JP2018142957A - Management device, program for making computer execute, and computer-readable recording medium recording program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a management device capable of determining a communication parameter for a secondary user to carry out radio communication while avoiding interference in real time with high accuracy.SOLUTION: A machine learning unit 102 generates teacher data from communication history of secondary users and history of interference notice, reduces deviation of the number of pieces of interference data and the number of pieces of non-interference data, classifies communication parameters into a class without interference and a class with interference in an n-dimensional space made up of n parameters on the basis of updated teacher data mutually associating presence or absence of interference, positions of terminals suffering interference, and the communication parameters, and determines a determination boundary having equal distance from two classes by machine learning. A parameter determination unit 103 searches for a point where a gain function becomes the maximum using distance from the determination boundary as an index on a (n-m) dimensional hyperplane where unchangeable m parameters are fixed of the n parameters, and determines a set of communication parameters not to interfere with a primary user.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a management apparatus, a program to be executed by a computer, and a computer-readable recording medium on which the program is recorded.

  In a cognitive radio system or a frequency sharing mobile communication system, in the frequency band where the primary user, which is a licensed radio communication system, exists, the communication band is secured by utilizing the space in time and space. , Improve communication capacity.

  However, since a wireless station that communicates using the primary user's frequency uses the primary user's frequency as the secondary user, there is a restriction that it does not interfere with the primary user's communication. The Therefore, the secondary user needs to set transmission power etc. appropriately.

  Interference with the primary user is affected not only by the frequency channel, transmission power and modulation method of the secondary user, but also by the positional relationship between the primary user and the secondary user, surrounding buildings, plants and weather, etc. Receive. For this reason, it is difficult to accurately model the interference amount and derive a theoretically appropriate parameter. Moreover, since the positional relationship, weather, and the like change with time, it is necessary to set appropriate parameters in real time so as not to interfere with the primary user.

  Conventionally, the following techniques are known as techniques for setting parameters of secondary users.

  A number of sensors are used to detect signals transmitted by the primary user, a database of frequency availability at a certain point and time, and the parameters of the secondary user are set using the database (non-patent) Reference 1).

  The secondary user performs a reception operation before signal transmission, and performs transmission with confidence that the primary user is not using the frequency channel (Non-Patent Document 2). That is, the secondary user performs carrier sense, and performs transmission when wireless communication is not performed as a result of carrier sense.

  The secondary user calculates the expected value of the amount of interference to the primary user from the received signal power from the primary user, and dynamically adjusts the transmission power by PID control to avoid interference (non-patent) Reference 3).

  The secondary user obtains an estimated value of the amount of interference with the primary user from the propagation model, and controls the transmission power so that the estimated value is equal to or less than a predetermined value (Non-Patent Document 4).

JP 2013-212609 A

FCC, “Second report and order and memorandum opinion and order,” Docket no. 08-260, Nov. 2008. A. Tsertou and D.I. Laurenson, “Revisiting the Hidden Terminal Problem in a CSMA / CA Wireless Network,” IEEE Trans. Mob. Comput., Vol. 7, no. 7. Pp 817-831, Jul. 2008. Motoki Matsui, et al. , “PID-controlled power adjustment algorithm considering radio interference in cognitive radio networks (wireless),” IEICE technical report. NS, Network System 110.448 (2011): 217-222. Takeo Fujii, “Multi-channel multi-hop cognitive wireless network that avoids interference by power limitation,” IEICE Transactions B91.11 (2008): 1369-1379. H. He and E. Garcia, “Learning from Imbalanced Data,“ IEEE Trans. Knowl. Data Eng., Vol. 21, no. 9, pp1263-1284, 2009. Zhenyu Wang, E.K.Tameh, and A.R.Nix, “Joint Shadowing Process in Urban Peer-to-Peer Radio Channels,” IEEE Trans. Veh., Vol. 57, no. 1, pp. 52-64, Jan. 2008. K. Karmarkar, "A New Polynomial-Time Algorithm for Linear Programming," Combinatorica, 4 (1984), 373-395. Glover, Fred (1986). "Future Paths for Integer Programming and Links to Artificial Intelligence". Computers and Operations Research 13 (5): 533-549.doi: 10.1016 / 0305-0548 (86) 90048-1. Cox, DR (1958). "The regression analysis of binary sequences (with discussion)". J Roy Stat Soc B 20: 215-242. Koby Crammer, Alex Kulesza and Mark Dredze, “Adaptive Regularization of Weight Vectors,” NIPS2009.

  However, in the technique described in Non-Patent Document 1, when the primary user is only receiving, it is not possible to consider interference at the reception point. Further, if the transmission power is limited, it is difficult to find the frequency channel even if there is a usable frequency channel. Furthermore, although radio wave propagation fluctuates due to the influence of surrounding buildings, it is difficult for a database to follow such real-time changes.

  Further, in the technique described in Non-Patent Document 2, as in the technique described in Non-Patent Document 1, when the primary user is only receiving, it is not possible to consider interference at the reception point. In addition, if the primary user and the secondary user's transmission terminal are separated and the primary user's receiving terminal is close to the secondary user, interference occurs at the primary user's receiving terminal. Occurs. This is a hidden terminal problem and is an unavoidable problem in avoiding interference by carrier sense.

  Furthermore, in the technique described in Non-Patent Document 3, it is assumed that the primary user is a transmission / reception terminal and the secondary user can receive the signal. However, as described in the technique in Non-Patent Document 2 as well. Regarding interference with the primary user's receiving terminal, it is impossible to avoid interference because a signal from the primary user's receiving terminal cannot be received.

  Furthermore, in the technique described in Non-Patent Document 4, various parameters relating to the primary user, specifically, propagation attenuation constant, transmission power, maximum communication distance, desired signal-to-noise ratio, and the like are acquired in advance. There is a need. When there are a large number of primary users, it is difficult to obtain these parameters for all primary users.

  According to the embodiment of the present invention, a management device is provided that allows a secondary user to accurately determine communication parameters for performing wireless communication while avoiding interference in real time.

  In addition, according to the embodiment of the present invention, there is provided a program for causing a computer to execute, in real time, accurately determining communication parameters for a secondary user to perform wireless communication while avoiding interference.

  Furthermore, according to the embodiment of the present invention, a computer having recorded thereon a program for causing a computer to execute a real time accurate determination of communication parameters for a secondary user to avoid interference and perform wireless communication A readable recording medium is provided.

(Configuration 1)
According to the embodiment of the present invention, the management device includes receiving means, generating means, updating means, machine learning means, communication parameter determining means, and transmitting means. The receiving means receives an interference notification including the interfered terminal position, which is the position of the interfered terminal, and the presence / absence of interference from a primary user who is a wireless communication system in which the ratio of data indicating interference is a constant value. To do. The generation means includes a label indicating the presence / absence of interference based on a communication history of a secondary user, which is a wireless communication system that performs wireless communication using the frequency band of the primary user, and an interference notification history, and a position of the interfered terminal And teacher data in which communication parameters are associated with each other. The updating means includes the number of interference data in which a label indicating that there is interference, a interfered terminal position, and a communication parameter are associated with each other, a label indicating that there is no interference, the interfered terminal position, and a communication parameter. A bias reduction process for reducing a bias with the number of non-interference data associated with each other is applied to the teacher data, and updated teacher data is generated by updating the teacher data. The machine learning means, based on the update teacher data, in the n-dimensional space composed of n parameters (n is an integer of 2 or more) including the communication parameters, the communication class is the second class that has interference with the first class that has no interference. A decision boundary, which is a boundary for classifying the first class and the second class and is present at an equal distance from the first and second classes, is determined by machine learning. The communication parameter determination means determines at least one of a communication parameter set in which interference does not occur in the primary user and a communication parameter set in which interference occurs in the primary user, using the determination boundary. The transmission means transmits the determined communication parameter set to the secondary user's terminal group.

  The management apparatus according to the embodiment of the present invention generates update teacher data in which the deviation between the number of interference data and the number of non-interference data is reduced, and performs machine learning based on the generated update teacher data to determine a decision boundary. And determining at least one of a communication parameter set in which interference does not occur in the primary user and a communication parameter set in which interference occurs in the primary user, using the determined decision boundary. As a result, the secondary user can determine whether or not interference occurs based on the determined communication parameter.

  Therefore, it is possible to accurately determine communication parameters for the secondary user to avoid interference and perform wireless communication in real time.

(Configuration 2)
In the configuration 1, the communication parameter determination means sets a (n−m) -dimensional hyperplane in which m parameters (m is an integer satisfying 1 ≦ m <n) that are not changeable among the n parameters are set, A point on the (nm) dimensional hyperplane that maximizes or minimizes the gain function is searched using the distance between the set point on the (nm) dimensional hyperplane and the decision boundary as an index, and the search is performed ( A communication parameter set in which interference does not occur in the primary user is determined from the points on the (n−m) -dimensional hyperplane using constraints.

  As described above, the management apparatus determines a communication parameter set of the secondary user that does not interfere with the primary user.

  Therefore, it is possible to accurately determine communication parameters for the secondary user to avoid interference and perform wireless communication in real time.

(Configuration 3)
In Configuration 1 or Configuration 2, the update unit updates the teacher data by increasing the number of interference data and decreasing the number of non-interference data to generate updated teacher data.

  By increasing the number of interference data and decreasing the number of non-interference data, the deviation between the number of interference data and the number of non-interference data is reduced, so the number of non-interference data is larger than the number of interference data. In a wireless communication environment, a communication parameter set of a secondary user that does not interfere with the primary user can be accurately determined in real time.

(Configuration 4)
In Configuration 3, the update unit generates pseudo interference data from the interference data included in the teacher data, and deletes non-interference data included in the teacher data to generate updated teacher data.

  Since the pseudo interference data is generated from the interference data included in the teacher data, the pseudo data can be generated at a point where interference can occur due to the correlation of shadowing in wireless communication. As a result, the communication parameter set of the secondary user that does not interfere with the primary user can be determined with high accuracy.

(Configuration 5)
In the configuration 4, the updating unit uses the interference data included in the teacher data as a reference point, and the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a predetermined distribution. In addition to generating data, update teacher data is generated by deleting non-interference data existing outside the circle centered on the reference point.

  Generates pseudo data so that it follows the same distribution as the interference data that is the reference point, and deletes non-interference data that exists at a certain distance from the interference data, so that pseudo data is generated where interference can occur. In addition, non-interference data that is not considered is deleted.

  Therefore, it is possible to accurately determine the communication parameter set of the secondary user that does not interfere with the primary user. In addition, the amount of calculation can be reduced.

(Configuration 6)
In Configuration 5, the updating unit generates the pseudo interference data so that the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a normal distribution when the variance is changed.

  Pseudo interference data is generated by statistical processing.

  Therefore, pseudo data can be easily generated at a point where interference can occur.

(Configuration 7)
In any one of Configurations 1 to 6, the management device further includes first and second risk level assigning means. The first risk level assigning unit estimates the received power in consideration of the terrain data when the coordinate point of the update teacher data received from the update unit is the transmission point and the primary user's position is the reception point. The first interference occurrence risk is estimated based on the estimated received power, and the estimated first interference occurrence risk is added to the updated teacher data. Upon receiving the decision boundary determined based on the updated teacher data to which the first interference occurrence risk level is given, the second risk level assigning means divides the update teacher data existing area into a plurality of areas, Estimate the received power when the coordinate point of one divided area is the transmission point and the position of the primary user is the reception point, and estimate the second risk of occurrence of interference based on the estimated received power Then, the process of giving the estimated second risk of occurrence of interference to one area and generating the determination data is executed for all of the plurality of areas. Then, the machine learning means determines a determination boundary by machine learning based on the updated teacher data to which the first interference occurrence risk is assigned. The communication parameter determination means includes a determination boundary determined based on the updated teacher data to which the first interference occurrence risk is assigned, determination data, and a plurality of areas to which the second interference occurrence risk is assigned. Based on the above, it is determined whether each of the plurality of areas is a communication prohibited area, and at least one of the communication prohibited area and the communication permitted area is determined as a communication parameter set.

  Since the decision boundary is determined using the first and second interference occurrence risks determined in consideration of the terrain data, the interference data and the non-interference data are not compared with the case where the first and second interference occurrence risks are not used. Interference data can be easily separated.

(Configuration 8)
In Configuration 7, the communication parameter determination unit compares the determination data with the determination boundary, and determines, as a communication prohibited area, an area corresponding to the determination data classified as the area that causes interference among the plurality of areas.

  Since the relative position between each area of the plurality of areas and the determination boundary is determined via the determination data, it is possible to accurately determine whether each area is a communication prohibited area.

(Configuration 9)
In Configuration 7 or Configuration 8, the first risk level assigning unit estimates the estimated received power as the first interference occurrence risk level, or the estimated received power is the interference data closest to the point where the received power is obtained. The division result obtained by dividing by the received power at the position is estimated as the first interference occurrence risk.

  The estimated received power or the result of dividing the estimated received power by the received power at the position of the interference data closest to the point where the received power is obtained is set as the first interference occurrence risk level.

  Accordingly, the first interference occurrence risk can be estimated based on the terrain data.

(Configuration 10)
In any one of Configurations 7 to 9, the first risk assigning unit holds a plurality of radio wave propagation characteristics as radio wave propagation characteristics between the primary user and the secondary user, and The most suitable radio wave propagation characteristic is selected from a plurality of radio wave propagation characteristics, and an estimation process for estimating received power using the selected radio wave propagation characteristics is executed.

  Accordingly, the received power can be estimated in conformity with the terrain data.

(Configuration 11)
In the configuration 10, the second risk level assigning unit performs the same estimation process as the estimation process in the first risk level giving unit, and estimates the received power.

  Therefore, it is possible to estimate the received power that is the basis of the second risk of occurrence of interference in accordance with the terrain data.

(Configuration 12)
In any one of Configurations 7 to 11, the terrain data includes an elevation map or an aerial photograph.

  The first and second interference occurrence risks can be estimated according to the radio wave propagation environment.

(Configuration 13)
According to the embodiment of the present invention, the program to be executed by the computer is a program for causing the computer to execute determination of a communication parameter set for performing wireless communication, wherein the receiving means is a data indicating interference. A first step of receiving an interference notification including a location of an interfered terminal, which is a location of a terminal that has undergone interference, and presence / absence of interference from a primary user who is a wireless communication system having a constant ratio A label indicating the presence / absence of interference based on a communication history of a secondary user and a history of interference notification, which is a wireless communication system that performs wireless communication using the frequency band of the primary user, A second step of generating teacher data in which communication parameters are associated with each other, and the updating means includes a label indicating that there is interference, an interfered terminal position, and a communication parameter. Bias reduction processing for reducing the deviation between the number of interference data associated with each other, the number of non-interference data with which the label indicating that there is no interference, the position of the interfered terminal, and the communication parameter are associated with each other A third step of generating updated teacher data, which is applied to the teacher data and updating the teacher data, and the machine learning means includes n (n is an integer of 2 or more) including communication parameters based on the updated teacher data. This is a boundary for classifying communication parameters into a first class without interference and a second class with interference in an n-dimensional space consisting of parameters, and exists at the same distance from the first and second classes. A fourth step of determining a decision boundary, which is a boundary, by machine learning, and the communication parameter determination means uses the determination boundary so that no interference occurs in the primary user A fifth step of determining at least one of a communication parameter set and a communication parameter set in which interference occurs in the primary user, and a transmitting means transmits the determined communication parameter set to a terminal group of secondary users This is a program for causing a computer to execute the sixth step.

  By causing the computer to execute the program, update teacher data that reduces the deviation between the number of interference data and the number of non-interference data is generated, machine learning is performed based on the generated update teacher data, and the decision boundary is determined Is done. Then, using the determined decision boundary, at least one of a communication parameter set in which interference does not occur in the primary user and a communication parameter set in which interference occurs in the primary user is determined. As a result, the secondary user can determine whether or not interference occurs based on the determined communication parameter.

  Therefore, it is possible to accurately determine communication parameters for the secondary user to avoid interference and perform wireless communication in real time.

(Configuration 14)
In the configuration 13, in the fifth step, the communication parameter determination unit fixes (n−m) dimensions in which m parameters (m is an integer satisfying 1 ≦ m <n) among n parameters are fixed in the fifth step. A hyperplane is set, and a point on the (nm) dimensional hyperplane that maximizes or minimizes the gain function is determined using the distance between the set point on the (nm) dimensional hyperplane and the decision boundary as an index. Then, a communication parameter set in which interference does not occur in the primary user is determined from the searched point on the (nm) -dimensional hyperplane using the constraint condition.

  In this way, the communication parameter set of the secondary user that does not interfere with the primary user is determined.

  Therefore, it is possible to accurately determine communication parameters for the secondary user to avoid interference and perform wireless communication in real time.

(Configuration 15)
In the configuration 13 or the configuration 14, in the third step, the update unit updates the teacher data by increasing the number of interference data and decreasing the number of non-interference data to generate updated teacher data.

  By increasing the number of interference data and decreasing the number of non-interference data, the deviation between the number of interference data and the number of non-interference data is reduced, so the number of non-interference data is larger than the number of interference data. In a wireless communication environment, a communication parameter set of a secondary user that does not interfere with the primary user can be accurately determined in real time.

(Configuration 16)
In the configuration 15, in the third step, the update unit generates pseudo interference data based on the interference data included in the teacher data, and generates update teacher data by deleting non-interference data included in the teacher data. To do.

  Since the pseudo interference data is generated from the interference data included in the teacher data, the pseudo data can be generated at a point where interference can occur due to the correlation of shadowing in wireless communication. As a result, the communication parameter set of the secondary user that does not interfere with the primary user can be determined with high accuracy.

(Configuration 17)
In the configuration 16, in the third step, the updating means uses the interference data included in the teacher data as a reference point, and the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point are predetermined. The pseudo interference data is generated so as to follow the distribution, and non-interference data existing outside the circle centered on the reference point is deleted to generate update teacher data.

  Generates pseudo data so that it follows the same distribution as the interference data that is the reference point, and deletes non-interference data that exists at a certain distance from the interference data, so that pseudo data is generated where interference can occur. In addition, non-interference data that is not considered is deleted.

  Therefore, it is possible to accurately determine the communication parameter set of the secondary user that does not interfere with the primary user. In addition, the amount of calculation can be reduced.

(Configuration 18)
In the configuration 17, in the third step, the updating unit performs pseudo-simulation so that the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a normal distribution when the variance is changed. Generate interference data.

  Pseudo interference data is generated by statistical processing.

  Therefore, pseudo data can be easily generated at a point where interference can occur.

(Configuration 19)
In any one of Configurations 13 to 18, the program has a case where the first risk level assigning unit uses the coordinate point of the update teacher data received from the updating unit as the transmission point and the position of the primary user as the reception point In addition, the received power in consideration of the terrain data is estimated, the first interference occurrence risk is estimated based on the estimated received power, and the estimated first interference occurrence risk is given to the updated teacher data. When the seventh step and the second risk level assigning unit receive the decision boundary determined based on the update teacher data to which the first interference occurrence risk level is given, a plurality of update teacher data existence regions are defined. It divides into regions, and estimates the received power when the coordinate point of the divided one region is the transmission point and the primary user's position is the reception point, and the second based on the estimated received power Interference hazard And causing the computer to further execute an eighth step of executing the process of generating the determination data by assigning the estimated second interference occurrence risk to one area for all of the plurality of areas, In the fourth step, the machine learning means determines a decision boundary by machine learning based on the updated teacher data to which the first interference occurrence risk is given. In the fifth step, the communication parameter determination means Based on the decision boundary determined based on the updated teacher data to which the first interference occurrence risk is assigned, the determination data, and the plurality of areas to which the second interference occurrence risk is assigned, It is determined whether each of the areas is a communication prohibited area, and at least one of the communication prohibited area and the communication permitted area is determined as a communication parameter set.

  Since the decision boundary is determined using the first and second interference occurrence risks determined in consideration of the terrain data, the interference data and the non-interference data are not compared with the case where the first and second interference occurrence risks are not used. Interference data can be easily separated.

(Configuration 20)
In the configuration 19, in the fifth step, the communication parameter determination means compares the determination data with the determination boundary, and communicates the region corresponding to the determination data classified as the region causing the interference among the plurality of regions. It is determined as a prohibited area.

  Since the relative position between each area of the plurality of areas and the determination boundary is determined via the determination data, it is possible to accurately determine whether each area is a communication prohibited area.

(Configuration 21)
In the configuration 19 or the configuration 20, the first risk level assigning unit estimates the estimated received power as the first interference occurrence risk level in the seventh step, or obtains the estimated received power as the received power. A division result obtained by dividing by the received power at the position of the interference data closest to the point is estimated as the first interference occurrence risk.

  The estimated received power or the result of dividing the estimated received power by the received power at the position of the interference data closest to the point where the received power is obtained is set as the first interference occurrence risk level.

  Accordingly, the first interference occurrence risk can be estimated based on the terrain data.

(Configuration 22)
In any one of Configurations 19 to 21, the first risk assigning unit holds a plurality of radio wave propagation characteristics as radio wave propagation characteristics between the primary user and the secondary user in the seventh step. The radio wave propagation characteristics that best match the terrain data are selected from a plurality of radio wave propagation characteristics, and an estimation process for estimating received power using the selected radio wave propagation characteristics is executed.

  Accordingly, the received power can be estimated in conformity with the terrain data.

(Configuration 23)
In the structure 22, the second risk level assigning unit estimates the received power in the eighth step by executing the same estimation process as the estimation process in the seventh step.

  Therefore, it is possible to estimate the received power that is the basis of the second risk of occurrence of interference in accordance with the terrain data.

(24)
In any one of Configuration 19 to Configuration 23, the terrain data includes an elevation map or an aerial photograph.

(Configuration 25)
According to the embodiment of the present invention, the recording medium is a computer-readable recording medium in which the program according to any one of Configurations 13 to 24 is recorded.

  Communication parameters for secondary users to perform wireless communication while avoiding interference can be accurately determined in real time.

It is the schematic which shows the communication apparatus in embodiment of this invention. It is the schematic which shows the structure in Embodiment 1 of the management apparatus shown in FIG. It is the schematic which shows the structure of the machine learning part shown in FIG. It is a conceptual diagram of a management table. It is a conceptual diagram of teacher data. It is a figure for demonstrating the production | generation method of pseudo data. It is a figure for demonstrating the method to delete excess data. It is a figure for demonstrating the production | generation method of update teacher data. It is a conceptual diagram of a decision boundary. It is a figure for demonstrating the method of calculating | requiring a decision boundary. It is a figure for demonstrating the method of calculating | requiring a decision boundary. It is a figure for demonstrating the determination method of a communication parameter set using a determination boundary. 5 is a flowchart in the first embodiment for explaining operations of a primary user's terminal and a management apparatus. It is a flowchart for demonstrating the detailed operation | movement of step S11 of FIG. It is a figure which shows the relationship between the ratio of the interference point outside a prohibition area | region, and the area of a prohibition area | region. It is the schematic which shows the structure in Embodiment 2 of the management apparatus shown in FIG. It is the schematic which shows the structure of the machine learning part and parameter determination part which are shown in FIG. It is a figure which shows the relationship between the electric field strength and distance in an Okumura-Sakai model. FIG. 10 is a conceptual diagram of a decision boundary in the second embodiment. It is a conceptual diagram of a decision boundary the case of not using the interference occurrence risk p _1. It is a figure for demonstrating the determination method of a communication prohibition area | region. It is a flowchart in Embodiment 2 for demonstrating operation | movement of a primary user's terminal and a management apparatus. It is a flowchart for demonstrating the detailed operation | movement of step S11A of FIG. It is a flowchart for demonstrating the detailed operation | movement of step S12A of FIG.

  Embodiments of the present invention will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and description thereof will not be repeated.

  FIG. 1 is a schematic diagram showing a communication device according to an embodiment of the present invention. Referring to FIG. 1, there are wireless communication systems WLC1 and WLC2. The wireless communication system WLC1 is a licensed wireless communication system and is referred to as a “primary user”. The wireless communication system WLC1 includes a wireless station 1 and terminals 2 and 3. The radio station 1 is a primary user's radio station, and the terminal 3 is a primary user's terminal. The wireless station 1 and the terminals 2 and 3 perform wireless communication with each other in the licensed frequency band.

  The wireless communication system WLC2 is a wireless communication system that performs wireless communication in the frequency band of the primary user, and is referred to as a “secondary user”. The wireless communication system WLC2 includes a wireless station 11 and terminals 12 and 13. The radio station 11 is a secondary user radio station, and the terminals 12 and 13 are secondary user terminals. The wireless station 11 and the terminals 12 and 13 perform wireless communication with each other in the primary user's frequency band.

  Each of the terminals 2 and 3 detects the received signal strength when the signals from the terminals 12 and 13 are received. Each of the terminals 2 and 3 determines that there is interference when the received signal strength is 83 dBm or more, and determines that there is no interference when the received signal strength is less than 83 dBm.

  When each of the terminals 2 and 3 determines that there is interference, for example, the time when it is determined that there is interference using a built-in timer while detecting its own position using GPS (Global Positioning System) Is detected. Since the detected position is a position when the terminals 2 and 3 receive interference, it is called “interfered terminal position”.

  Each of the terminals 2 and 3 generates an interference notification including the time, the presence / absence of interference, and the interfered terminal position, and transmits the generated interference notification to the radio station 1.

  The wireless station 1 receives the interference notification from the terminals 2 and 3 and transmits the received interference notification to the wireless station 11.

  The management device 10 is disposed in the wireless station 11. Then, the management device 10 receives an interference notification from the wireless station 1. In addition, the management device 10 performs wireless communication with the terminals 12 and 13, and holds the time when the wireless communication is performed and the history of the wireless communication (communication history of the secondary user) in association with each other.

  Then, the management apparatus 10 extracts the time, presence / absence of interference, and the location of the interfered terminal from the interference notification, and stores the extracted time, presence / absence of interference, and the location of the interfered terminal in a management table described later in association with each other. At the same time, the stored time and the communication history of the secondary user are associated with each other and stored in a management table described later.

  The management apparatus 10 repeats the above operation, and stores the presence / absence of interference, the location of the interfered terminal, and the communication history of the secondary user in the management table in association with each time.

  The management apparatus 10 creates teacher data based on the management table by a method described later, determines a communication parameter set that does not interfere with the primary user based on the created teacher data, and determines The communication parameter set is transmitted to the terminals 12 and 13.

  When receiving the communication parameter set from the management device 10, the terminals 12 and 13 perform wireless communication with the wireless station 11 using the received communication parameter set.

[Embodiment 1]
FIG. 2 is a schematic diagram showing the configuration of the management apparatus 10 shown in FIG. 1 in the first embodiment. With reference to FIG. 2, the management apparatus 10 includes an information management unit 101, a machine learning unit 102, a parameter determination unit 103, and a parameter notification unit 104.

  The information management unit 101 includes a timer (not shown) and receives an interference notification from the wireless station 1.

  Further, the information management unit 101 manages communication parameters (communication parameters for secondary users) used for wireless communication with the terminals 12 and 13 in association with the time. The communication parameters are, for example, transmission power, transmission terminal position, carrier frequency, bandwidth, modulation scheme, multiple access scheme, antenna sector, and the like.

  Then, the information management unit 101 stores in the management table the presence / absence of interference and the location of the interfered terminal included in the interference notification in association with the time, and also manages the communication parameters in the secondary user in association with the time. To store. That is, the information management unit 101 manages the history of interference notification (the presence / absence of interference and the location of the interfered terminal position) and the communication history of the secondary user (communication parameter history of the secondary user) in time series. . The information including the history of the interference notification and the communication history of the secondary user in time series is referred to as a “management table”.

  When the information management unit 101 newly receives an interference notification, the information management unit 101 adds the interference notification history and the communication history of the secondary user to the management table and updates the management table. Then, the information management unit 101 outputs the updated management table to the machine learning unit 102.

  The machine learning unit 102 receives the management table from the information management unit 101, and creates teacher data by a method described later based on the received management table. Then, the machine learning unit 12 updates the teacher data by a method described later, and generates updated teacher data. Thereafter, the machine learning unit 102 performs machine learning based on the updated teacher data, and is a boundary for classifying the communication parameter into the class CL1 having no interference and the class CL2 having interference, and the classes CL1 and CL2. Determine a decision boundary that is a boundary equidistant from both.

  Then, the machine learning unit 102 outputs the determined determination boundary to the parameter determination unit 103.

  The parameter determination unit 103 receives a determination boundary from the machine learning unit 102 and uses the received determination boundary to determine a communication parameter set that does not cause interference in the primary user by a method described later. Then, parameter determination unit 103 outputs the determined communication parameter set to parameter notification unit 104.

  The parameter notification unit 104 receives the communication parameter set from the parameter determination unit 103 and transmits the received communication parameter set to the terminals 12 and 13.

  FIG. 3 is a schematic diagram showing the configuration of the machine learning unit 102 shown in FIG. With reference to FIG. 3, the machine learning unit 102 includes a generation device 1021, a pseudo data generation device 1022, an excess data deletion device 1023, an update device 1024, and a determination device 1025.

The generation device 1021 receives the management table from the information management unit 101, and generates teacher data by a method described later based on the received management table. Then, the generation device 121 outputs the generated teacher data to the pseudo data generation device 1022 and the excess data deletion device 1023.

  The pseudo data generation device 1022 receives the teacher data from the generation device 1021, and generates pseudo data from the received teacher data by a method described later. Then, the pseudo data generation apparatus 1022 outputs the generated pseudo data to the update apparatus 1024.

  The excessive data deleting device 1023 receives the teacher data from the generating device 1021, and deletes the excessive data from the received teacher data. Then, the excess data deletion device 1023 outputs the teacher data from which the excess data has been deleted to the update device 1024.

  The update device 1024 receives pseudo data from the pseudo data generation device 1022 and receives teacher data from which excess data has been deleted from the excess data deletion device 1023. Then, the update device 1024 updates the teacher data by adding pseudo data to the teacher data from which excess data has been deleted, and generates updated teacher data. Then, the update device 1024 outputs the update teacher data to the determination device 1025.

  The determination device 1025 receives the update teacher data from the update device 1024, performs machine learning based on the received update teacher data, and classifies the communication parameters into the class CL1 having no interference and the class CL2 having interference. And a decision boundary which is a boundary equidistant from both of the classes CL1 and CL2. Then, the determination device 1025 outputs the determined determination boundary to the parameter determination unit 103.

  FIG. 4 is a conceptual diagram of the management table. Referring to FIG. 4, management table Table-CTL includes a primary user interference notification history and a secondary user communication history. The primary user interference notification history and the secondary user communication history are associated with each other.

  The history of the interference notification of the primary user includes time 1, the presence / absence of interference, and the interfered terminal position. Time 1, the presence / absence of interference, and the interfered terminal position are associated with each other.

  The communication history of the secondary user includes time 2, transmission power, and center frequency. Time 2, transmission power, and center frequency are associated with each other.

  In the above description, the secondary user communication parameters have been described as including transmission power, transmission terminal location, carrier frequency, bandwidth, modulation scheme, multiple access scheme, antenna sector, and the like. However, in FIG. 4, only the transmission power and the center frequency are shown as secondary user communication parameters for easy explanation. Actually, the communication history of the secondary user includes the transmission power, the transmission terminal position, the carrier frequency, the bandwidth, the modulation scheme, the multiple access scheme, the antenna sector, and the like described above.

In FIG. 4, at times t0, t3, and t4, there is no interference, and the transmission power and the center frequency for the secondary user at that time are shown. Since there is no interference, “unknown” is stored in the interfered terminal position. And “no interference”, “unknown”, 2
The transmission power and the center frequency of the next user are associated with times t0, t3, and t4.

  Also, in FIG. 4, at times t1 and t2, there is interference, and the position of the interfered terminal at that time, the transmission power and the center frequency of the secondary user are shown. Then, “with interference”, interfered terminal position ((x1, y1), (x2, y2)), the transmission power of the secondary user and the center frequency of the secondary user are associated with times t1 and t2. Yes.

  As described above, the management table Table-CTL has a configuration in which the presence / absence of interference of the primary user, the position of the interfered terminal, and the communication parameter of the secondary user are associated with the time.

  When receiving the interference notification from the radio station 1, the information management unit 101 extracts time 1, presence (or absence) of interference, and interfered terminal position from the interference notification. Then, the information management unit 101 stores the time 1, the presence (or absence) of interference, and the interfered terminal position in association with each other in the management table Table-CTL. The transmission power of the user and the center frequency of the secondary user are associated with each other and stored in the management table Table-CTL.

  The information management unit 101 performs this operation every time an interference notification is received, and creates and stores a management table Table-CTL. When the information management unit 101 receives a new interference notification, the information management unit 101 performs the above operation to update the management table Table-CTL, stores the updated management table Table-CTL, and outputs it to the machine learning unit 102. .

  FIG. 5 is a conceptual diagram of teacher data. Referring to FIG. 5, teacher data TCHR_D includes the presence / absence of interference, the interfered terminal position, transmission power, and the center frequency. The presence / absence of interference, the interfered terminal position, the transmission power, and the center frequency are associated with each other.

  In FIG. 5, for ease of explanation, the transmission power and the center frequency are shown as the communication parameters of the secondary user, but actually the transmission power is used as the communication parameter of the secondary user. , Transmitting terminal position, carrier frequency, bandwidth, modulation scheme, multiple access scheme, antenna sector, etc. are stored in the teacher data TCHR_D.

  In the teacher data TCHR_D, “1” indicates that there is interference, and “−1” indicates that there is no interference.

  In the teacher data TCHR_D, when there is no interference, the interfered terminal position is represented by (0, 0).

  The presence / absence of interference (−1 or 1), the interfered terminal position ((0, 0) or (x1, y1), etc.), transmission power (7 dBm, 10 dBm, etc.) and center frequency (2.42 GHz, etc.) are: They are associated with each other.

  When the generation device 1021 of the machine learning unit 102 receives the management table Table-CTL from the information management unit 101, the generation device 1021 refers to the time 1 and time 2 of the received management table Table-CTL, and the time 1 and time 2 are The presence / absence of interference, the interfered terminal position, the transmission power, and the center frequency are extracted, and the teacher data TCHR_D is created by associating the extracted interference presence / absence, interfered terminal position, transmission power, and center frequency with each other. In this case, the teacher data TCHR_D is created by representing the presence of interference as “1”, the absence of interference as “−1”, and the interfered terminal position when there is no interference as (0, 0).

  As a result, the machine learning unit 102 creates teacher data TCHR_D shown in FIG.

  In the teacher data TCHR_D, the flag (“1”) indicating that there is interference and the interfered terminal position, transmission power, and center frequency associated with the flag (“1”) constitute “interference data”. . In the teacher data TCHR_D, the flag (“−1”) indicating that there is no interference and the interfered terminal position, transmission power, and center frequency associated with the flag (“−1”) are “non-interfering”. Data ".

  In FIG. 5, only the transmission power and the center frequency are shown as communication parameters of the secondary user, but actually, the communication history of the secondary user includes the transmission power, the transmission terminal position, the carrier frequency, the band described above. It consists of width, modulation system, multiple access system, antenna sector, and the like. Therefore, the flag (“1”) indicating that there is interference, the interfered terminal position associated with the flag (“1”), and the communication parameters of the secondary user (transmission power, transmission terminal position, carrier frequency, (Bandwidth, modulation scheme, multiple access scheme, antenna sector, etc.) constitute “interference data” and are associated with a flag (“−1”) indicating no interference and a flag (“−1”). The interfered terminal location and secondary user communication parameters (transmission power, transmission terminal location, carrier frequency, bandwidth, modulation scheme, multiple access scheme, antenna sector, etc.) constitute “non-interfering data”.

  FIG. 6 is a diagram for explaining a method of generating pseudo data. FIG. 6 shows a two-dimensional space composed of two parameters for easy understanding of the description. The “distance from the interference source” is a parameter that cannot be changed (cannot be changed).

  Referring to FIG. 6, teacher data is arranged in a two-dimensional space. White circles represent non-interference data, and black circles represent interference data. Thus, the number of interference data is very small compared to the number of non-interference data. In actual communication, since non-interference data indicates normal communication, there is abundant history, but interference data is generated only when there is an interference notification, so that there is very little. Therefore, there is a bias between the number of interference data and the number of non-interference data.

  In machine learning, learning data that has a large bias in the probability of occurrence between classes belonging to different labels (presence or absence of interference) is called asymmetric data. Learning with asymmetric data significantly reduces the learning accuracy. (Non-patent Document 5).

  Therefore, in the embodiment of the present invention, the interference data (black circle) is used as a reference point, and pseudo data that is interference data is generated from the reference point.

  For example, the pseudo data generation apparatus 1022 changes the variance of the normal distribution as a parameter, and generates the pseudo data so that the interference data and the pseudo data are distributed according to the normal distribution having the changed variance.

  Further, the pseudo data generation device 1022 generates pseudo data so that the interference data and the pseudo data fall within a uniform distribution range.

  As described above, the pseudo data generation device 1023 adds noise (noise that follows a normal distribution having a changed dispersion or noise that falls within a uniform distribution range) to interference data (reference point). Generate pseudo data.

  In general, outdoor wireless communication is affected by shadowing due to shielding by a building or the like, but it is known that shadowing has a spatial correlation (Non-Patent Document 6). Therefore, interference is likely to occur near the interference data.

  Therefore, by generating the pseudo data by appropriately setting the noise to be added to the interference data (reference point), there is a problem that the pseudo data can be generated at a point where interference cannot occur originally, which can occur in the conventional oversampling method. It is possible to increase the ratio of interference data in the teacher data after avoiding it.

  When coordinates in a three-dimensional space are used as feature quantities, the distribution of the normal distribution or the range of the uniform distribution is a distance (generally about 20 m) at which the spatial correlation of shadowing in wireless communication is 0.5. Must be set.

  As described above, the pseudo data generation apparatus 1022 generates pseudo data from the interference data (reference point) included in the teacher data.

  FIG. 7 is a diagram for explaining a method of deleting excess data. FIG. 7 also shows a two-dimensional space composed of two parameters for easy understanding. The “distance from the interference source” is a parameter that cannot be changed (cannot be changed).

  Referring to FIG. 7, teacher data is arranged in a two-dimensional space. White circles represent non-interference data, and black circles represent interference data. The excessive data deletion device 1023 of the machine learning unit 102 creates a reference circle CRC (n-1 dimensional hypersphere in the n-dimensional plane) centered on a reference point that is interference data, and a reference circle CRC (n in the n-dimensional plane). The non-interference data of the area outside the (-1D hypersphere) is deleted. By deleting excess data by this method, while avoiding the problem of non-interfering data that should be learned in order to obtain an originally communicable region that can occur with the conventional undersampling method, The ratio of non-interfering data can be reduced.

  The reference circle CRC (n-1 dimensional hypersphere in the n-dimensional plane) is, for example, the same distance from half of the distance to the nearest other reference point so that the distance from the reference point is sufficiently separated on the n-dimensional plane. Or, when using coordinates in a three-dimensional space as a feature amount, the distance at which the spatial correlation of shadowing is 0.5 or less (generally about 20 m) or more, and the distance at which the spatial correlation is 0 (generally about 100 m) or less Set as a guide.

  FIG. 8 is a diagram for explaining a method of generating updated teacher data. Referring to FIG. 8, the update device 1024 of the machine learning unit 102 receives pseudo data (see (a) in FIG. 8) from the pseudo data generation device 1022, and excess data (part of non-interference data) is deleted. The teacher data (see FIG. 8B) is received from the excessive data deletion device 1023.

  Then, the update device 1024 adds pseudo data (see FIG. 8A) to the teacher data (see FIG. 8B) from which excess data (part of the non-interference data) has been deleted, thereby updating the teacher. Data (see FIG. 8C) is generated. Then, the update device 1024 outputs update teacher data to the determination device 1025.

  As described above, in the embodiment of the present invention, pseudo data is generated from interference data (reference point) to increase the number of interference data, and non-interference data existing outside the reference circle CRC around the reference point. Is deleted to reduce the number of non-interfering data, thereby reducing the bias between the interference data and the non-interfering data.

  When determining the transmission power of the secondary user that does not cause interference in the primary user, learning is performed using the update teacher data TCHR_D_UP, and the communication parameters are classified into the class CL1 having no interference and the class CL2 having interference. There is a need.

  For example, a support vector machine (SVM: Support Vector Machine) is used as a machine learning device suitable for classifying communication parameters into two classes CL1 and CL2.

  SVM is one of the pattern recognition models using supervised learning, and learns the boundary that most boldly divides the learning sample (update teacher data). In order to learn this boundary, an n-dimensional space constituted by n (n is an integer of 2 or more) parameters is set.

  As a result of learning using the learning sample (update teacher data) arranged in the n-dimensional space, the boundary that most boldly divides the learning sample (update teacher data) is the sample (data) closest to the boundary. And the distance (margin) is determined to be the maximum.

  A feature of SVM is to search for a hyperplane having a maximum margin among many candidate planes that can separate a plurality of points in an n-dimensional space.

  This margin means the minimum value of the distance from the hyperplane to a plurality of points. When trying to classify a plurality of points into two classes CL1 and CL2 while maximizing this margin, some margins belonging to class CL1 are eventually obtained. The hyperplane must be positioned such that the minimum value of the distances to the points of and the minimum values of the distances to some points belonging to class CL2 are equal.

  Such a hyperplane is a hyperplane having a maximum margin, and is referred to as a “decision boundary” in the embodiment of the present invention.

  FIG. 9 is a conceptual diagram of a decision boundary. In FIG. 9, a two-dimensional space composed of two parameters is shown for easy understanding.

  When determining the transmission power of the secondary user in which no interference occurs in the primary user, the presence / absence of interference and the transmission power are selected from the update teacher data. Further, as a parameter that cannot be changed (cannot be changed), for example, a distance from the interference source is selected.

  Then, when the distance between the transmission power and the interference source is a parameter, the presence of interference is represented by a black circle and the absence of interference is represented by a white circle, as shown in FIG.

  Then, in the two-dimensional space, a determination boundary BD that divides a plurality of black circles and a plurality of white circles is determined by the following method. In FIG. 9, the hyperplane having the maximum margin is a straight line on a two-dimensional plane for easy understanding. If the ideal decision boundary cannot be expressed by a straight line (or a plane), data conversion is performed using a kernel function, for example, a Gaussian kernel (RGB kernel), and processing is performed so that the decision boundary can be expressed by a plane. The subsequent procedure is the same.

  10 and 11 are diagrams for explaining a method for obtaining a decision boundary.

  The identification line that distinguishes the group without interference (white circle group) and the group with interference (black circle group) with the maximum margin is unclear where it passes, but a positive example that creates the maximum margin somewhere ( It is assumed that there is data at the end of each group of a white circle group without interference) and a negative example (black circle group with interference). This edge data is called a support vector.

  Let two lines passing through the support vector be ax + by + c = 1 and ax + by + c = −1. Since the two lines are parallel, the coefficients are the same.

  The area indicated by ax + by + c ≦ 1 is a range in which positive example (white circle group without interference) exists, and the area indicated by ax + by + c ≧ −1 is negative example (black circle group with interference) data. Is within the range (see FIG. 10).

  Here, two inequalities that are necessary conditions for drawing the identification line are obtained. These are combined into one. That is, in the positive example, the inequality is multiplied by “1” to be ax + by + c ≦ 1, and in the negative example, the inequality is multiplied by “−1” to be −1 (ax + by + c) ≧ −1.

  Then, if a variable t (t = 1 or t = −1) indicating whether it is a positive example or a negative example is used, t (ax + by + c) ≦ 1.

Distance L ₁ perpendicular line drawn to the identification line from each of the support vector positive cases and negative cases are represented by the following equation.

  As a result, the position of the identification line based on the maximum margin can be determined by obtaining a, b, and c that minimize the denominator of d while satisfying the condition of t (ax + by + c) ≦ 1 for all the teacher data ( FIG. 11). Determining the position of the identification line based on this maximum margin is equivalent to determining the equidistant position from both the positive example (white circle group without interference) and the negative example (black circle group with interference). To do.

When there are i (i = 1, 2, 3, 4,...) Update teacher data, t ₁ (ax ₁ + by ₁ + c) ≦ 1, t ₂ (ax ₂ + by ₂ + c) ≦ 1,. .., T _i (ax _i + by _i + c) ≦ 1 is used to obtain a, b, and c that give the minimum value of a ² + b ² .

The problem of obtaining a parameter for determining an identification line based on the maximum margin is that t _i (ax _i + by _i + c) ≦ 1 (i = 1 to 1), assuming two variables x and y and N update teacher data. N), it becomes a constrained minimization problem for obtaining a, b, c that minimizes a ² + b ² .

A convenient method for solving such a constrained minimization problem is Lagrange's undetermined multiplier (coefficient) method. The a ² + b ² to be minimized is set to f (a, b) = a ² + b ² , and the constraint condition t _i (ax _i + by _i + c) ≦ 1 is set to g (x, y) = t _i (ax _i + by _i + c ) If −1 ≦ 0, the Lagrange's undetermined multiplier method can be applied to convert the following equation (2) into equation (3).

  In equations (2) and (3), s. t. Means “obey the conditions described on the right side of st.” (The same applies hereinafter).

Equation (2) _{is, g 1 (a, b,} c) ≦ 0, g 2 (a, b, c) ≦ 0, ···, g n (a, b, c) ≦ 0 according f (a, This is a main problem because it means obtaining a and b that minimize b).

The expressions (3a) and (3b) in the expression (3) are the results of applying Lagrange's undetermined multiplier method. Equation (3b) represents a, b, c that minimizes L (a, b, c, μ ₁ ,..., Μ _n ). φ is a symbol indicating one function, similarly to f.

Equation (3c), (3d) is for showing the dual _{problem, φ (μ 1, ···,} μ n) μ 1 to maximize, ..., by obtaining the mu _n, the main problem It is guaranteed that a, b, and c that give the minimum value are also obtained.

In the following, in order to simplify the explanation, the main problem is created with two teacher data, and the function L is created by applying Lagrange's undetermined multiplier method. However, a ² + b ² is multiplied by 1/2 for convenience of calculation.

t _i (ax _i + by _i + c) ≦ 1 is transformed into t _i (ax _i + by _i + c) −1 ≦ 0. As a result, the following equation is obtained.

  In order to create a dual problem function to be maximized, L is minimized for a, b, and c. Therefore, L is partially differentiated with respect to a, b, and c, and the result is set to 0. Then, the following equation is obtained.

  Then, the following equation is obtained from the upper two equations of equation (5).

In Expressions (5) and (6), t _i is the sign of the i-th updated teacher data, and x _i and y _i are the x-coordinate and y-coordinate of the i-th teacher data, so t _i , x _i and y _i are known.

Therefore, if μ _i that maximizes the function of the dual problem is known, a and b can be obtained from Equation (6).

  Therefore, when all the conditions (equation (5)) obtained by partial differentiation are substituted for L in order to create a dual problem, the following equation is obtained.

  As a result, the dual problem is as follows.

Maximization of the function φ (μ ₁ , μ ₂ ) of the dual problem can be obtained by an interior point method (Non-Patent Document 7) that is a solution method of the convex quadratic programming problem. When emphasizing the speed for finding a solution, an evolutionary calculation such as a tabu search method (Non-Patent Document 8) may be used while considering the constraint conditions.

  The interior point method is a sequential iterative solution method that generates a point sequence that converges to an optimal solution within a constraint region of an optimization problem.

  The tabu search method is a kind of local search method that searches for a certain solution in the vicinity of the currently obtained solution, prevents the convergence to the local solution by not searching for a while as a tabu for the recently searched solution, The optimal solution is obtained by performing a global search of the solution space.

Therefore, points for increasing the function φ (μ ₁ , μ ₂ ) of the equation (8) are sequentially generated within the region of the constraint condition of the equation (8) (the equation described on the right side of the _st .), And the function φ (μ _1, μ ₂₎ optimal solution (μ _1, μ ₂₎ to maximize the seek.

In addition, when importance is attached to the speed for finding a solution, the optimum for performing the wide area search and maximizing the function φ (μ ₁ , μ ₂ ) while preventing the convergence to the local solution in consideration of the constraint condition of Equation (8). Find the solutions (μ ₁ , μ ₂ ).

If μ ₁ and μ ₂ can be obtained, a and b can be obtained from Equation (6) using known t ₁ , t ₂ , x ₁ , x ₂ , y ₁ , and y ₂ .

And, c is the x coordinate of a support vector and _{x S,} when the y-coordinate and _{y S,} obtained from _{ax S + by S + c =} 1 ( or _{ax S + by S + c =} -1).

  As a result, a decision boundary BD for identifying the classes CL1 and CL2 with the maximum margin is determined.

In the above description, the two-dimensional case where the teacher data is represented by x and y has been described. However, when the teacher data is composed of n dimensions, the x axis is x ₁ , the y axis is x ₂ , the z axis is x ₃ , ..., the axis of the nth dimension is expressed by _xn .

Further, a, also the coefficient of b such, _{w 1,} w 2 corresponds to n dimensions, ..., represented by _{w n.}

Since (x ₁ , x ₂ ,..., X _n ) and (w ₁ , w ₂ ,..., W _n ) are vectors, X = (x ₁ , x ₂ ,..., X _n ), W = (w ₁ , w ₂ ,..., w _n ).

Further, a ² + b ² becomes w ₁ ² + w ₂ ² +... + W _n ² . The square root of w ₁ ² + w ₂ ² +... + W _n ² is referred to as a (Euclidean) norm.

A place where the same dimensions such as x _i and x _j are multiplied and added is represented by X ^T X as an inner product of vectors. X ^T means the transposed matrix of X.

  Furthermore, when the number of teacher data is n, “2” is replaced with “n”.

  Therefore, when the teacher data consists of n dimensions, the identification line (= determination boundary BD) having the maximum margin is determined using the following equations (9) to (16).

  The determination device 1025 of the machine learning unit 102 determines the determination boundary BD by the method described above, and outputs the determined determination boundary to the parameter determination unit 103.

  FIG. 12 is a diagram for explaining a method for determining a communication parameter set using the determination boundary BD.

  For n parameters, an (n−m) -dimensional hyperplane is set with m parameters (m is an integer satisfying 1 ≦ m <n) that cannot be changed, such as an interference occurrence position, fixed. At this time, a certain point on the (n−m) -dimensional hyperplane represents a parameter set. This is called a parameter setting surface.

  An interference allowable threshold and an (n−1) -dimensional hyperplane (minimum communication condition plane) are set as thresholds for giving a parameter setting range. The specific determination method is as follows.

[Minimum communication conditions]
The minimum communication condition plane is obtained by smoothly connecting a minimum parameter set satisfying a desired signal-to-noise ratio at the receiving station of the secondary user. If the parameter set is closer to the decision boundary than the minimum communication condition surface, it is guaranteed that the secondary user can communicate without interfering with the primary user.

[Interference tolerance threshold]
The interference tolerance threshold indicates allowable interference. The distance from the decision boundary BD is k times the minimum value (k is a real number satisfying 0 <k <1) with respect to the minimum value of the distance between the observation point (the point included in the update teacher data) and the decision boundary BD. This value is set as an allowable interference threshold. For example, k is 0.5.

  Then, a point for maximizing or minimizing a certain gain function is searched on the parameter setting surface. At this time, the following constraint conditions are set as the constraint conditions.

  (1) The point at which the gain function is maximized or minimized is present on the interference-free side with respect to the decision boundary BD and is located on the decision boundary BD side with respect to the lowest communication condition surface.

(2) In addition, the distance L ₁ between the point and the decision boundary BD to maximize or minimize the gain function is in the range satisfying above interference tolerance threshold.

  As the gain function, for example, the following is assumed.

(I) Channel capacity of the secondary user's terminal The channel amount C is determined by C = Blog (1 + SINR). B is a bandwidth and depends on the frequency bandwidth to be used. SINR is a signal-to-noise ratio, and is determined by transmission power, frequency band, and the like.

(Ii) The communicable distance of the secondary user's terminal is determined depending on the frequency band to be used, transmission power, and the like.

(Iii) Bit error rate It is determined depending on transmission power, modulation scheme, frequency bandwidth, and the like.

  With reference to FIG. 12, the determination method of a communication parameter set is demonstrated concretely. In FIG. 12, the distance from the interference source is used as a parameter that cannot be changed. The distance from the interference source is obtained as the distance between the interfered terminal position included in the updated teacher data and the position of the secondary user's terminal that has given interference.

  Then, the obtained distance from the interference source is fixed and the parameter setting surface is set in a two-dimensional space. After that, the above-described minimum communication condition plane and interference tolerance threshold are set in a two-dimensional space.

  In FIG. 12, since the update teacher data is two-dimensional, the parameter setting surface, the minimum communication condition surface, and the interference tolerance threshold are straight lines.

  Then, a point where the gain function is maximized or minimized is searched on the parameter setting surface (straight line), and a point satisfying the above-described constraints (1) and (2) is selected as a candidate point from the searched points. Extract.

  When there are a plurality of candidate points, an arbitrary point of the plurality of candidate points is determined as a communication parameter set.

  On the other hand, when there is one candidate point, that one candidate point is determined as a communication parameter set.

  The parameter determination unit 103 determines a communication parameter set for performing wireless communication without interfering with the primary user's terminal by using the determination boundary BD by using the determination boundary BD.

  Then, parameter determination unit 103 outputs the determined communication parameter set to parameter notification unit 104.

  The changeable parameters include, for example, the transmission power of the base station or terminal in the secondary user, the transmission terminal position, the carrier frequency, the bandwidth, the modulation scheme, the multiple access scheme, the antenna sector, and the like.

  The parameters that cannot be changed include, for example, the position of the primary user's terminal, the distance from the primary user's terminal, the modulation method in the primary user, and the multiple access method of the primary user. .

  When the teacher data consists of three or more dimensions, the decision boundary BD, the minimum communication condition plane, and the interference tolerance threshold are planes.

  Therefore, the minimum communication condition described in the claims consists of an (n-1) -dimensional line or surface.

  FIG. 13 is a flowchart in the first embodiment for explaining the operations of the primary user's terminals 2 and 3 and the management apparatus 10.

  Referring to FIG. 13, the primary user terminal (at least one of terminals 2 and 3) detects interference (step S <b> 1), and transmits an interference notification to radio station 1. Then, the radio station 1 transmits the interference notification received from the terminal (at least one of the terminals 2 and 3) to the management apparatus 10 (step S2).

  On the other hand, when there is no interference, the information management unit 101 of the management apparatus 10 causes wireless communication between the wireless station 11 and the terminals 12 and 13 (step S3), and communication history with the secondary user's terminals 12 and 13 Is stored (step S4).

  Then, the information management unit 101 receives an interference notification from the radio station 1 (step S5), extracts the interference information and the communication history, and creates a management table Table-CTL (step S6). Then, the information management unit 101 outputs the management table Table-CTL to the machine learning unit 102.

  The generation device 1021 of the machine learning unit 102 receives the management table Table-CTL from the information management unit 101, collates the interference information included in the received management table Table-CTL with the communication history and time, and performs the teaching by the method described above. Data is created (step S7).

  Thereafter, the generation device 1021 of the machine learning unit 102 outputs the created teacher data to the pseudo data generation device 1022 and the excess data deletion device 1023.

  The excess data deletion device 1023 receives the teacher data from the generation device 1021, and deletes the excess data (some non-interference data) by the above-described method based on the interference data (reference point) included in the received teacher data. (Step S8). Then, the excess data deletion device 1023 outputs the teacher data from which the excess data has been deleted to the update device 1024.

  Further, the pseudo data generation device 1022 receives the teacher data from the generation device 1021, and generates the pseudo data by the above-described method from the interference data (reference point) included in the received teacher data (step S9). Then, the pseudo data generation apparatus 1022 outputs the generated pseudo data to the update apparatus 1024.

  The update device 1024 receives pseudo data from the pseudo data generation device 1022 and receives teacher data from which excess data has been deleted from the excess data deletion device 1023. Then, the update device 1024 updates the teacher data by adding pseudo data to the teacher data from which excess data has been deleted, and generates updated teacher data (step S10). Then, the update device 1024 outputs update teacher data to the determination device 1025.

  Thereafter, the determination device 1025 of the machine learning unit 102 inputs the update teacher data to the machine learner, and determines or updates the determination boundary BD using the SVM by the method described above (step S11).

  Then, the machine learning unit 102 outputs the determined determination boundary BD to the parameter determination unit 103.

  Upon receipt of the decision boundary BD, the parameter determination unit 103 determines a communication parameter set for preventing interference with the primary user's terminal by using the received decision boundary BD (step S12). ), And outputs the determined communication parameter set to the parameter notification unit 104.

  Then, the parameter notification unit 104 transmits the communication parameter set to the terminals 12 and 13 of the secondary user (Step S13).

  The terminals 12 and 13 receive the communication parameter set, and update the parameters used for wireless communication using the received communication parameter set (step S14).

  As a result, a series of operations is completed.

  In step S11, the decision boundary BD is determined or updated because the flowchart shown in FIG. 13 is executed each time the primary user's terminals 2 and 3 detect interference. This is because, when executed for the second time or later, the update teacher data is also updated, and the already determined decision boundary BD is updated using the updated update teacher data.

  Further, in the flowchart shown in FIG. 13, it has been described that pseudo data is generated, excess data is deleted, pseudo data is added to teacher data from which excess data has been deleted, and teacher data TCHR_D is updated to update teacher data TCHR_D_UP. (See Steps S8 to S10) In the embodiment of the present invention, the present invention is not limited to this. Excess data (partially non-interfering data) is deleted from the teacher data, and excess data (partially non-interfering data) The update teacher data TCHR_D_UP may be generated by generating pseudo data from the interference data (reference point) of the teacher data from which is deleted.

  FIG. 14 is a flowchart for explaining the detailed operation of step S11 of FIG.

  Referring to FIG. 14, the updating device 1024 of the machine learning unit 102 determines the objective function and the constraint condition shown in Expression (10) after step S10 in FIG. 13 (step S111).

  Then, the update device 1024 converts the main problem into a dual problem by applying the Lagrange's undetermined multiplier method to the objective function and the constraints as shown in Expression (11) (step S112).

Thereafter, the update device 1024 solves the dual problem using equations (12) to (16), and obtains μ _i that maximizes the function of the dual problem (step S113).

Subsequently, the update unit 1024, _{w 1,} w 2 that gives the minimum value of the main problems with the updating teacher data mu _i, · · ·, obtains the _{w n} (step S114). More specifically, the update device 1024, known _t i indicated by the updated training _data, and x i1 _{~x in,} and mu _i was obtained by substituting the equation _{(14) w 1, w 2} , · ..., seek _{w n.}

After step S114, the update device 1024 sets w ₁ as x ₁ coordinates, x ₂ coordinates,..., X _n coordinates of a certain support vector as x _1S , x _{2S ,.} Request _{x 1S + w 2 x 2S +} ··· + w n x nS + c = 1 ( or _{w 1 x 1S + w 2 x} 2S + ··· + w n x nS + c = -1) from the intercept c. Then, the update unit 1024, the determined _{w 1,} w 2, · · ·, positive by substituting the _{w n} and c to _{w 1 x 1 + w 2 x} 2 + ··· + w n x n + c = 0 A plane (or line) having the maximum margin (distance L ₁ ) with both the example and the negative example is obtained, and the obtained plane (or line) is determined as a decision boundary (step S115).

  Thereafter, the series of operations proceeds to step S12 in FIG.

  As described above, the management device 10 updates the update teacher data (the history information of communication parameters in the data transmission of the secondary user and the interference of the primary user in the data transmission) with reduced bias between the interference data and the non-interference data. Communication parameter set that does not interfere with the primary user's terminal (see S6 to S12), and the secondary user's communication parameter history information and the interference information A communication parameter set that does not interfere with the primary user's terminal can be accurately determined in real time from only the presence / absence history information.

  In addition, the management apparatus 10 receives only an interference notification including the presence / absence of interference, the position of the interfered terminal, and the time when the interference is detected from the primary user, and interferes with the primary user's terminal. Since no communication parameter set is determined, an appropriate communication parameter set can be set even if the wireless communication setting of the primary user is unknown.

  Furthermore, since the flowcharts shown in FIGS. 13 and 14 are executed every time interference is detected in the primary user's terminal, the primary user's terminal is interfered with in response to a real-time change in the interference state. Communication parameter sets that are not given can be set.

  Furthermore, if the decision boundary using the update teacher data is sufficiently learned, the communication parameter set can be determined with a small amount of calculation.

  Furthermore, although the risk when interference is given varies depending on the primary user, the learned decision boundary is uniquely determined, so even if the primary user changes, appropriate communication parameters can be set. . That is, the communication parameter set determination method described above can have robustness against changes of the primary user.

  Japanese Patent Laying-Open No. 2013-221609 (Patent Document 1) discloses a technique for selecting an appropriate communication parameter according to the surrounding state of a wireless communication device using a support vector machine (SVM).

  This technology associates the situation, parameters, and performance (communication performance) with each other and stores them in the history database, and refers to the history database to obtain communication parameters that provide good communication performance in the current surrounding situation. To choose.

  Then, the peripheral situation (context) is determined using a support vector machine (SVM).

  As described above, in Patent Document 1, the support vector machine (SVM) is used to determine the surrounding situation (context). That is, in Patent Document 1, a support vector machine (SVM) is used to classify surrounding situations (contexts).

  On the other hand, in the embodiment of the present invention, a support vector machine (SVM) is used to classify communication parameters that cause interference and communication parameters that do not cause interference.

  Therefore, the object to be classified using the support vector machine (SVM) is completely different between the embodiment of the present invention and Patent Document 1, and the communication parameter is classified using the support vector machine (SVM) in Patent Document 1. There is no suggestion to do.

  Further, in Patent Document 1, a communication parameter between wireless communication apparatuses that perform wireless communication with each other is selected. However, in the embodiment of the present invention, a secondary user is not a wireless communication partner. Communication parameters that do not interfere with the communication are determined.

  Accordingly, the determination of communication parameters in the first embodiment is completely different from the selection of communication parameters in Patent Document 1.

  FIG. 15 is a diagram illustrating the relationship between the ratio of interference points outside the prohibited region and the area of the prohibited region. In FIG. 15, the vertical axis represents the area of the prohibited region, and the horizontal axis represents the ratio of interference points outside the prohibited region. White circles indicate the relationship between the ratio of interference points outside the prohibited area and the area of the prohibited area in Embodiment 1, and black circles indicate the ratio of interference points outside the prohibited area and the area of the prohibited area in the fixed area allocation method. Shows the relationship.

  Referring to FIG. 15, interference occurs from the default communication prohibited area (for example, a circle centered on a radar in which the ratio of interference data outside the prohibited area is 10%), and machine learning proceeds. The area of the prohibited area increases, and the ratio of interference data outside the prohibited area decreases. When learning progresses and the ratio of interference data outside the prohibited area reaches a target value (for example, 1%, etc.), the method according to the first embodiment differs from the conventional method (fixed region allocation method), and inherently interference. Because it is difficult for pseudo data to be generated at a point where data cannot occur and non-interference data that should be learned to obtain a communicable area is deleted, the area of the communication prohibited area is more than necessary. It is not set too much.

  As another excellent effect, since the input data for machine learning can be reduced in the excess data deletion apparatus 1023, the amount of calculation for machine learning can be reduced.

  In the first embodiment, the functions of the information management unit 101, the machine learning unit 102, the parameter determination unit 103, and the parameter notification unit 104 may be realized by the program PROG_A.

  In this case, the program PROG_A includes steps S3 to S13 (including steps S111 to S115) described above.

  The management device 10 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The ROM stores a program PROG_A.

  When determining the communication parameter set, the CPU runs the program PROG_A out of the ROM. Then, the CPU stores history information of communication parameters of the secondary user in the RAM and stores various calculation results when determining the communication parameter set in the RAM.

  Therefore, the program PROG_A is a program for causing a computer (CPU) to determine a communication parameter set in order to perform wireless communication.

  In the first embodiment, the program PROG_A may be provided by being recorded on a recording medium such as a CD or a DVD. The user sets a recording medium on which the program PROG is recorded in the computer, and the computer (CPU) reads the program PROG_A from the recording medium and executes it.

  Therefore, the recording medium on which the program PROG_A is recorded is a recording medium that can be read by a computer (CPU).

  In the above, it has been described that the machine learning unit 102 determines the decision boundary BD using a support vector machine (SVM). However, in the first embodiment, the machine learning unit 102 is not limited to this, and the machine learning unit 102 may perform logistic regression ( Non-Patent Document 9) or AROW (Adaptive Regularization of Weight Vectors) (Non-Patent Document 10) may be used to determine the decision boundary BD. Generally, the decision boundary BD is determined using the update teacher data TCHR_D_UP as an input. The determination boundary BD may be determined using any machine learning device as long as it is a machine learning device.

  In the above description, the transmission power that does not interfere with the primary user's terminals 2 and 3 is determined. However, in the first embodiment, the present invention is not limited to this, and the primary user's terminals 2 and 3 are determined. In general, any communication parameter set may be determined as long as it is a communication parameter set that does not cause interference to the terminals 2 and 3 of the primary user. Good.

  Further, in the above description, the parameter that cannot be changed is “transmission power”. However, in Embodiment 1, the parameters other than “transmission power” are not limited to this, and may be parameters that cannot be changed. Good.

  Further, in the above description, the management apparatus 10 is described as being installed in the radio station 11, but in the first embodiment, the management apparatus 10 is not limited to this and is disposed at a position different from the radio station 11. May be. In this case, the management apparatus 10 may receive an interference notification from the wireless station 11, may receive an interference notification from the wireless station 1, or may directly receive an interference notification from the terminals 2 and 3.

  Furthermore, in the above description, it has been described that the management apparatus 10 receives the interference notification from the radio station 1, but in the first embodiment, the management apparatus 10 directly receives the interference notification from the terminals 2 and 3. You may receive it.

  Furthermore, in the above, when determining a communication parameter set that does not interfere with the primary user's terminals 2 and 3, a minimum communication condition plane and an allowable interference threshold are set, and the minimum communication condition plane is set on the parameter setting plane. In the first embodiment, the point that is closer to the decision boundary BD and the distance from the decision boundary BD is larger than the distance between the interference allowable threshold and the decision boundary BD is selected as the candidate point of the communication parameter set. However, the present invention is not limited to this, and a point belonging to the class CL1 having no interference with respect to the decision boundary BD is selected as a candidate point of the communication parameter set on the parameter setting surface without setting the minimum communication condition surface and the allowable interference threshold value. May be.

  Even if the secondary user's terminals 12 and 13 perform wireless communication using the communication parameter set determined in this manner, the primary user's terminals 2 and 3 do not interfere with each other. It is because it can achieve.

  Furthermore, in the above, generating pseudo data from the interference data (reference point) included in the teacher data, deleting non-interference data outside the reference circle CRC centered on the interference data (reference point), However, in the first embodiment, the present invention is not limited to this, and pseudo data is generated from interference data (reference points) included in the teacher data. The teacher data may be updated by applying only the generation, and the updated teacher data TCHR_D_UP may be generated, and only the non-interference data outside the reference circle CRC centered on the interference data (reference point) is deleted. The teacher data may be updated by application to generate updated teacher data TCHR_D_UP. Generally, interference data (reference point) included in the teacher data The teacher data is updated by applying at least one of generating pseudo data and deleting non-interference data outside the reference circle CRC centered on the interference data (reference point), and updating the teacher data TCHR_D_UP Should be generated.

  Apply at least one of generating pseudo data from interference data (reference point) included in the teacher data and deleting non-interference data outside the reference circle CRC centered on the interference data (reference point). However, the deviation between the number of interference data and the number of non-interference data can be reduced.

  Further, in the above description, in the wireless communication environment where the number of non-interference data is larger than the number of interference data, the update teacher data TCHR_D_UP is generated by increasing the number of interference data and decreasing the number of non-interference data. In the first embodiment, the present invention is not limited to this, and in a wireless communication environment in which the number of interference data is larger than the number of non-interference data, the number of non-interference data is increased by generating pseudo data composed of non-interference data. And updating the teacher data by applying at least one of reducing the number of interference data by deleting the interference data outside the reference circle centered on the non-interference data (reference point). Data TCHR_D_UP may be generated.

  That is, in the first embodiment, the update teacher data TCHR_D_UP may be generated by applying a bias reduction process for reducing the bias between the number of interference data and the number of non-interference data to the teacher data.

  Further, in the above description, the primary user has been described as a licensed radio communication system. However, in Embodiment 1, the primary user is not limited to this, and the proportion of data indicating interference is low. Any wireless communication system having a constant value may be used.

[Embodiment 2]
In the first embodiment, in order to determine the communication parameters of the secondary user, only the secondary user's communication information is handled as the feature quantity of machine learning, and the secondary user interferes with the primary user. Reflection and shielding by the terrain and buildings in the radio wave transmission path are not considered. Accordingly, an inappropriate communication parameter may be set depending on the characteristics of the propagation path such as the presence or absence of a line-of-sight communication path between the primary user and the secondary user.

  Specifically, in the case where the prospect between the secondary user and the primary user is poor due to mountains, buildings, etc., in the first embodiment, although communication is originally possible, the communication prohibited area May be set as. In the first embodiment, propagation environment characteristics are used in the pseudo data generation device and the excess data deletion device, but only statistical values such as the average, variance, and correlation distance of spatial correlation shadowing are handled. Thus, it is difficult to completely adapt to actual propagation path characteristics.

  Therefore, in the second embodiment, a communication prohibited area where secondary user communication is prohibited is determined in accordance with actual propagation path characteristics.

  FIG. 16 is a schematic diagram showing the configuration of the management apparatus 10 shown in FIG. 1 in the second embodiment. Referring to FIG. 16, management apparatus 10A according to the second embodiment is obtained by replacing machine learning unit 102 of management apparatus 10 shown in FIG. 2 with machine learning unit 102A and replacing parameter determination unit 103 with parameter determination unit 103A. Others are the same as those of the management apparatus 10.

  The machine learning unit 102A generates update teacher data in the same manner as the machine learning unit 102 described above, and outputs the generated update teacher data to the parameter determination unit 103A. The machine learning unit 102A receives the update teacher data to which the interference occurrence risk is assigned from the parameter determination unit 103A, performs machine learning based on the update teacher data to which the interference occurrence risk is assigned, and interferes with the communication parameter. A decision boundary which is a boundary for classifying into a class CL1 having no interference and a class CL2 having interference and is equidistant from both the classes CL1 and CL2 is determined. Then, machine learning unit 102A outputs the determined determination boundary to parameter determining unit 103A. The machine learning method for determining the decision boundary is the same as the machine learning method in the first embodiment.

  When the parameter determination unit 103A receives the update teacher data from the machine learning unit 102A, the parameter determination unit 103A estimates the received power when the coordinate of the update teacher data is the transmission point and the position of the primary user is the reception point. The risk of occurrence of interference is estimated based on the received power. Then, the parameter determination unit 103A gives the estimated interference occurrence risk to the updated teacher data and outputs it to the machine learning unit 102A.

  Thereafter, upon receiving the decision boundary from the machine learning unit 102A, the parameter determination unit 103A divides the region (two-dimensional region) covering the update teacher data into a plurality of meshes, and the center coordinates of each of the divided meshes Is estimated, and the estimated interference occurrence risk is assigned to each central coordinate. Then, the parameter determination unit 103A determines whether each mesh is a communication prohibited area based on the center coordinates to which the risk of occurrence of interference is given and the determination boundary, and the determination result is sent to the parameter notification unit 104. Output to.

  FIG. 17 is a schematic diagram illustrating configurations of the machine learning unit 102A and the parameter determination unit 103A illustrated in FIG.

  Referring to FIG. 17, machine learning unit 102A is the same as machine learning unit 102 shown in FIG. 3 except that determination device 1025 of machine learning unit 102 shown in FIG. 3 is replaced with determination device 1027. is there.

  Parameter determination unit 103A includes a terrain data feature extraction device 1026 and a communication prohibited region determination device 1028.

  In the machine learning unit 102A, the update device 1024 generates update teacher data by the method described above, and outputs the generated update teacher data to the terrain data feature extraction device 1026 of the parameter determination unit 103A. The interference data and non-interference data constituting the update teacher data are represented by (x, l) including the coordinate x of the update teacher data and the label l. Therefore, the update device 1024 outputs the update teacher data (x, l) to the terrain data feature extraction device 1026 of the parameter determination unit 103A.

  The terrain data feature extraction device 1026 of the parameter determination unit 103A holds an elevation map or an aerial photograph in the communication range of the primary user and the secondary user as terrain data in advance. This altitude map is a map that takes into account topography or structures such as undulations of mountains and valleys. Further, the terrain data feature extraction device 1026 receives update teacher data (x, l) from the update device 1024 of the machine learning unit 102A.

Upon reception of the update teacher data (x, l), the terrain data feature extraction apparatus 1026 receives the update when the coordinate x of the update teacher data (x, l) is the transmission point and the position of the primary user is the reception point. The power is estimated, and the interference occurrence risk p ₁ is estimated based on the estimated received power. Then, the terrain data feature extraction apparatus 1026 assigns the estimated interference occurrence risk p ₁ to the update teacher data (x, l) and updates the teacher data (x, p ₁ ) to which the interference occurrence risk p ₁ is assigned. , L), and the generated update teacher data (x, p ₁ , l) is output to the communication prohibited area determining device 1028 and the determining device 1027 of the machine learning unit 102A.

Further, the terrain data feature extraction device 1026 has center coordinates (x _{G_r} , y _{G_r} ) (r) of a plurality of meshes constituting the region (two-dimensional region) covering the update teacher data (x, p ₁ , l). Is an integer satisfying 1 ≦ r ≦ R, and R is an integer of 2 or more representing the number of meshes) from the communication prohibited area determining device 1028, the center coordinates (x _{G_r} , y _{G_r} ) are used as transmission points. , the received power in the case where a reception point position of the primary user estimated, estimates the interference occurrence risk p ₂ on the basis of the received power thereof estimated. Then, the terrain data feature extraction apparatus 1026 generates the determination data (x _{G_r} , y _{G_r} , p ₂ ) by assigning the interference occurrence risk p ₂ to the center coordinates (x _{G_r} , y _{G_r} ), and generates the generated data The determination data (x _{G — r} , y _{G — r} , p ₂ ) is output to the communication prohibited area determination device 1028.

The determination device 1027 of the machine learning unit 102A receives the update teacher data (x, p ₁ , l) from the terrain data feature extraction device 1026 of the parameter determination unit 103A, and receives the received update teacher data (x, p ₁ , l). Is a boundary for classifying a communication parameter into a class CL1 having no interference and a class CL2 having interference, and is a boundary equidistant from both of the classes CL1 and CL2. To decide. Then, the determination device 1027 outputs the determined determination boundary to the communication prohibited area determination device 1028 of the parameter determination unit 103A.

The communication prohibited area determining device 1028 of the parameter determining unit 103A receives update teacher data (x, p ₁ , l) from the terrain data feature extracting device 1026. Then, upon receiving the decision boundary from the decision device 1027 of the machine learning unit 102A, the communication prohibited region decision device 1028 converts the region (two-dimensional region) covering the update teacher data (x, p ₁ , l) into a plurality of meshes. The center coordinates (x _{G_r} , y _{G_r} ) of each of the divided meshes are output to the terrain data feature extraction device 1026.

Then, _{upon receiving the} determination data (x _{G_r} , y _{G_r} , p ₂ ) from the topographic data feature extraction device 1026, the communication prohibited area determination device 1028 determines the determination data (x _{G_r} , y _{G_r} , p ₂ ). And the mesh corresponding to the determination data classified as the region causing the interference is determined as the communication prohibited region. Then, the communication prohibited area determination device 1028 outputs the set of meshes determined as the communication prohibited area to the parameter notification unit 104.

  In the management apparatus 10A, the parameter notification unit 104 transmits a set of meshes determined to be a communication prohibited area to a terminal group of secondary users.

A method of estimating the interference occurrence risk p _{1 in} the terrain data feature extraction apparatus 1026 will be described. The terrain data feature extraction device 1026 refers to the coordinate x of the update teacher data (x, l), and determines the received power when the coordinate x is the transmission point and the primary user's position is the reception point as follows. Estimate by any of the methods.
(I) Received power is estimated using a radio wave propagation simulation.
(II) Received power is estimated using the Okumura-Kashiwa model.

  When estimating the received power using the radio wave propagation simulation of (I), the terrain data feature extraction apparatus 1026 takes altitude map information with a ray tracing simulator such as Ralab and sets the communication parameters similar to those of the secondary user. Thus, the received power at the coordinates of the primary user is estimated.

  Further, when estimating received power using the Okumura-Hagi model of (II), the terrain data feature extraction apparatus 1026 holds a plurality of radio wave propagation models in the Okumura-Hagi model, and can be used for altitude maps or aerial photographs. A radio wave propagation model closest to the topographic data extracted based on the selected radio wave propagation model is selected from a plurality of radio wave propagation models, and the received power at the coordinates of the primary user is estimated using the selected radio wave propagation model.

  FIG. 18 is a diagram showing the relationship between electric field strength and distance in the Okumura-Kashiwa model. In FIG. 18, the vertical axis represents the electric field strength, and the horizontal axis represents the distance. Curve k1 shows the relationship between electric field strength and distance in open land, curve k2 shows the relationship between electric field strength and distance in the suburbs, and curve k3 shows the relationship between electric field strength and distance in small and medium cities. A curve k4 shows the relationship between electric field strength and distance in a large city.

  Curves k1 to k4 shown in FIG. 18 are expressed by the following equations.

In Expression (17), A and B are common to the curves k1 to k4, and a (h _m ) and C are different for each area. L _CH is the electric field strength, d is the distance, and _hm is the antenna height of the secondary user.

  A and B in Expression (17) are represented by the following expressions.

In equation (18), f is the frequency of the propagating radio wave, and h _b is the primary user's antenna height.

When the radio wave propagation environment is an open ground (curve k1), a (h _m ) and C are expressed by the following equations.

When the radio wave propagation environment is a suburb (curve k2), a (h _m ) and C are expressed by the following equations.

If the radio wave propagation environment is small city (curve _k3), a (h m) and C is expressed by the following equation.

When the radio wave propagation environment is a large city (curve k4), a (h _m ) and C are expressed by the following equations.

As shown in Expression (22), in a large city, a (h _m ) varies depending on the frequency.

  The terrain data feature extraction apparatus 1026 holds Expressions (17) to (22). When the terrain data is closest to the open land, the terrain data feature extraction apparatus 1026 is represented by a curve k1 by Expressions (17), (18), and (19). Radio wave propagation characteristics are calculated, and received power at the position of the primary user is estimated based on the calculated radio wave propagation characteristics.

  Further, when the terrain data is closest to the suburbs, the terrain data feature extraction apparatus 1026 calculates the radio wave propagation characteristic indicated by the curve k2 by the equations (17), (18), and (20), and calculates the calculated radio wave propagation characteristic. Based on this, the received power at the position of the primary user is estimated.

  Furthermore, when the terrain data feature extraction device 1026 is closest to the small and medium city, the terrain data feature extraction device 1026 calculates the radio wave propagation characteristic indicated by the curve k3 by the equations (17), (18), and (21), and the calculated radio wave propagation characteristic. Based on the above, the received power at the position of the primary user is estimated.

  Furthermore, when the topographic data is closest to the big city, the topographic data feature extraction device 1026 calculates the radio wave propagation characteristics indicated by the curve k4 by the equations (17), (18), and (22), and the calculated radio wave propagation characteristics. Based on the above, the received power at the position of the primary user is estimated.

When the received power at the location of the primary user is estimated, the terrain data feature extracting apparatus 1026 estimates the interference occurrence risk p ₁ based on the estimated received power.

More specifically, the terrain data characteristic extraction unit 1026, as the interference occurrence risk p ₁ the received power estimated. Also, topographical data characteristic extraction unit 1026, the received power estimation to the interference occurrence risk p ₁ a value obtained by dividing the received power in an interference data having the closest coordinates on the received power.

Even if the risk of occurrence of interference p ₁ is estimated by either of these two methods, the larger the value is, the higher the received power is, or the higher the possibility that interference will occur.

Then, the terrain data feature extraction apparatus 1026 estimates the interference occurrence risk p ₁ for each of the interference data and the non-interference data constituting the update teacher data (x, l), and the interference occurrence risk p ₁ is given. Update teacher data (x, p ₁ , l) is generated.

Incidentally, topographical data characteristic extraction unit 1026, an interference occurs risk p ₁ of the interference occurrence risk p ₂ was estimated by the same method as the estimation method, determination data interference occurrence risk p ₂ is assigned (x _{G_r,} y _{G — r} , p ₂ ) is generated.

  FIG. 19 is a conceptual diagram of a decision boundary in the second embodiment. Referring to FIG. 19, a white circle represents class C1 without interference, and a black circle represents class C2 with interference. Also, the dotted circle represents the class C2 with interference as with the black circle.

When the interference data and the non-interference data are plotted on the space consisting of three parameters of the distance from the interference source, the transmission power, and the risk of occurrence of interference P ₁ , the interference data and the non-interference data are represented by a decision boundary BD consisting of a plane. To be separated. Here, the black circle and the dotted circle are on the back side of the decision boundary BD on the paper surface of FIG. 19, and the white circle is on the near side of the decision boundary BD on the paper surface of FIG.

FIG. 20 is a conceptual diagram of a decision boundary when the interference occurrence risk p ₁ is not used. Referring to FIG. 20, a white circle represents class C1 without interference, and a black circle represents class C2 with interference. When interference data and non-interference data are plotted on a plane composed of two parameters of the distance from the interference source and the transmission power, the decision boundary BD is a straight line. Then, although the interference data and the non-interference data are tried to be separated by the decision boundary BD, the interference data and the non-interference data may not be separated by the decision boundary BD.

However, as shown in FIG. 19, by adding the interference occurrence risk level p ₁ , the decision boundary BD can be set in a high-dimensional space, and the interference data and the non-interference data can be more accurately separated.

  A determination boundary determination method according to the second embodiment will be described. The determination device 1027 of the machine learning unit 102A determines the determination boundary BD using the same machine learning method as the machine learning method in the first embodiment.

  In the second embodiment, the space for determining the decision boundary BD is a three-dimensional space composed of three parameters as shown in FIG. 19, and a plane passing through the support vector is assumed. Accordingly, two planes passing through the support vector are set to ax + by + cz + d = 1 and ax + by + cz + d = −1. Since the two planes are parallel, the coefficients are the same.

  The area indicated by ax + by + cz + d ≦ 1 is a range where data of a positive example (white circle group without interference) exists, and the area indicated by ax + by + cz + d ≧ −1 is data of a negative example (group of black circles with interference). Is the range that exists.

  Here, two inequalities that are necessary conditions for drawing the identification surface are obtained, and these are combined into one. That is, in the positive example, the inequality is multiplied by “1” to be ax + by + cz + d ≦ 1, and in the negative example, the inequality is multiplied by “−1” to be −1 (ax + by + cz + d) ≧ −1.

  Then, if a variable t (t = 1 or t = −1) indicating whether it is a positive example or a negative example is used, t (ax + by + cz + d) ≦ 1.

Distance L ₂ perpendicular line drawn to the identification surface from the respective support vector positive cases and negative cases are represented by the following equation.

As a result, the position of the identification plane based on the maximum margin is obtained by obtaining a, b, c, d that minimizes the denominator of L ₂ while satisfying the condition of t (ax + by + cz + d) ≦ 1 for all of the teacher data. Can be determined. Determining the position of the identification surface based on this maximum margin is equivalent to determining the equidistant position from both the positive example (white circle group without interference) and the negative example (black circle group with interference). To do.

When the update teacher data is i (i = 1, 2, 3, 4,...), T ₁ (ax ₁ + by ₁ + cz ₁ + d) ≦ 1, t ₂ (ax ₂ + by ₂ + cz ₂ + d) ≦ 1,..., T _i (ax _i + by _i + cz _i + d) ≦ 1 is used to obtain a, b, c, and d that give the minimum value of a ² + b ² + c ² using i inequalities Become.

The problem of obtaining parameters for determining an identification plane based on the maximum margin is that if three variables x, y, and z and N update teacher data are provided, t _i (ax _i + by _i + cz _i + d) ≦ 1 ( It becomes a constrained minimization problem for obtaining a, b, c, d that minimizes a ² + b ² + c ² under the constraint condition of i = 1 to N).

A convenient method for solving such a constrained minimization problem is Lagrange's undetermined multiplier (coefficient) method. It is assumed that a ² + b ² + c ² to be minimized is f (a, b, c) = a ² + b ² + c ² , and the constraint condition t _i (ax _i + by _i + cz _i + d) ≦ 1 is g (x, y) = When t _i (ax _i + by _i + cz _i + d) −1 ≦ 0, Lagrange's undetermined multiplier method can be applied to convert the following equation (24) into equation (25).

  In equations (24) and (25), s. t. Means “obey the conditions described on the right side of st.” (The same applies hereinafter).

Equation (24) is expressed as g ₁ (a, b, c, d) ≦ 0, g ₂ (a, b, c, d) ≦ 0,..., G _n (a, b, c, d) ≦ This is a main problem because it means obtaining a, b, c that minimizes f (a, b, c) according to 0.

Expressions (25a) and (25b) in Expression (25) are the results of applying Lagrange's undetermined multiplier method. Equation (25b) represents a, b, c, d that minimizes L (a, b, c, d, μ ₁ ,..., Μ _n ). φ is a symbol indicating one function, similarly to f.

Equation (25c), (25d) are indicative of the dual _{problem, φ (μ 1, ···,} μ n) μ 1 to maximize, ..., by obtaining the mu _n, the main problem It is guaranteed that a, b, c, and d that give the minimum value are also obtained.

In the following, in order to simplify the explanation, the main problem is created with two teacher data, and the function L is created by applying Lagrange's undetermined multiplier method. However, a ² + b ² + c ² is multiplied by 1/2 for convenience of calculation.

t _i (ax _i + by _i + cz _i + d) ≦ 1 is transformed into t _i (ax _i + by _i + cz _i + d) −1 ≦ 0. As a result, the following equation is obtained.

  In order to create a dual problem function to be maximized, L is minimized for a, b, c, and d. Therefore, L is partially differentiated with respect to a, b, c, and d, and the result is set to 0. Then, the following equation is obtained.

  Then, the following equation is obtained from the upper three equations of equation (27).

In equations (26) and (27), t _i is the sign of the i-th updated teacher data, and x _i , y _i , and z _i are the x-coordinate, y-coordinate, and z-coordinate of the i-th teacher data. As such, t _i , x _i , y _i , and z _i are known.

Therefore, if μ _i that maximizes the function of the dual problem is known, a, b, and c can be obtained from Expression (28).

  Therefore, if all the conditions (formula (27)) obtained by partial differentiation are substituted for L in order to create a dual problem, the following formula is obtained.

  As a result, the dual problem is as follows.

Maximization of the function φ (μ ₁ , μ ₂ ) of the dual problem can be obtained by an interior point method (Non-Patent Document 7) that is a solution method of the convex quadratic programming problem. When emphasizing the speed for finding a solution, an evolutionary calculation such as a tabu search method (Non-Patent Document 8) may be used while considering the constraint conditions.

  The interior point method is a sequential iterative solution method that generates a point sequence that converges to an optimal solution within a constraint region of an optimization problem.

  The tabu search method is a kind of local search method that searches for a certain solution in the vicinity of the currently obtained solution, prevents the convergence to the local solution by not searching for a while as a tabu for the recently searched solution, The optimal solution is obtained by performing a global search of the solution space.

Therefore, points for increasing the function φ (μ ₁ , μ ₂ ) of the equation (30) are sequentially generated within the region of the constraint condition of the equation (30) (the equation described on the right side of the _st .), And the function φ (μ _1, μ ₂₎ optimal solution (μ _1, μ ₂₎ to maximize the seek.

In addition, when importance is attached to the speed for finding a solution, an optimum for performing a wide-area search and maximizing the function φ (μ ₁ , μ ₂ ) while preventing the convergence to a local solution in consideration of the constraint condition of Expression (30). Find the solutions (μ ₁ , μ ₂ ).

If μ ₁ and μ ₂ can be obtained, a, b, a and b are obtained from the equation (28) using known t ₁ , t ₂ , x ₁ , x ₂ , y ₁ , y ₂ , z ₁ , z ₂ . c can be obtained.

Then, d is the x-coordinate of a support vector and _{x S,} y coordinates and _{y S,} if the z-coordinate and _{z S, ax S + by S} + cz S + d = 1 ( or _{ax S + by S + cz S} + d = -1).

  As a result, a decision boundary BD for identifying the classes CL1 and CL2 with the maximum margin is determined.

  The other description of the method for obtaining the decision boundary in the second embodiment is the same as the description of the method for obtaining the decision boundary in the first embodiment.

  Therefore, the determination device 1027 determines the determination boundary BD by the method described above, and outputs the determined determination boundary BD to the communication prohibited area determination device 1028 of the parameter determination unit 103A.

FIG. 21 is a diagram for explaining a method of determining a communication prohibited area. Referring to FIG. 21, upon receiving the decision boundary BD from decision device 1027, communication prohibited region decision device 1028 divides the region covering update teacher data (x, p ₁ , l) into a plurality of meshes Msh_r. Each mesh Msh_r has, for example, a square shape with sides of 1 m to 10 m.

Transmission prohibition region determining unit 1028, the determination data _{(x G_r, y G_r, p} 2) receives the terrain data characteristic extraction unit 1026, the determination data _{(x G_r, y G_r, p} 2) each decision boundary Compared with the BD, the mesh Msh_r corresponding to the determination data classified as the region causing the interference is determined as the communication prohibited region.

In FIG. 21, if it is assumed that the region above the decision boundary BD is an area that causes interference and the region below the decision boundary BD is an area that does not cause interference, the determination is arranged above the decision boundary BD. The mesh Msh_r (hatched mesh) corresponding to the data (x _{G_r} , y _{G_r} , p ₂ ) is determined as a communication prohibited area, and the determination data (x _{G_r} , y arranged below the determination boundary BD) _The mesh Msh_r corresponding to _{G_r} , p ₂ ) is determined as the communication permitted area.

  FIG. 22 is a flowchart in the second embodiment for explaining operations of the primary user's terminal and the management apparatus. The flowchart shown in FIG. 22 is the same as the flowchart shown in FIG. 13 except that steps S11 to S14 in the flowchart shown in FIG. 13 are replaced with steps S11A, S12A, S13A, and S14A, respectively.

  Referring to FIG. 22, when a series of operations is started, the above-described steps S1 to S10 are sequentially executed.

After step S10, the machine learning unit 102A determines the decision boundary by using the interference occurrence risk _{p 1} (step S11A), and outputs the determined boundaries the determined to the parameter determining unit 103A.

  The parameter determination unit 103A receives the determination boundary from the machine learning unit 102A, and determines the communication prohibited area using the received determination boundary (step S12A). Then, parameter determination unit 103A outputs the communication prohibited area to parameter notification unit 104, and parameter notification unit 104 transmits the communication prohibited area received from parameter determination unit 103A to the terminal group of the secondary user (step S13A). ).

  The secondary user's terminal group receives the communication prohibited area, and updates the communication prohibited area based on the received communication prohibited area (step S14A). As a result, the operations of the primary user's terminal and the management apparatus are completed.

  FIG. 23 is a flowchart for explaining detailed operation of step S11A of FIG.

  Referring to FIG. 23, after step S10 of FIG. 22, the terrain data feature extraction device 1026 of the parameter determination unit 103A extracts the terrain data of the transmission path between the coordinates of the primary user and the secondary user. (Step S116).

  Then, the terrain data feature extraction apparatus 1026 estimates the received power when the position of the primary user is the reception point and the position of the secondary user is the transmission point in consideration of the terrain data by the method described above. (Step S117).

Thereafter, the terrain data feature extraction apparatus 1026 estimates the interference occurrence risk p ₁ based on the estimated received power by the method described above (step S118).

Then, the terrain data feature extraction apparatus 1026 generates update teacher data (x, p ₁ , l) to which the estimated interference occurrence risk p ₁ is assigned (step S119), and the generated update teacher data (x, p ₁ , 1) is output to the determination device 1027 of the machine learning unit 102A.

The determination device 1027 of the machine learning unit 102A receives the update teacher data (x, p ₁ , l) from the terrain data feature extraction device 1026 and based on the received update teacher data (x, p ₁ , l) 14 step S111 to step S115 are executed to determine the decision boundary (step S120), and the decided decision boundary is output to the communication prohibited area determination device 1028 of the parameter determination unit 103A. Thereafter, the series of operations proceeds to step S12A in FIG.

  FIG. 24 is a flowchart for explaining the detailed operation of step S12A of FIG.

Referring to FIG. 24, after step S11A in FIG. 22, communication prohibited area determining device 1028 in parameter determining unit 103A receives update teacher data (x, p) when receiving a determination boundary from determining device 1027 in machine learning unit 102A. An area covering ₁ , 1) is divided into a plurality of meshes (step S121).

Then, the communication prohibited area determination device 1028 extracts the center coordinates ( _{xG_r} , _{yG_r} ) of each of the divided meshes (step S122), and the extracted center coordinates ( _{xG_r} , _{yG_r} ). Is output to the terrain data feature extraction apparatus 1026.

The terrain data feature extraction apparatus 1026 receives a plurality of center coordinates (x _{G_r} , y _{G_r} ) from the communication prohibited area determination apparatus 1028. Then, the terrain data feature extraction apparatus 1026 performs a plurality of processes for estimating the received power when the center coordinates ( _{xG_r} , _{yG_r} ) are the transmission point and the primary user's position is the reception point by the above-described method. The process is executed for all the center coordinates ( _{xG_r} , _{yG_r} ) (step S123).

Subsequently, the geographic data characteristic extraction unit 1026, estimated by the method described above the interference occurrence risk p ₂ on the basis of the received power of the estimated (step S124).

Then, the terrain data feature extraction device 1026 generates the determination data (x _{G_r} , y _{G_r} , p ₂ ) by assigning the estimated interference occurrence risk p ₂ to the central coordinates (x _{G_r} , y _{G_r} ) ( step S125), the generated determination data _{(x G_r, y G_r,} outputs _{p 2)} to the transmission prohibition region determining unit 1028.

Transmission prohibition region determining unit 1028, the determination data _{(x G_r, y G_r, p} 2) receiving from the terrain data feature extraction unit 1026, the received determination data _{(x G_r, y G_r, p} 2) determining boundaries And the mesh corresponding to the determination data classified as the region where interference occurs is determined as the communication prohibited region (step S126). Thereafter, the series of operations proceeds to step S13A in FIG.

  In the second embodiment, the functions of the information management unit 101, the machine learning unit 102A, the parameter determination unit 103A, and the parameter notification unit 104 may be realized by a program PROG_B.

  In this case, the program PROG_B includes the above-described steps S3 to S10, S11A, S12A, and S13A (including steps S111 to S115, steps S116 to S120, and steps S121 to S126).

  The management device 10 includes a CPU, a ROM, and a RAM. The ROM stores a program PROG_B and terrain data.

  When determining the communication prohibited area, the CPU executes the program PROG_B out of the ROM. And CPU memorize | stores the various calculation results at the time of determining a communication prohibition area | region in RAM.

  Therefore, the program PROG_B is a program for causing a computer (CPU) to determine a communication prohibited area in order to perform wireless communication.

  In the second embodiment, the program PROG_B may be provided by being recorded on a recording medium such as a CD or a DVD. The user sets the recording medium on which the program PROG_B is recorded in the computer, and the computer (CPU) reads the program PROG_B from the recording medium and executes it.

  Therefore, the recording medium on which the program PROG_B is recorded is a recording medium that can be read by a computer (CPU).

  Other explanations in the second embodiment are the same as those in the first embodiment.

  According to the above-described first and second embodiments, the management device according to the embodiment of the present invention provides a wireless communication system in which the ratio of data indicating interference outside the communication prohibited area is a constant value. Receiving means for receiving an interference notification including the position of the interfered terminal, which is the position of the terminal that has received interference, and the presence or absence of interference, and a wireless communication system that performs wireless communication using the frequency band of the primary user Based on the user's communication history and interference notification history, a means for generating teacher data that correlates the label indicating the presence / absence of interference, the position of the interfered terminal, and the communication parameter, and indicates that there is interference The number of interference data in which the label, the interfered terminal position, and the communication parameter are associated with each other, and the label indicating that there is no interference, the interfered terminal position, and the communication parameter are associated with each other. An updating unit that applies a bias reduction process for reducing the bias with the number of data to the teacher data and generates updated teacher data in which the teacher data is updated, and n (n is 2) including the communication parameter based on the updated teacher data. A boundary for classifying communication parameters into a first class without interference and a second class with interference in an n-dimensional space composed of (integer) parameters, and the first and second classes Machine learning means for determining a decision boundary, which is a boundary existing at an equal distance from the machine learning, and a communication parameter set in which interference does not occur in the primary user and interference in the primary user using the decision boundary Communication parameter determining means for determining at least one of the communication parameter sets, and sending the determined communication parameter set to a terminal group of secondary users It is sufficient that a transmission means for.

  Further, a program for causing a computer to execute according to an embodiment of the present invention is a program for causing a computer to execute determination of a communication parameter set for performing wireless communication. A first notification is received from a primary user who is a wireless communication system in which a ratio of data indicating interference is a constant value, including an interfered terminal position that is a position of the interfered terminal and presence / absence of interference. And a label indicating the presence or absence of interference based on the communication history of the secondary user, which is a wireless communication system that performs wireless communication using the primary user's frequency band, and the history of interference notification. A second step of generating teacher data in which the interference terminal position and the communication parameter are associated with each other; and a label indicating that the update means has interference and the interfered terminal And the number of interference data in which communication parameters and communication parameters are associated with each other, and the number of non-interference data in which the label indicating that there is no interference, the interfered terminal position, and communication parameters are associated with each other. A third step of applying a bias reduction process to reduce the teacher data to generate updated teacher data in which the teacher data is updated, and the machine learning means includes n (n is 2) including communication parameters based on the updated teacher data. A boundary for classifying communication parameters into a first class without interference and a second class with interference in an n-dimensional space composed of (integer) parameters, and the first and second classes A fourth step of determining a decision boundary, which is a boundary existing equidistant from the machine, by machine learning, and the communication parameter determination means uses the determination boundary in the primary user A fifth step of determining at least one of a communication parameter set in which interference does not occur and a communication parameter set in which interference occurs in the primary user; and the transmission means transmits the determined communication parameter set to the terminal of the secondary user What is necessary is just a program for making a computer perform the 6th step transmitted to a group.

  The “communication prohibited area” in the second embodiment corresponds to a communication parameter set in which interference occurs in the primary user, and the “communication permitted area” in the second embodiment causes interference in the primary user. Corresponds to the communication parameter set.

  In the embodiment of the present invention, the information management unit 101 that receives the interference notification constitutes a “reception unit”.

  Further, in the embodiment of the present invention, the generation device 1021 that generates teacher data by the above-described method constitutes “generation means”.

Furthermore, according to the embodiment of the present invention, the above-described method is used to determine the number of interference data in which the label indicating that there is interference, the interfered terminal position, and the communication parameter are associated with each other, and that there is no interference. Applying a bias reduction process for reducing the bias of the number of non-interfering data in which the indicated label, the interfered terminal position, and the communication parameter are associated with each other to the teacher data, and generating updated teacher data by updating the teacher data The pseudo data generation device 1022, the excess data deletion device 1023, and the update device 1024 constitute “update means”.

  Furthermore, in the embodiment of the present invention, the determination device 1025 for determining the decision boundary BD by the above-described method constitutes “machine learning means”.

  Furthermore, in the embodiment of the present invention, the parameter determination unit 103 that determines a communication parameter set that does not interfere with the primary user by the above-described method constitutes a “parameter determination unit”.

  Furthermore, in the embodiment of the present invention, the parameter notification unit 104 constitutes “transmission means”.

Furthermore, in embodiments of the present invention estimates the interference occurrence risk p _1, topographical data characteristic extraction unit 1026 for imparting interference occurrence risk p ₁ was the estimated update training data (x, l) is The “first risk degree giving means” is configured.

Furthermore, in the embodiment of the present invention, the interference occurrence risk p ₂ is estimated, and the estimated interference occurrence risk p ₂ is _assigned to the center coordinates (x _{G — r} , y _{G — r} ) to determine data (x _{G — r).} , Y _{G — r} , p ₂ ), the terrain data feature extraction apparatus 1026 constitutes “second risk assigning means”.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description of the embodiments but by the scope of claims for patent, and is intended to include meanings equivalent to the scope of claims for patent and all modifications within the scope.

  The present invention is applied to a management apparatus, a program to be executed by a computer, and a computer-readable recording medium on which the program is recorded.

  1,11 radio station 2,3,12,13 terminal 10,10A management device 101 information management unit 102 102A machine learning unit 103 103A parameter determination unit 104 parameter notification unit 1021 generation device 1022 Pseudo data generation device, 1024 excess data deletion device, 1024 update device, 1025, 1027 determination device, 1026 terrain data feature extraction device, 1028 communication prohibited region determination device.

Claims (25)

Receiving means for receiving, from a primary user who is a wireless communication system in which the ratio of data indicating interference is a fixed value, an interference notification including the location of the interfered terminal, which is the location of the interfered terminal, and the presence or absence of interference; ,
A label indicating presence / absence of interference based on a communication history of a secondary user who is a wireless communication system that performs wireless communication using the frequency band of the primary user and the history of the interference notification, and the position of the interfered terminal Generating means for generating teacher data in which communication parameters are associated with each other;
The number of interference data in which the label indicating that there is interference, the interfered terminal position, and the communication parameter are associated with each other, the label indicating that there is no interference, the interfered terminal position, and the communication parameter, An update unit that applies a bias reduction process for reducing a bias with the number of non-interference data associated with each other to the teacher data, and generates update teacher data in which the teacher data is updated;
Based on the update teacher data, the communication parameter includes an n-dimensional space composed of n (n is an integer of 2 or more) parameters, and the communication parameter includes a first class that has no interference and a second class that has interference. Machine learning means for determining, by machine learning, a decision boundary that is a boundary that is classified as follows and is a boundary that is equidistant from the first and second classes;
Communication parameter determining means for determining at least one of a communication parameter set in which interference does not occur in the primary user and a communication parameter set in which interference occurs in the primary user, using the determination boundary;
A management apparatus comprising: a transmission unit that transmits the determined communication parameter set to the terminal group of the secondary user.   The communication parameter determining means sets (n−m) -dimensional hyperplane in which m parameters (m is an integer satisfying 1 ≦ m <n) that cannot be changed among the n parameters are set, and the setting is performed. The point on the (nm) dimensional hyperplane that maximizes or minimizes the gain function is searched using the distance between the point on the (nm) dimensional hyperplane and the decision boundary as an index, and the search is performed. The management apparatus according to claim 1, wherein a communication parameter set that does not cause interference in the primary user is determined from a point on an (nm) -dimensional hyperplane using a constraint condition.   The update means updates the teacher data by increasing the number of the interference data and decreasing the number of the non-interference data, and generates the update teacher data according to claim 1 or 2. Management device.   The management according to claim 3, wherein the update unit generates pseudo interference data from interference data included in the teacher data, and generates the updated teacher data by deleting non-interference data included in the teacher data. apparatus.   The update means uses the interference data included in the teacher data as a reference point, and the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a predetermined distribution. The management apparatus according to claim 4, wherein the update teacher data is generated by generating interference data and deleting non-interference data existing outside a circle centered on the reference point.   The update means generates the pseudo interference data so that the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a normal distribution when the variance is changed. The management device according to claim 5. When the coordinate point of the update teacher data received from the updating means is the transmission point and the primary user's position is the reception point, the received power is estimated in consideration of the terrain data, and the estimated received power is A first risk generation means for estimating a first risk of occurrence of interference based on the estimated first risk of risk of interference to the updated teacher data;
When receiving a decision boundary determined based on the updated teacher data to which the first interference occurrence risk is given, the update teacher data is divided into a plurality of regions, and the coordinates of the divided one region The received power is estimated when the point is the transmission point and the position of the primary user is the reception point, the second interference occurrence risk is estimated based on the estimated received power, and the estimated second A second risk level assigning unit that executes the process of generating the determination data by giving the interference occurrence risk level to the one area for all of the plurality of areas,
The machine learning means determines the decision boundary by machine learning based on the updated teacher data to which the first interference occurrence risk is given,
The communication parameter determining means is provided with a determination boundary determined based on the updated teacher data to which the first interference occurrence risk is assigned, the determination data, and the second interference occurrence risk. 2. It is determined whether each of the plurality of areas is a communication prohibited area based on a plurality of areas, and at least one of a communication prohibited area and a communication permitted area is determined as the communication parameter set. The management apparatus according to claim 6.   The communication parameter determining means compares the determination data with the determination boundary, and determines, as the communication prohibited area, an area corresponding to the determination data classified as an area causing interference among the plurality of areas. The management apparatus according to claim 7.   The first risk level assigning unit estimates the estimated received power as the first interference occurrence risk level, or the estimated received power at a position of interference data closest to a point where the received power is obtained. The management apparatus according to claim 7 or 8, wherein a division result obtained by dividing the received power is estimated as the first interference occurrence risk level.   The first risk assigning means holds a plurality of radio wave propagation characteristics as radio wave propagation characteristics between the primary user and the secondary user, and the radio wave propagation characteristics that best match the topographic data. The management apparatus according to any one of claims 7 to 9, wherein an estimation process for selecting the received power from the plurality of radio wave propagation characteristics and estimating the reception power using the selected radio wave propagation characteristics is performed.   The management apparatus according to claim 10, wherein the second risk level assigning unit performs the same estimation process as the estimation process in the first risk level granting unit to estimate the received power.   The management device according to any one of claims 7 to 11, wherein the topographic data is an altitude map or an aerial photograph. A program for causing a computer to execute determination of a communication parameter set for performing wireless communication,
The receiving means receives an interference notification including the interfered terminal position, which is the position of the interfered terminal, and the presence / absence of interference from a primary user who is a wireless communication system in which the ratio of data indicating interference is a constant value. A first step to:
The generation means includes a label indicating presence / absence of interference based on a communication history of a secondary user, which is a wireless communication system that performs wireless communication using the frequency band of the primary user, and the history of the interference notification, A second step of generating teacher data in which the interference terminal position and the communication parameter are associated with each other;
The updating means includes a label indicating that there is interference, the number of interference data in which the interfered terminal position and the communication parameter are associated with each other, a label indicating that there is no interference, the interfered terminal position, and the Applying a bias reduction process that reduces a bias with the number of non-interference data associated with communication parameters to each other to generate updated teacher data in which the teacher data is updated;
Based on the updated teacher data, the machine learning means causes the communication parameter to interfere with the first class having no interference in an n-dimensional space composed of n (n is an integer of 2 or more) parameters including the communication parameter. A fourth step of determining by machine learning a decision boundary that is a boundary for classification into a second class and is a boundary that is equidistant from the first and second classes;
A communication parameter determining means uses the determination boundary to determine at least one of a communication parameter set that does not cause interference in the primary user and a communication parameter set that causes interference in the primary user. Steps,
A program for causing a computer to execute a sixth step of transmitting the determined communication parameter set to the terminal group of the secondary user.   In the fifth step, the communication parameter determination means fixes m parameters (m is an integer satisfying 1 ≦ m <n) that cannot be changed among the n parameters. A point on the (nm) dimensional hyperplane that sets a plane and maximizes or minimizes the gain function using the distance between the set point on the (nm) dimensional hyperplane and the determined boundary as an index The communication parameter set in which interference does not occur in the primary user is determined from a point on the searched (n−m) -dimensional hyperplane using a constraint condition. Program to let you.   The update means updates the teacher data to generate the updated teacher data by increasing the number of the interference data and decreasing the number of the non-interference data in the third step. Or the program for making the computer of Claim 14 perform.   In the third step, the update unit generates pseudo interference data from interference data included in the teacher data, and deletes non-interference data included in the teacher data to generate the updated teacher data. A program for causing a computer according to claim 15 to be executed.   In the third step, the updating means uses the interference data included in the teacher data as a reference point, and the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point are predetermined. 17. The computer according to claim 16, wherein the pseudo-interference data is generated so as to follow the distribution of the non-interference data existing outside a circle centered on the reference point, and the update teacher data is generated by deleting the non-interference data. A program to be executed.   In the third step, the updating means is configured so that the pseudo interference data generated based on the interference data of the reference point and the interference data of the reference point follow a normal distribution when variance is changed. The program for making a computer run of Claim 17 which produces | generates pseudo | simulation interference data. When the first risk level assigning means uses the coordinate point of the updated teacher data received from the updating means as the transmission point and the primary user's position as the reception point, the received power is estimated in consideration of the terrain data. And a seventh step of estimating a first interference occurrence risk based on the estimated received power and assigning the estimated first interference occurrence risk to the updated teacher data;
When the second risk level assigning unit receives the decision boundary determined based on the update teacher data to which the first interference occurrence risk level is given, the update risk data existence area is divided into a plurality of areas. The received power is estimated when the coordinate point of the divided one region is used as a transmission point and the position of the primary user is used as a reception point, and a second interference occurrence risk is determined based on the estimated received power. And executing the process of generating determination data by assigning the estimated second risk of occurrence of interference to the one area for all of the plurality of areas. Let
In the fourth step, the machine learning means determines the determination boundary by machine learning based on the updated teacher data to which the first interference occurrence risk is given.
In the fifth step, the communication parameter determining means determines the determination boundary determined based on the updated teacher data to which the first interference occurrence risk is given, the determination data, and the second interference. And determining whether each of the plurality of areas is a communication prohibited area based on the plurality of areas to which the occurrence risk is given, and setting at least one of the communication prohibited area and the communication permitted area as the communication parameter set The program for making the computer of any one of Claim 13 to 18 determine as follows.   In the fifth step, the communication parameter determination unit compares the determination data with the determination boundary, and selects a region corresponding to the determination data classified as a region causing interference from the plurality of regions. The program for making the computer of Claim 19 determine as said communication prohibition area | region.   In the seventh step, the first risk level assigning unit estimates the estimated received power as the first interference occurrence risk level, or uses the estimated received power as a point where the received power is obtained. 21. A program for causing a computer according to claim 19 or 20 to estimate a division result obtained by dividing received power at a position of closest interference data as the first interference occurrence risk.   In the seventh step, the first risk assigning unit holds a plurality of radio wave propagation characteristics as radio wave propagation characteristics between the primary user and the secondary user, and the topographic data The radio wave propagation characteristic that most closely matches the radio wave propagation characteristic is selected from the plurality of radio wave propagation characteristics, and the estimation process for estimating the received power using the selected radio wave propagation characteristic is executed. A program for causing a computer described in the section to be executed.   The said 2nd risk provision means performs the estimation process same as the said estimation process in the said 7th step in the said 8th step, and estimates the said received power, It performs to the computer of Claim 22 Program to let you.   The computer program according to any one of claims 19 to 23, wherein the terrain data includes an altitude map or an aerial photograph.   A computer-readable recording medium on which the program according to any one of claims 13 to 24 is recorded.

Images