CN107426703B - Mobility prediction method based on fuzzy clustering in outdoor crowded place - Google Patents
Mobility prediction method based on fuzzy clustering in outdoor crowded place Download PDFInfo
- Publication number
- CN107426703B CN107426703B CN201710733379.0A CN201710733379A CN107426703B CN 107426703 B CN107426703 B CN 107426703B CN 201710733379 A CN201710733379 A CN 201710733379A CN 107426703 B CN107426703 B CN 107426703B
- Authority
- CN
- China
- Prior art keywords
- prediction
- user
- clustering
- track
- users
- 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.)
- Active
Links
Images
Classifications
-
- 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
- H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/06—Testing, supervising or monitoring using simulated traffic
-
- 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
- H04W4/029—Location-based management or tracking services
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a mobility prediction method based on fuzzy clustering in an outdoor crowded place, and belongs to the technical field of mobile communication. The method comprises the steps of dividing a scene into a plurality of prediction areas, dividing prediction time into a plurality of time periods, dividing user movement tracks into different groups in each prediction area according to the prediction time periods, removing irregular tracks by utilizing sequence pattern mining, finding out a frequent movement pattern sequence, and predicting next position information by matching with the user historical movement tracks, so that the prediction accuracy is improved. The invention ensures the continuous communication service of the users in the outdoor crowded area, effectively reduces the unnecessary switching times and increases the detention time of the users. The invention analyzes and models the user movement track of the outdoor crowd-sourced region, and is beneficial to analyzing and solving the user behavior problem of other related outdoor crowd-sourced regions.
Description
Technical Field
The invention belongs to the technical field of mobile communication, and particularly relates to a mobility prediction method based on fuzzy clustering in an outdoor crowded place.
Background
With the explosive increase of data traffic in outdoor crowded areas, the prediction of the movement track and behavior of users can ensure the service quality of the communication network in the areas, and the prediction result can be applied to network deployment planning, resource allocation and mobility management. In an outdoor place with dense people flow, frequent movement of a user can cause the user to switch among different cells, and in order to ensure continuous service, resource allocation of a network is optimized through accurate prediction of a user track and smooth switching is ensured.
In outdoor crowded places, due to the huge number of users and the complex and varied behavior characteristics of users, how to quickly predict the next position of the user must be considered. Meanwhile, the behavior characteristics of the user will present different characteristics in different time and space, and how to comprehensively consider the factors influencing the motion trajectory must also be solved.
In the prior art, the next position of a user is predicted according to the movement characteristics of a single user, and the prediction precision of a new user is reduced; in the prediction scheme according to the multi-user movement characteristics, the influence factor of irregular movement tracks and the influence of space-time two dimensions are not considered, and the prediction precision is also reduced.
In the prediction scheme for predicting the next position of the user according to the moving track of the single user, for example, reference [1] (reference [1 ]: x.zhu, m.li, w.xia, and h.zhu, "a novel handoff algorithm-m for handover networks," China Communications, vol.13, No.8, pp.136-147, and Aug 2016), a handover algorithm based on speed and service perception is proposed, and the speed of the user is predicted by using a markov prediction model, so as to determine the next position area of the user to select a handover base station. But this mechanism only considers the current speed and direction of a single user, which will result in a reduction in prediction accuracy for an irregularly moving user.
In The prediction scheme for predicting The next position of a single user according to The moving tracks of multiple users, such as reference [2] (reference [2 ]: j. jeong, m.leonte, and a. property, "Cluster-aided location predictions," in IEEE info com 2016-The 35th Annual international conference Computer Communications, April2016, pp.1-9.), a method for predicting The next position area of a new user by using The tracks of all The mobile users is proposed, and The moving tracks of all The users are classified by a clustering algorithm and then similar moving tracks are extracted to improve The prediction accuracy. However, in the multi-user movement trajectory, the irregular user movement trajectory will reduce the overall prediction accuracy.
Under the condition that switching frequently occurs, the switching process can be optimized by utilizing the mobility prediction technology, and unnecessary switching times are reduced. However, in an outdoor crowded environment, the huge number of users and the variable user behaviors make accurate prediction more difficult, and the existing prediction model is no longer suitable for the mobility prediction in the scene.
Disclosure of Invention
Aiming at mobility prediction in outdoor crowded places, the invention provides a mobility prediction method based on fuzzy clustering in the outdoor crowded places in order to solve the problem that the existing prediction model is not applicable. The method of the invention adopts a clustering algorithm to divide the prediction area, divides the prediction time according to the characteristics of the research scene, clusters the movement track mode of the user in each prediction area, eliminates irregular movement tracks and finds out frequent sequences based on a sequence mode mining technology, thereby improving the movement prediction precision.
The invention provides a mobility prediction method based on fuzzy clustering in an outdoor crowded place, which comprises the following steps:
step 1, an initialization phase, comprising: dividing the whole area into a plurality of prediction areas; dividing time into a plurality of prediction time slots according to the characteristics of a researched scene, grouping user tracks contained in each time slot segment in each prediction area, and dividing the user tracks into different movement mode groups; and mining a frequent movement pattern sequence in each divided group of user tracks by using the sequence pattern.
Step 3, a switching execution stage, comprising: and determining a Base Station (BS) connected with the next position according to the predicted position in the prediction stage, and switching the optimal target BS.
The invention has the advantages that:
1. the mobility prediction scheme can realize accurate prediction of the user movement track in the outdoor crowded area, ensures the continuous communication service of the users in the outdoor crowded area, and decides the optimal target switching base station by predicting the next position area of the users, thereby ensuring the good continuous communication service of the users. According to the simulation result, the switching optimization is carried out based on the proposed prediction scheme, so that unnecessary switching times are effectively reduced, and the residence time of a user is increased.
2. The invention comprehensively considers the factors of two dimensions of the prediction time and the prediction space, realizes the detailed division of the user movement characteristics of the predicted area, and can more accurately predict the next position area of the user. The invention analyzes and models the user movement track of the outdoor crowd-sourced region, and is beneficial to analyzing and solving the user behavior problem of other related outdoor crowd-sourced regions.
Drawings
FIG. 1 is a mobility prediction scenario for an outdoor crowded place of the present invention;
FIG. 2 is an algorithmic flow chart of the movement prediction scheme of the present invention;
FIG. 3 illustrates an average number of handovers per prediction period for a user in accordance with an embodiment of the present invention;
FIG. 4 is a diagram illustrating the number of times users switch between different prediction regions according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating average user retention times in different prediction regions according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The invention provides a prediction scheme aiming at outdoor crowded places for researching the application of mobility prediction in outdoor scenes, and the next position area of a user is predicted by fully considering the factors of two dimensions of time and space, so that the fluency of switching of user equipment between outdoor communication cells is optimized, and the continuous service of the user is ensured.
In the outdoor people flow dense area, the invention introduces the mobility prediction technology to optimize the switching process and ensure the continuous communication service of the user.
As shown in fig. 1, taking a campus scenario as an example, the present invention divides the whole campus into a plurality of prediction areas according to people flow aggregation, divides the prediction time into a plurality of time periods according to the class work and rest time, divides the user movement trajectories into different groups in each prediction area according to the prediction time periods, eliminates irregular trajectories by using sequence pattern mining, finds out frequent movement pattern sequences, and predicts next location information by matching with the user historical movement trajectories, thereby improving the prediction accuracy. The mobility prediction method based on fuzzy clustering in the outdoor crowded place has the overall flow as shown in FIG. 2, the method divides a new user into the belonged prediction region according to the predicted time and the position of the new user, matches the most suitable movement track and predicts the next position region, and decides the optimal switching target base station according to the prediction result. The steps are explained below.
Step 1, an initialization stage, which mainly comprises two parts, wherein the first part is used for carrying out prediction area division, and the second part is used for grouping user tracks in each prediction area.
Firstly, by analyzing a mobile user gathering area, dividing the whole area into a plurality of prediction areas, grouping user tracks contained in each area, dividing time into a plurality of prediction time slots according to the characteristics of a researched scene, and predicting according to the movement characteristics of users in each time slot.
The method classifies the researched area into a plurality of prediction areas according to the distribution of the people stream based on a Kmeans clustering algorithm. The process of prediction region division includes step 1.1 to step 1.4.
Step 1.1, selecting the central position of the dense people flow area as a clustering center, and extracting position coordinate information from historical movement tracks of all users as a clustering data set.
Setting initialized cluster centersWherein u isj=(xj,yj) And expressing the position coordinates of the jth clustering center, and expressing the number of the selected clustering centers by K. Let the extracted set of user position coordinates be expressed asWherein s isi=(xi,yi) The ith user position in the set is represented, and N represents the number of positions.
And the number K of the clustering centers is determined according to the research scene and is associated with the distribution of the user in the research scene. Since the number of clusters in the prediction region partition is small, traversal from 2 to 10 can be adopted to determine the optimal number of cluster centers. And initially selecting the center position of the crowd dense area as a clustering center.
And 1.2, dividing all the positions according to the distance from the position of each user to the clustering center. Each cluster center corresponds to a partition region. To the collectionThe distances from the position to all the cluster centers are calculated, and the position is divided into the area corresponding to the cluster center with the minimum distance.
User location siDistance cluster center ujIs d(s)i,uj) As follows:
finding the user position siThe minimum distance from each cluster center is set as the distance from the cluster centerMore recently, the following:
cican represent the area c to which the user belongsi。
Step 1.3: update the cluster centers as follows:
in the formula (3), the jth cluster center ujUpdating according to the position divided into the jth area, wherein the numerator represents the sum of all position coordinates belonging to the jth area, and the denominator represents the number of the position coordinates in the jth area.
Step 1.4: the objective function is solved as follows:
wherein, J(c,u)And the sum of the distances between each user position and the clustering center of each belonging area is represented. Wherein the distance isCan be calculated according to the formula (1).
Step 1.5, judging whether the target function converges to the minimum value, if so, finishing iteration, finishing prediction region division, and outputting each clustering center and the position in each region; otherwise, the step 1.2 is returned to and executed.
And in each iteration, updating and recording the minimum value of the objective function, and when the objective function value is not changed during the iteration, representing that the minimum value is converged and ending the iteration.
And (3) grouping the user tracks in each prediction area by using an FCM (Fuzzy C-means) algorithm, wherein in each prediction position area, the user moving tracks contained in each time period are extracted and classified, and the mobile users with similar tracks are classified into corresponding groups. The specific steps comprise step 2.1-step 2.2.
Step 2.1: for a certain prediction region, extracting the track sequence of L users contained in a set time period in the prediction region
Wherein P isl=<pl1,pl2,…,…plw,…,plW>For the track of user l, W represents the total number of positions in the track of user l, and setThe number of the clustering centers of the tracks of the L users is T, and the total number of the clustering centers is T. The track of one user only belongs to one track cluster center in one time period. And (4) clustering the central position of the initial track, and selecting the central position of the crowd dense area.
Step 2.2: and solving the membership degree of each user and the track clustering center according to the target function.
The objective function J is:
wherein, | | | represents solving Euclidean distance, m represents membership factor, etlRepresenting the user l as belonging to the trajectory cluster center vtDegree of membership.
Degree of membership etlCalculated according to the following formula:
step 2.3: updating the central position of each track cluster, wherein the formula is as follows:
step 2.4: judging whether the target function converges to the minimum value, if not, continuing to execute the step 2.2, calculating the target function according to the formulas (5) and (6), and updating the track clustering center according to the formula (7); if so, namely when the target function is continuously reduced and the final convergence is not changed, namely the target function converges to the minimum value, the iteration is stopped.
And after the iteration is finished, outputting T track clustering centers and the user tracks of all the track clustering centers. Thus, the study area is divided into several prediction areas and the user movement trajectories of each area are classified according to the prediction time.
And then mining the movement tracks of all groups of users in each prediction area by using a sequence mode to find out the frequent movement track mode of the users in the same group. And clustering the movement tracks of the users in each prediction area, and mining by using a sequence mode to find out frequent movement tracks so as to eliminate irregular movement tracks.
And 2, a prediction stage.
Firstly, adding the current position of a user into a position observation sequence, and selecting a prediction area corresponding to the current time period of the user according to the historical movement position track of the user. The user can be divided into corresponding prediction areas according to the position coordinates where the user is currently located.
The users are then classified into corresponding groups of similar movement patterns within the predicted area. And selecting the frequent movement pattern sequence with the highest matching degree as a prediction sequence by matching the historical movement track of the user with the frequent movement pattern sequence in the group.
Dividing users into corresponding similar moving mode groups in a time slot in a prediction area by calculating the membership e of the users to each track clustering center in the prediction areatlTo determine the group to which the user belongs.
And finally, according to and predicting the next position information of the user. The next position information is predicted by matching with the historical movement track of the user, so that the accuracy of prediction is improved.
Step 3, switching execution stage.
And determining the base station BS connected with the next position according to the predicted position obtained in the prediction stage. As shown in FIG. 2, let the currently connected base station be labeled as BSqAnd the decided base station to be connected is marked as BSp. When the BS is decidedpAnd the current connection BSqAre identical to each otherWhen the BS is decided, the switching operation is not performed, and the next position prediction is performed in the prediction stage, if the BS is decidedpAnd the current connection BSqDifferent, handover to target BSpAnd returning to the prediction stage to continue predicting the next position.
Examples
As shown in fig. 1, in a campus dense traffic field, a Small-cell Base Station (SBS) is deployed in the area, and the coverage area is 100 m. The mobile user trajectory is extracted from the user data in reference [3] (reference 3: i.rhee, m.shin, s.hong, k.lee, s.kim, and s.chong, "Crawdad dataset ncsu/mobilitymodels," downloaded from http:// Crawdad. org/Crawdad/ncsu/mobilitymodels/20090723/, Jul2009.) and the position coordinates of the user are recorded every 30 seconds. And dividing the user tracks in the scene into 6 groups according to the comparison of the simulation performances, and counting the switching times of the users in each area and the residence time of the users in each cell range.
To demonstrate the performance of the mobility prediction optimized handover scheme (MPSDM) proposed herein, two handover mechanisms were chosen for comparison.
Mechanism 1 (MP-IUM): the scheme is based on the single-user movement characteristics and carries out the prediction of the next position information of the user.
Mechanism 2 (MP-MUM): according to the scheme, the next position information of the user is predicted based on the multi-user movement characteristics, and the influence of the movement characteristics of the user changing along with time and the motion trail of an irregular moving user on the prediction precision is not considered.
As shown in fig. 3, is the average number of handovers per prediction period for all users. Wherein, the abscissa represents 9 the prediction time period, and the ordinate represents the switching times, it can be seen from the figure that the switching times of the method of the present invention in each time period are less than those of the other two comparison schemes. At a time period t1,t4And t7It can be seen that the number of switching times is greater than those in other periods because the three periods are the time to go to class and eat, and the increase in the number of switching times is caused by the large movement of the stream of people. As shown in fig. 4, the number of times the user switches in different predicted areas is shown. WhereinThe abscissa represents the area region and the ordinate represents the average switching times as can be seen from the figure, the switching times of the method of the present invention in each region are all smaller than the comparison scheme. The MP-IUM handover frequency is the largest because the MP-IUM is based on the moving track of a single user, and when the irregular movement of the user is predicted to increase, the prediction accuracy will decrease, and the decided handover base station is not optimal, which will result in frequent handover. The MP-MUM classifies similar movement tracks through the movement tracks of multiple users, and carries out corresponding movement group matching on new users to predict the next position. The method divides the prediction area, more accurately predicts the concentrated area, and divides the prediction time according to the characteristics of the research scene, thereby integrating the movement characteristics of the user in each position area and time period, eliminating irregular movement tracks, improving the prediction precision and reducing unnecessary switching times.
As shown in fig. 5, the average residence time of the user in different prediction regions is shown. Where the abscissa represents the prediction region and the ordinate represents the mean residence time it can be seen from the figure that the residence time of the method of the invention is greater than for the other two comparison schemes. Because the prediction accuracy of the comparison scheme on the next position of the user is reduced, the corresponding switching times are increased, so that the residence time of the user in the same cell is reduced. The method improves the accuracy of prediction and decides the optimal target base station by comprehensively considering the influence factors of space-time two dimensions, thereby reducing unnecessary switching times and improving the residence time of users in the same cell.
Claims (4)
1. A mobility prediction method based on fuzzy clustering in an outdoor crowded place is characterized by comprising the following steps:
step 1, an initialization phase, comprising: dividing the whole area into a plurality of prediction areas; dividing time into a plurality of prediction time slots according to the characteristics of the studied scene, grouping user tracks in each time slot in each prediction area, and dividing the user tracks into different movement mode groups; mining a frequent movement pattern sequence in each group of divided user tracks by using a sequence pattern;
in the step 1, the prediction region is divided based on a Kmeans clustering algorithm, and the method specifically comprises the following steps:
step 1.1, selecting K clustering centers, and extracting position coordinates from historical movement tracks of all users as a clustering data set; k is a positive integer;
step 1.2, for each position in the clustering data set, dividing the position into an area corresponding to the clustering center with the minimum distance according to the distance from the position to each clustering center;
step 1.3, updating the position of each clustering center according to the position divided into each region;
step 1.4, solving an objective function, wherein the objective function is the sum of the distances between each user position and the clustering center of each region to which the user position belongs;
step 1.5, judging whether the target function converges to the minimum value, if so, finishing iteration, finishing prediction region division, and outputting user position coordinates of each clustering center and each prediction region; otherwise, the step 1.2 is executed;
in step 1, grouping the user trajectories in each time slot in each prediction region by using a fuzzy C-means algorithm, which specifically includes:
step 2.1, setting a prediction area, extracting the track sequence of L users contained in a time slot in the prediction area, and setting T track clustering centers of the L users, wherein the T track clustering centers are expressed as { v1,v2,…,vt,…,vT}; l, T is a positive integer;
step 2.2, calculating an objective function J of each user about each track clustering center as follows:
wherein, PlRepresents the track of the user l, | | | represents the Euclidean distance, m represents the membership degreeA factor; e.g. of the typetlRepresenting the user l as belonging to the trajectory cluster center vtThe calculation formula of the membership degree of (c) is as follows:
step 2.3, updating the central position of each track cluster, wherein the formula is as follows:
step 2.4, judging whether the target function converges to the minimum value, if not, continuing to execute the step 2.2; otherwise, stopping iteration, and outputting T track clustering centers and the user tracks of all the track clustering centers;
step 2, a prediction stage, comprising: adding the current position of the user into a position observation sequence, and selecting a prediction area corresponding to the current time period of the user according to the historical movement track of the user; dividing the user into similar movement mode groups in corresponding prediction areas, and selecting a frequent movement mode sequence with the highest matching degree as a prediction sequence by matching the historical movement track of the user with the frequent movement mode sequence in the groups for predicting the next position of the user;
step 3, the switching execution stage comprises: and according to the predicted position in the prediction stage, deciding the base station connected with the next position, and switching the optimal target base station.
2. The method of claim 1, wherein K is in the range of [2,10], and the optimal number of cluster centers is determined by traversing values in the range.
3. The method according to claim 1, wherein in the step 2, the users are classified into groups of similar moving patterns in the corresponding prediction areas, and the groups to which the users belong are determined by calculating the membership of the users with respect to the clustering centers of the tracks in the prediction areas.
4. The method according to claim 1, wherein in the step 3, the performing the optimal target base station handover specifically comprises: when the decided base station is the same as the base station connected currently, switching operation is not carried out, and then the next position is predicted in a prediction stage; and when the decided base station is different from the base station connected currently, switching to the decided base station, and then returning to the prediction stage to predict the next position.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710733379.0A CN107426703B (en) | 2017-08-24 | 2017-08-24 | Mobility prediction method based on fuzzy clustering in outdoor crowded place |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710733379.0A CN107426703B (en) | 2017-08-24 | 2017-08-24 | Mobility prediction method based on fuzzy clustering in outdoor crowded place |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107426703A CN107426703A (en) | 2017-12-01 |
CN107426703B true CN107426703B (en) | 2020-04-03 |
Family
ID=60434937
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710733379.0A Active CN107426703B (en) | 2017-08-24 | 2017-08-24 | Mobility prediction method based on fuzzy clustering in outdoor crowded place |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107426703B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110021161B (en) * | 2018-01-09 | 2021-12-21 | 株式会社日立制作所 | Traffic flow direction prediction method and system |
CN108416293A (en) * | 2018-03-08 | 2018-08-17 | 中国科学院大学 | A kind of specific group's Assembling Behavior early detection and aggregation ground prediction technique and system |
CN109041217B (en) * | 2018-09-21 | 2020-01-10 | 北京邮电大学 | Hierarchical mobility prediction method in heterogeneous network |
CN109635059A (en) * | 2018-11-23 | 2019-04-16 | 武汉烽火众智数字技术有限责任公司 | People's vehicle association analysis method and system based on track similarity mode |
CN111148180A (en) * | 2020-01-02 | 2020-05-12 | 维沃移动通信有限公司 | Method for connecting base station and terminal equipment |
CN111372190B (en) * | 2020-01-23 | 2021-11-02 | 鹏城实验室 | Machine learning model, method, device and storage medium for mobile terminal position tracking |
CN112629550B (en) * | 2020-10-13 | 2024-03-01 | 北京三快在线科技有限公司 | Method and device for predicting obstacle track and model training |
CN112600728B (en) * | 2020-12-07 | 2022-06-14 | 南昌交通学院 | 5G mobile base station flow prediction analysis system based on big data |
CN113065690B (en) * | 2021-03-19 | 2023-06-30 | 浙江财经大学 | Traffic prediction method and device |
CN113932282B (en) * | 2021-11-11 | 2024-03-01 | 武汉中绘炫图信息科技有限公司 | Urban heating supervision method and system |
CN116842285B (en) * | 2023-07-27 | 2024-05-03 | 中国人民解放军陆军工程大学 | Target accompanying mode mining method based on space-time track data |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103444163A (en) * | 2011-02-05 | 2013-12-11 | 苹果公司 | Method and apparatus for mobile location determination |
CN104039009A (en) * | 2014-06-23 | 2014-09-10 | 西北大学 | Target wireless device positioning method based on regional division |
CN105635968A (en) * | 2016-01-18 | 2016-06-01 | 广州杰赛科技股份有限公司 | Hotspot area identification method based on time unit and predication method and device |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150172216A1 (en) * | 2013-12-18 | 2015-06-18 | International Business Machines Corporation | Determining rules for partitioning internet connection bandwidth |
-
2017
- 2017-08-24 CN CN201710733379.0A patent/CN107426703B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103444163A (en) * | 2011-02-05 | 2013-12-11 | 苹果公司 | Method and apparatus for mobile location determination |
CN104039009A (en) * | 2014-06-23 | 2014-09-10 | 西北大学 | Target wireless device positioning method based on regional division |
CN105635968A (en) * | 2016-01-18 | 2016-06-01 | 广州杰赛科技股份有限公司 | Hotspot area identification method based on time unit and predication method and device |
Non-Patent Citations (2)
Title |
---|
Lijun Ma;Kangqing Lin,Meijiao Guan,Meiyan Lin.The prediction of carbon emission in all provinces of China with the K-means cluster based Logistic model.《2017 International Conference on Service Systems and Service Management》.2017, * |
参与式感知系统中基于社会关系的移动用户位置预测算法;于瑞云,夏兴有,李婕,周岩,王兴伟;《计算机学报》;20150228;第38卷(第2期);374-385 * |
Also Published As
Publication number | Publication date |
---|---|
CN107426703A (en) | 2017-12-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107426703B (en) | Mobility prediction method based on fuzzy clustering in outdoor crowded place | |
Liu et al. | Mobility modeling, location tracking, and trajectory prediction in wireless ATM networks | |
US11109283B1 (en) | Handover success rate prediction and management using machine learning for 5G networks | |
CN107046700B (en) | Method and device for predicting base station switching of mobile terminal | |
CN108900980B (en) | Resource allocation optimization method based on mobility prediction in heterogeneous network | |
Chen et al. | Predicting a user's next cell with supervised learning based on channel states | |
Lin et al. | Anticipatory mobility management by big data analytics for ultra-low latency mobile networking | |
Tung et al. | Big data and machine learning driven handover management and forecasting | |
CN110149595B (en) | HMM-based heterogeneous network user behavior prediction method | |
CN103810270B (en) | Optimized tracking area partition method and device | |
Mehbodniya et al. | A fuzzy MADM ranking approach for vertical mobility in next generation hybrid networks | |
Prado et al. | ECHO: Enhanced conditional handover boosted by trajectory prediction | |
Parvaresh et al. | Deep Q-learning-enabled deployment of aerial base stations in the presence of mobile users | |
Li et al. | Mobile user location prediction based on user classification and markov model | |
CN107466012A (en) | Network switching optimization method based on mobility prediction under home network indoors | |
Stoyanova et al. | Algorithmic approaches for vertical handoff in heterogeneous wireless environment | |
Fazio et al. | A new Markov-based mobility prediction scheme for wireless networks with mobile hosts | |
Nyalapelli et al. | Recent Advancements in Applications of Artificial Intelligence and Machine Learning for 5G Technology: A Review | |
CN109673015B (en) | Tracking area planning method based on spectral clustering | |
Wong et al. | An adaptive distance-based location update algorithm for PCS networks | |
Kaleem et al. | Application of fuzzy TOPSIS for weighting the system attributes in overlay networks | |
Zhohov et al. | Reducing latency: Improving handover procedure using machine learning | |
Taheri et al. | Clustering techniques for dynamic location management in mobile computing | |
Chen et al. | Support vector machine based mobility prediction scheme in heterogeneous wireless networks | |
Halgamuge | Performance evaluation and enhancement of mobile and sensor networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |