, a candidate network set CNS_1 for switching and a cooperative network set CNS_2 for jumping are generated in advance according to a predicted position, then a network topology and a terminal motion state are adopted to define a jumping factor, the scores of each network in the CNS_1 and the CNS_2 are calculated through a multi-attribute decision algorithm for correcting the interval number of Jaccard similarity, and finally, an optimal switching strategy is generated for the terminal according to the jumping factor and a network scoring result.
2. The method for determining the handover in a heterogeneous ultra-dense wireless network according to claim 1, wherein the step 101 periodically collects the signal strength and the network bandwidth of the vehicle-mounted terminal in the current network, and calculates the handover trigger factor α, and specifically includes:
201. received signal strength: at time t, the received signal strength of the terminal i accessing the network j is expressed as:
RSS ij (t)＝P j -ηlgdis ij (t)+μ (0,σ) (1)
Wherein P is j Representing the wireless signal transmit power of network j, η representing the path loss factor, dis ij (t) represents the distance from the terminal i to the network j at t moments, mu (0,σ) To obey a gaussian random variable with a mean of 0 and a variance of σ;
202. network bandwidth: the network bandwidth obtained by the terminal i accessing the network j at the t-th moment can be expressed as:
Figure FDA0003988496870000011
wherein u represents the number of vehicle-mounted terminals accessed into the network j at the t-th moment, and B r Representing bandwidth per resource block, prb ij (t) represents the number of resource blocks allocated to the network j by the in-vehicle terminal i, PRB j The maximum number of available resource blocks of the network j is represented, so the handover trigger α of the vehicle-mounted terminal i in the original access network j at the t-th moment can be represented as:
Figure FDA0003988496870000021
wherein when the RSS is lower than the set threshold value RSS th And hysteresis margin (Hysteresis Margin, HM) or bandwidth below the minimum bandwidth requirement required by terminal i to operate the service
Figure FDA0003988496870000022
When alpha is ij (t) =1, requiring a trigger switch, otherwise, α ij (t) =0, not triggered.
3. The method for determining the handover in a heterogeneous ultra-dense wireless network according to claim 1, wherein the step 102 collects motion data of a vehicle through a positioning device on the vehicle and records the motion data in a historical track database, and when the vehicle-mounted terminal triggers the handover, the position of the terminal at the next two moments is predicted through an IKF model by using the historical track information of the vehicle, and the method specifically comprises the following steps:
301. Theoretical prediction: let the state of the vehicle at time t be S (t) = [ la (t), lo (t)] T Where la (t) is latitude data at time t and lo (t) is longitude data; if the optimal estimation state at time t-1 is S' (t-1), estimating the prediction state at time t according to the theoretical model is:
Figure FDA0003988496870000023
wherein F represents a state transition matrix for describing how the state at the previous time transitions to the next state,
Figure FDA0003988496870000025
representing the prediction noise, the error is represented as a covariance matrix:
Figure FDA0003988496870000024
by Q p And (3) representing noise of the prediction model, and deriving a transmission process of an adjacent moment error covariance matrix by using the formula (4) to the formula (5) is as follows:
P′(t)＝FP(t-1)F T +Q p (6)
302. GPS measurement: the state observed by the GPS positioning equipment at the moment t is marked as Z (t), the observation matrix is marked as S (t), and the observation noise is marked as Q g The transformation process of the predicted state to the observed state of the vehicle at time t can be expressed as:
Z(t)＝HS(t)+Q g (7)
303. and (5) updating the state: the predicted state and the observed state at the time t are obtained in the formula (4) and the formula (7), and the predicted value is corrected through the observed value, so that the corrected optimal estimated state is obtained as follows:
Figure FDA0003988496870000031
wherein the method comprises the steps of
Figure FDA0003988496870000032
For the residual error of the actual observation value and the expected observation value, K (t) is the Kalman gain at the time t, and the calculation process is as follows:
K(t)＝P′(t)H T (HP′(t) T H T +Q g ) -1 (9)
The effect of the Kalman gain is to balance the predicted state covariance P with the observed state covariance Q g To determine the ratio of the effect of the predictive model and the observation model in the prediction process; after obtaining the kalman gain, for the next prediction, the noise covariance matrix P (t) of the best estimation state needs to be updated, wherein the e is a unit matrix;
P(t)＝(Ε-K(t)H)P′(t) (10)
304. prediction model improvement: the Kalman filter is improved by introducing an attenuation memory filtering method, and an improved Kalman filter position prediction model is obtained.
4. The method for determining handover adaptation in a heterogeneous ultra-dense wireless network according to claim 1, wherein the step 103 generates a candidate network set cns_1 for handover and a cooperative network set cns_2 for hopping, specifically comprising:
assuming that the vehicle-mounted terminal i triggers switching at the t-th moment, the position of the vehicle-mounted terminal at the t+1th moment can be predicted by utilizing the historical motion trail of the vehicle-mounted terminal at the previous t-th moment according to the IKF model, when a switching request of a user arrives, the background discovers networks in all connection ranges according to the predicted position, and the network obtained at the position is used as a candidate network set after triggering switching and is marked as CNS_1; similarly, according to the historical motion track of the previous t+1 time, the position of the t+2 time can be predicted, the network obtained by the position is used as a cooperation network set for jumping switching after switching triggering, and is denoted as CNS_2, and CNS_1 and CNS_2 are collectively called as an alternative network set.
5. The method for handover adaptive handover decision in a super dense heterogeneous wireless network according to claim 4, wherein the handover decision parameters specifically include: when the vehicle-mounted terminal triggers switching, a new network is decided to be accessed to the terminal in the alternative network sets CNS_1 and CNS_2; in the motion process of the vehicle-mounted terminal, the data transmission rate, the network delay and the packet loss rate are key indexes for measuring the performance of the access network, and the 3 parameters are used for evaluating the network performance; because of the influence of network topology and terminal motion state, accessing new target network may cause frequent switching of terminal, defining jump factor, marking the network which is easy to cause frequent switching in the alternative network set as the network which needs to be skipped.
6. The method for determining the handover in a heterogeneous ultra-dense wireless network according to claim 5, wherein the calculation formulas of the data transmission rate, the network delay and the packet loss rate specifically include:
data transmission rate: according to shannon's formula, the data transmission rate of the terminal accessing the network is related to the bandwidth and the signal-to-noise ratio parameter, and the data transmission rate e obtained by the terminal i accessing the network j at the t-th moment ij (t) can be expressed as:
e ij (t)＝B ij (t)×log 2 (1+SNR ij (t)) (11)
wherein B is ij (t) bandwidth resources allocated to terminal i by network j at time t, SNR ij (t) represents a signal-to-noise ratio, the value of which approximates the ratio of RSS to interference noise I in the network;
(2) Network delay: delay and netThe relation between the network loads is set as an exponential function relation; if the attribute time delay of the network j is d' j The time delay d of the terminal i accessing the network j at the t-th moment ij (t) can be expressed as:
Figure FDA0003988496870000041
(3) Packet loss rate: the packet loss rate refers to the ratio of the number of lost data packets to the number of all transmitted data packets in a certain time, and it is assumed that the number of data packets transmitted by the network j at the first t moments is ψ total The number of data packets for which acknowledgement is received is ψ ack The packet loss rate gamma of the terminal i accessing the network j at the t moment ij (t) can be expressed as:
Figure FDA0003988496870000042
(4) Jump factor: a hopping factor delta is defined to mark networks in the set of alternative networks that need to be skipped. After triggering the handover at the t-th moment, the terminal i hops by a factor delta at the j-th alternative network ij (t) can be expressed as:
Figure FDA0003988496870000051
in the method, in the process of the invention,
Figure FDA0003988496870000052
representing the coverage area, cl, of the candidate network j j Representing the chord length of the trajectory of the candidate network j, τ ij Indicating the residence time of terminal i in network j.
7. The method for determining the handover in a heterogeneous ultra-dense wireless network according to claim 6, wherein the interval number multi-attribute decision algorithm for correcting Jaccard similarity in step 103 specifically comprises the following steps:
(1) Constructing an interval number decision matrix: assuming that N networks are concentrated in the network to be evaluated, the number of the network attributes participating in the evaluation is M, acquiring the interval numbers of the M attributes of the N networks before decision, determining the interval number of each network attribute by the maximum and minimum values in multiple data sampling, and respectively acquiring the maximum and minimum values obtained by multiple sampling the kth attribute of the network j as follows
Figure FDA0003988496870000053
And->
Figure FDA0003988496870000054
The number of intervals of the kth attribute of network j can be expressed as + ->
Figure FDA0003988496870000055
The interval number decision matrix to be decided can be expressed as:
Figure FDA0003988496870000056
(2) Normalized attribute interval number: matrix of logarithmic interval numbers
Figure FDA0003988496870000057
Normalizing to obtain normalized matrix +.>
Figure FDA0003988496870000058
Wherein->
Figure FDA0003988496870000059
Formulas (15) and (16) are normalized procedures for benefit-type and cost-type network parameters, respectively:
Figure FDA00039884968700000510
Figure FDA00039884968700000511
(3) Determining interval type ideal scheme: for better measurement of the difference between networks, it is assumed that the interval type ideal scheme of each network attribute is Θ= [ Θ ] 1 ,Θ 2 ,...,Θ M ]Wherein Θ is k Can be expressed as:
Figure FDA0003988496870000061
(4) Calculate the revised Jaccard similarity: the Jaccard similarity is used for describing similarity and difference between sets, the greater the Jaccard similarity is, the higher the similarity of the sets is, and since the number of intervals can also be regarded as a set of numbers, the Jaccard similarity of the normalized attribute value with respect to the ideal solution Θ can be expressed as:
Figure FDA0003988496870000062
Since the similarity between two sections cannot be compared with the similarity between another section when the middle points of the sections are the same, the right end point of the section can be added into the calculation process of Jaccard to correct the sections, and the corrected Jaccard similarity can be expressed as:
Figure FDA0003988496870000063
thus, the modified Jaccard similarity for each network attribute value in the normalized decision matrix corresponding to an ideal solution can be represented by the matrix as ζ= (ζ) jk ) NM ；
(5) Determining the optimal weight of each network attribute: the weight of each network attribute in the judgment process is determined according to the minimum sum of the deviation, and the corresponding optimization model is as follows:
Figure FDA0003988496870000064
(6) Calculating comprehensive similarity: after the weight of each network attribute is obtained, the comprehensive similarity theta of the networks j in the network set to be evaluated can be obtained through weighted summation j ；
Figure FDA0003988496870000065
/>
8. The method for determining handover adaptation in a heterogeneous ultra-dense wireless network according to claim 7, wherein the generating the optimal handover policy specifically comprises:
after the vehicle-mounted terminal triggers the switching, an IKF position prediction model is adopted to generate a candidate network set CNS_1 and a cooperative network set CNS_2 of the terminal in advance, and the comprehensive similarity score of all networks in the two network sets can be calculated through an MJS-INMADM algorithm and recorded as theta 1 And theta 2 By combining the jump factor delta of each network in the alternative network set, an optimal switching strategy can be generated for the terminal, and the generation process is as follows:
selecting network O with highest comprehensive similarity from CNS_1 1 Wherein
Figure FDA0003988496870000071
If network O 1 Jump factor delta of (2) O1 = 1, optimal policy is to switch directly to network O 1 The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, network O with highest comprehensive similarity and jump factor value of 1 is obtained from CNS_2 2 Wherein->
Figure FDA0003988496870000072
The optimal strategy is to switch to network O directly 2 。/>
CN202211573834.2A 2022-12-08 2022-12-08 Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network Pending CN116095770A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211573834.2A CN116095770A (en) 2022-12-08 2022-12-08 Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211573834.2A CN116095770A (en) 2022-12-08 2022-12-08 Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network

Publications (1)

Publication Number Publication Date
CN116095770A true CN116095770A (en) 2023-05-09

Family

ID=86205421

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211573834.2A Pending CN116095770A (en) 2022-12-08 2022-12-08 Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network

Country Status (1)

Country Link
CN (1) CN116095770A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117241328A (en) * 2023-11-15 2023-12-15 四川长虹新网科技有限责任公司 Easymesh switching method
CN117857219A (en) * 2024-03-06 2024-04-09 深圳市永达电子信息股份有限公司 Kalman filtering-based network interference-free strategy control system and method
CN117914768A (en) * 2024-03-19 2024-04-19 中国科学院空天信息创新研究院 Service access path construction system for pass-through node
CN117857219B (en) * 2024-03-06 2024-06-04 深圳市永达电子信息股份有限公司 Kalman filtering-based network interference-free strategy control system and method

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117241328A (en) * 2023-11-15 2023-12-15 四川长虹新网科技有限责任公司 Easymesh switching method
CN117241328B (en) * 2023-11-15 2024-01-23 四川长虹新网科技有限责任公司 Easymesh switching method
CN117857219A (en) * 2024-03-06 2024-04-09 深圳市永达电子信息股份有限公司 Kalman filtering-based network interference-free strategy control system and method
CN117857219B (en) * 2024-03-06 2024-06-04 深圳市永达电子信息股份有限公司 Kalman filtering-based network interference-free strategy control system and method
CN117914768A (en) * 2024-03-19 2024-04-19 中国科学院空天信息创新研究院 Service access path construction system for pass-through node
CN117914768B (en) * 2024-03-19 2024-05-17 中国科学院空天信息创新研究院 Service access path construction system for pass-through node

Similar Documents

Publication Publication Date Title
CN116095770A (en) Cross-region cooperation self-adaptive switching judgment method in ultra-dense heterogeneous wireless network
Choi et al. Adaptive bandwidth reservation and admission control in QoS-sensitive cellular networks
Han et al. Artificial intelligence-based handoff management for dense WLANs: A deep reinforcement learning approach
Alkhawlani et al. Multi-criteria vertical handover by TOPSIS and fuzzy logic
RU2534740C2 (en) Method and device for determining terminal mobility state
CN112672366B (en) Vertical switching system based on personalized consumption preference in heterogeneous wireless network
US20220225127A1 (en) Proactive Mobile Network Optimization
CN111328065B (en) Dynamic programming-based mobility cooperative management method for 5G cloud access network
CN107613532B (en) Fuzzy vertical switching method based on lingering time prediction in vehicle heterogeneous network
Khalaf et al. A comprehensive approach to vertical handoff in heterogeneous wireless networks
CN106102099A (en) A kind of isomery car networking changing method based on residence time
Rajule et al. Survey of vertical handover decision algorithms
Mollel et al. Deep reinforcement learning based handover management for millimeter wave communication
Shi et al. A user mobility prediction method to reduce unnecessary handover for ultra dense network
Padmapriya et al. Handoff decision for multi-user multi-class traffic in MIMO-LTE-A networks
Mehbodniya et al. Wireless network access selection scheme for heterogeneous multimedia traffic
Alablani et al. An SDN/ML-based adaptive cell selection approach for HetNets: a real-world case study in London, UK
Boban et al. Measurement-based evaluation of uplink throughput prediction
CN108882377B (en) Cognitive-based LTE-R resource allocation method
Herculea et al. Straight: Stochastic geometry and user history based mobility estimation
EP3340707A1 (en) Channel selection device and method for wireless local area network
CN115278792B (en) Switching judgment method for sensing terminal access preference in ultra-dense heterogeneous wireless network
CN111988834A (en) Heterogeneous Internet of vehicles network selection method, system and device
CN112672393B (en) Vertical switching method for self-adaptive terminal requirements in ultra-dense heterogeneous wireless network
Ouyang et al. An vertical handover introduced by adaptive mechanism based on fuzzy analytic hierarchy process for heterogeneous networks

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