JP6204889B2 - COMMUNICATION AREA ESTIMATION SYSTEM, SERVER DEVICE, COMMUNICATION AREA ESTIMATION METHOD, AND COMMUNICATION AREA ESTIMATION PROGRAM - Google Patents

Description

  The present invention relates to a communication area estimation system, a server device, a communication area estimation method, and a communication area estimation program for estimating a wireless communication area.

  In recent years, with the spread of self-employed indoor wireless communication systems such as a wireless local area network (LAN), interference with adjacent access points and terminals thereunder has become a problem. In order to reduce these interferences, it is necessary to design an appropriate communication area for one's own access point while avoiding interference with a communication area of an adjacent access point. The communication area design is to determine an appropriate transmission power setting and an antenna installation position. In order to accurately design a communication area in a home environment where the influence of shadowing due to a shield such as furniture greatly varies depending on the antenna installation position, it is necessary to estimate indoor propagation characteristics at an arbitrary antenna installation position. For indoor propagation characteristics, there is a method of estimating a communication area using a statistical model expressed by the average propagation loss and the statistical properties of shadowing (see Non-Patent Document 1, for example).

  FIG. 11 is a block diagram showing a conventional communication area estimation system to which the method of Non-Patent Document 1 is applied. Moreover, FIG. 12 is a figure which shows the concept of the conventional communication area estimation system. As shown in FIG. 11, conventionally, a reception level is estimated using a uniform statistical model (Non-Patent Document 1) with the position of an access point (AP) and the distance from the access point as parameters. Was estimating the communication area. Thereby, the concentric communication area shown in FIG. 12 can be mapped. However, as shown in FIG. 12, in the conventional method, it is determined that the non-line-of-sight environment (triangular area shown in FIG. 12) cannot be communicated.

“RECOMMENDATION ITU-R P.1238-1 Propagation data and prediction methods for the planning of indoor radiocommunication systems and radio local area networks in the frequency range 900 MHz to 100 GHz”, October 2009

  As described above, in the conventional method described in Non-Patent Document 1, in an indoor propagation environment, the influence of multipath due to diffraction / reflection of an object such as a wall in a room or a fixture greatly varies depending on the area. It is difficult to estimate the propagation characteristics at any place necessary for communication area design. In particular, the conventional method that does not take into account multipath waves due to diffraction and reflection determines an out-of-sight environment in which direct waves are blocked as an incommunicable region, but there are actually communicable regions. As described above, since there are regions having different propagation characteristics even in the same area, there is a problem that a communication area estimation error is large in a unique model.

  Further, since propagation fluctuations due to objects such as walls and furniture are unique to each building / room, it is necessary to estimate propagation characteristics unique to each building / room. There is a site survey as a method for accurately estimating the inherent propagation characteristics. Site survey is to collect radio wave environment information while moving in the area with a wireless terminal for measurement at the actual location. To obtain data that satisfies sufficient accuracy in the site survey, Experts with propagation analysis skills needed to collect data after understanding the fluctuations in each area, which was a heavy burden on the acquirer.

  The present invention has been made in view of such circumstances, and provides a communication area estimation system, a server device, a communication area estimation method, and a communication area estimation program capable of estimating a communication area in wireless communication with high accuracy. For the purpose.

  The present invention is a communication area estimation system that includes a wireless terminal and a server device and estimates a communication area of the wireless terminal, and the wireless terminal is within a predetermined measurement range in accordance with an instruction from the server device. A data measurement unit that acquires reception level information for each area surrounded by an area boundary while it exists, a data measurement unit that acquires its own location information, and a data transmission unit that transmits the reception level information and the location information to the server device, The server device includes: a logical area database that stores area information logically divided in advance; an actual measurement database that stores actual measurement values; a categorization database that stores information on the results of categorizing acquired information; Based on the position information transmitted from the wireless terminal, the position of the wireless terminal is managed, and the wireless terminal A data acquisition instruction unit that transmits a data acquisition instruction to the wireless terminal when entering the measurement range to be determined, and the reception level information transmitted from the wireless terminal in response to the instruction A data reception unit that receives and stores the logically divided area information stored in the logical area database and the reception level information in the actual measurement database; and information stored in the actual measurement database. A learning unit that categorizes the reception level information according to a difference in feature quantity of the reception level information and stores a categorization result in the categorization database; and a radio that is assigned to each category based on the information stored in the categorization database. An area characteristic determination unit that estimates the maximum likelihood of the parameters of the propagation model, and the maximum likelihood estimated parameter. Characterized in that it comprises the area estimation unit that estimates a communication area of the radio from the propagation model of each category to the over data based.

  In the present invention, the learning unit maps the reception level information acquired on a feature space centered on the reception level information resulting from the characteristics of the propagation model to be generated, and a logical area associated with the reception level information. Using the learning vector quantization method with the information as teacher information, the received reception level information is categorized, the area is divided by the similarity in the feature space of the reference vector, and the position information of the reference vector in the feature space is The reception level information is categorized according to the difference in the feature amount by being stored in a categorization database.

  The present invention is a server device in a communication area estimation system that includes a wireless terminal and a server device, and estimates a communication area of the wireless terminal, and stores a logical area database that preliminarily logically divides area information; The wireless terminal manages the position of the wireless terminal based on the actual value database that stores the actual measured value, the categorized database that stores information on the result of categorizing the acquired data, and the position information transmitted from the wireless terminal. When the communication area is within the measurement range to be estimated, a data acquisition instruction unit that transmits a data acquisition instruction to the wireless terminal, and a reception transmitted from the wireless terminal in response to the instruction The level information is received and the area information logically divided into the logical area database and the reception level information are linked. The reception level information is categorized according to the feature amount of the reception level information using the data reception unit stored in the actual measurement value database and the information stored in the actual measurement value database, and the categorization result is categorized. Based on the learning unit stored in the database, the area characteristic determining unit that estimates the maximum likelihood of the radio propagation model parameter assigned to each category based on the information stored in the categorized database, and the maximum likelihood estimated parameter And an area estimation unit that estimates a wireless communication area from a propagation model of each category.

  The present invention is a communication area estimation method that includes a wireless terminal and a server device, and is performed by a communication area estimation system that estimates a communication area of the wireless terminal. The server device is an area that is logically divided in advance. A logical area database for storing information; an actual measurement database for storing actual measurement values; and a categorization database for storing information on the result of categorizing acquired data, wherein the server device transmits position information transmitted from the wireless terminal. A data acquisition instruction step of managing the position of the wireless terminal and transmitting a data acquisition instruction to the wireless terminal when the wireless terminal enters a measurement range in which a communication area should be estimated; The wireless terminal is surrounded by an area boundary while existing within a predetermined measurement range according to an instruction from the server device. A reception level information for each area; a data measurement step for acquiring own location information; a data transmission step for the wireless terminal to transmit the reception level information and the location information to the server device; In response to the instruction, the reception level information transmitted from the wireless terminal is received, and the logically divided area information stored in the logical area database and the reception level information are linked to the actual measurement value. A data receiving unit stored in a database and the server device categorize the reception level information according to a difference in feature quantity of the reception level information using information stored in the actual measurement value database, and categorize a categorization result. A learning step stored in a database, and the server device includes the categorized data An area characteristic determination step for estimating the maximum likelihood of a radio propagation model parameter to be assigned to each category based on information stored in the source; and the propagation of each category based on the parameter estimated by the server And an area estimation step for estimating a wireless communication area from the model.

  In the learning step, in the learning step, the reception level information acquired on a feature space centered on the reception level information resulting from the characteristics of the propagation model to be generated is mapped, and a logical area associated with the reception level information is mapped Using the learning vector quantization method with the information as teacher information, the received reception level information is categorized, the area is divided by the similarity in the feature space of the reference vector, and the position information of the reference vector in the feature space is The reception level information is categorized according to the difference in the feature amount by being stored in a categorization database.

  The present invention is a communication area estimation program for causing a computer to function as the server device.

  According to the present invention, the area is subdivided according to the characteristics of the propagation environment, and the parameters of the propagation model are estimated for each area. Therefore, it is possible to perform highly accurate communication area estimation in wireless communication. It is done.

It is a block diagram which shows the structure of 1st Embodiment of this invention. 4 is a diagram illustrating an example in which a training data set output from a data measurement unit 11 is stored in an actual measurement database 31. FIG. It is a flowchart which shows operation | movement of the feature space mapping part 24 shown in FIG. It is a flowchart which shows the processing operation of the vector learning part 26 shown in FIG. It is a figure which shows the example of the reference vector after learning. It is a figure which shows the model assumed by the experiment of a communication area estimation system. It is a figure which shows the output result of the vector learning part 26 shown in FIG. It is a figure which shows the output result of the area characteristic determination part 27 shown in FIG. It is a figure which shows the output result of the area estimation part 28 shown in FIG. It is explanatory drawing which shows a transmission model. It is a block diagram which shows the conventional communication area estimation system to which the method of a nonpatent literature 1 is applied. It is a figure which shows the concept of the conventional communication area estimation system.

<First Embodiment>
Hereinafter, a communication area estimation system according to a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the embodiment. In this figure, reference numeral 1 denotes a wireless terminal (hereinafter simply referred to as a terminal) used by a user. Reference numeral 2 denotes a server that estimates a communication area. The terminal 1 includes a data measurement unit 11 and a data transmission unit 12. The server 2 includes a terminal position monitoring unit 21, a data acquisition instruction unit 22, a data receiving unit 23, a feature space mapping unit 24, an initial value generation unit 25, a vector learning unit 26, an area characteristic determination unit 27, an area estimation unit 28, a terminal A location database 29, a logical area database 30, an actual measurement database 31, and a categorization database 32 are provided.

  The terminal location monitoring unit 21, the data acquisition instruction unit 22, and the data reception unit 23 constitute an acquisition unit 33. The feature space mapping unit 24, the initial value generation unit 25, and the vector learning unit 26 constitute a learning unit 34. The area characteristic determination unit 27 and the area estimation unit 28 constitute an estimation unit 35. The communication area estimation system shown in FIG. 1 is characterized in that the user's terminal 1 and the server 2 are linked and logical area information is assigned to the acquired data as teacher information.

ここで、Ｌ_{ｔｏｔａｌ}は減衰量、Ｎは減衰率、ｆは周波数（ＭＨｚ）、ｄは基地局と端末間の距離（ｍ）［ｄ＞１ｍ］、Ｌ_ｆは透過減衰率（ｄＢ）、ｎは基地局（アクセスポイント）と端末間の階数（ｎ≧１）である。 Next, operation | movement of the acquisition part 33 of the communication area estimation system shown in FIG. 1 is demonstrated. Here, the statistical model described in ITU-R is assumed as a propagation model. The statistical model is shown in equation (1).

Here, L _total is the amount of attenuation, N is the attenuation rate, f is the frequency (MHz), d is the distance between the base station and the terminal (m) [d> 1 m], L _f is the transmission attenuation rate (dB), n Is the rank (n ≧ 1) between the base station (access point) and the terminal.

  First, the terminal position monitoring unit 21 refers to the terminal position database 29 to acquire the terminal position of the terminal 1, and when the terminal 1 enters the measurement range A in which the communication area should be estimated, the acquisition start trigger is set as data. The data is output to the acquisition instruction unit 22. The data acquisition instruction unit 22 instructs the data measurement unit 11 of the terminal 1 to acquire data.

In response to this, the data measurement unit 11 of the terminal 1 receives the reception level, phase, throughput, sensor information, signal-to-noise power ratio, distance attenuation amount, delay profile, arrival in accordance with the instruction of the data acquisition instruction unit 22 of the server 2. Acquire any information of wave direction and own position information. The delay profile can be obtained by a method of performing a correlation operation between a received signal and a known signal, or a method of performing an inverse discrete Fourier transform from a frequency spectrum, and a delay spread can be calculated therefrom (for example, see Reference 1).
Reference 1: Koki Shibata, “Delay Profile Estimation Method in OFDM Transmission”, Journal of the Institute of Image Information and Television Engineers, Vol.60, No.10, pp.1672-1680, 2006.

  For throughput measurement, a throughput analysis tool is used to generate UDP / TCP traffic from the terminal 1 on the transmission side and measure the throughput on the reception side. Alternatively, the arrival direction distribution may be estimated using an array antenna. In addition, the data transmission unit 12 of the terminal 1 transmits the data acquired by the data measurement unit 11 to the data reception unit 23 of the server 2 and stores the position information in the terminal position database 29 of the server 2.

  The server 2 receives the acquired data from the terminal 1 by the data receiving unit 23, and in order to subdivide according to the characteristics of each area, the logical area information held in the logical area database 30 (such as the inside / outside of the line of sight, near the wall) The training data set is stored in the measured value database 31 by associating the location name) with the teacher information. FIG. 2 is a diagram illustrating an example in which the training data set output from the data measurement unit 11 is stored in the actual measurement value database 31. In FIG. 2, the transmitting station may be a wireless LAN terminal or a sensor terminal.

  Next, the operation of the learning unit 34 shown in FIG. 1 will be described. In the learning unit 34, the feature space mapping unit 24 maps the feature space about the data that can be acquired by the terminal 1 as information on the environment side based on the information in the actual measurement database 31. The feature space may be selected in all combinations, or the feature space may be determined by the user selecting manually.

FIG. 3 is a flowchart showing the operation of the feature space mapping unit 24. The feature space mapping unit 24 accumulates data for each selected vector (input vector) (step S1), calculates a relative value of the accumulated data (step S2), normalizes it, and maps it to the feature space (step S1). S3). The relative value is calculated using the maximum value and the minimum value, or calculated using the average value and the standard deviation. That is, the relative value V ′ is calculated using the formula (2) or the formula (3). The input vector is V, the minimum value is V _min , the maximum value is V _max , the average value is V 値 (￣ is on V), and the standard deviation is σ.

As a method of categorizing the feature space by learning, there is LVQ (Learning Vector Quantization). LVQ treats an input data set as a vector, plots it in a feature space with the values of each element of the vector as an axis, and makes the area represented by a reference vector, thereby assigning the feature space to the input vector. This is a technique for drawing an optimal boundary line for subdividing by information. This operation has been proven to form a Bayesian identification boundary, which is the theoretically optimal area identification boundary (see, for example, Reference 2).
Reference 2: T. kohonen, GBarna, and R. Chrisley, “Statistical pattern recognition with neural networks”, Proc. ICNN, val. I, pp. 62-68, 1988

Next, the initial value generation unit 25 outputs a random value with the same elements as the training data as the initial value of the reference vector. The vector learning unit 26 estimates the barycentric position of a reference vector representing an area using categorized machine learning. FIG. 4 is a flowchart showing the processing operation of the vector learning unit 26 shown in FIG. Here, the input vector is V _n and the reference vector is W _m . First, when a learning step is started (step S11), the vector learning unit 26 performs an iterative process for each input vector (step S12). The vector learning unit 26 searches for the reference vector W _m (m = 1,..., M) closest to the input vector V _n using the Euclidean distance equation (4) (step S13). Here, t indicates the number of learning steps.

Next, the vector learning unit 26 determines whether or not the category of the input vector and the searched reference vector are the same (step S14), and compares the searched logical area information assigned to W _m and V _n to be the same. In this case, W _m is brought closer to V _n using the equation (5) (step S15), and if different, W _m and V _n are moved away using the equation (6) (step S16). α (t) is a monotone decreasing function that decreases in accordance with the number of learning steps, and is defined as in equation (7) when a linear function is used. α ₀ is an initial learning rate, 0 <α ₀ <1 is satisfied, and l _max is a preset maximum number of learning times.

  Next, the vector learning unit 26 repeats the movement of the reference vector for all the input vectors (n = 1,..., N), and then resets the logical area information of the reference vector (step S17). The method of resetting is to search the reference vectors that are closest to all input vectors, and for each reference vector, the number searched for the input vectors is totaled for each area, and the logical area information searched for the most input vectors is obtained. Assign to that reference vector.

  The movement of the reference vector and the resetting of the logical area information of the reference vector are repeated as the maximum learning number set in advance as one iteration process. FIG. 5 is a diagram illustrating an example of a reference vector after learning. As shown in FIG. 5, a reference vector and logical area information are output for each selected feature space and stored in the categorization database 32.

Next, the operation of the 35 estimation unit shown in FIG. 1 will be described. The area characteristic determination unit 27 searches for a reference vector closest to each input vector, and assigns the same logical area information as the reference vector to the input vector. For each area, the transmission / reception terminal position is extracted from the terminal position database 29 and the reception level is extracted from the categorization database 32. The propagation distance d and the reception level RSSI are input, and the power loss coefficient a satisfying the evaluation function of equation (8) is fixed. Among the combinations with the constant b, the combination that minimizes the evaluation function is selected. I is the number of data used.

  Next, the area estimation unit 28 calculates the absolute position of the access point, the distance from the access point, and the distance obtained for each area in the area characteristic determination unit 27 and the reception level, and calculates the relationship of the target wireless device. Estimate the communication area. The power loss coefficient a calculated for each area is set as the attenuation factor in the equation (1), the fixed constant b is set as the loss due to the frequency, the transmission attenuation factor and the constant term in the equation (1), and the propagation loss from the absolute position of the access point is Calculate a reception level map for each device.

  This implementation process includes an iterative process in which the area characteristic determination unit 27 calculates an area correct answer rate (%) by comparing reference vector groups having different times in the categorized database 32 and updates the logical area database 30 based on the correct answer rate. Thus, the logical area information may be set dynamically. The area correct answer rate calculation process compares the logical area information of the reference vector obtained at the time based on each measurement range defined in the logical area database 30 with the logical area information obtained at another time, The rate at which the logical area information matches is calculated as the correct answer rate.

  FIG. 6 is a diagram illustrating a model assumed in an experiment of the communication area estimation system. In this experiment, the boundary of the area is defined as a wall surface, and the learning unit 34 maps it onto the feature space around each reception level of the access points (AP1 to AP4) installed in the areas 1, 2, 4 and 5, The center of gravity of the reference vector representing the area was estimated, and the correct answer rate for each area was calculated. FIG. 7 is a diagram illustrating an experimental result of the communication area estimation system described above and an output result of the vector learning unit 26. FIG. 8 is a diagram illustrating an experiment result of the communication area estimation system according to the first embodiment and an output result of the area characteristic determination unit 27. FIG. 9 is a diagram illustrating an output result of the area estimation unit 28.

  It can be seen from FIG. 7 that the reference vector representing the area is calculated by the vector learning unit 26 of the learning unit 34 and the identification boundary dividing the category can be estimated. FIG. 9 shows the relationship between the distance from the absolute position of the access point 3 (AP3) and the reception level for each area. Actual measurement values are indicated by ◯, ×, □, Δ, +, and estimated values for each area are indicated by lines. At this time, since the estimated reference vector is used to estimate the parameters of the propagation model for each area and the reception level can be estimated for each subdivided area, the accuracy is improved.

<Second embodiment>
Next, a communication area estimation system according to the second embodiment of the present invention will be described. In the present embodiment, in order to accurately estimate the transmission loss of building materials, the Fresnel transmission coefficient is assumed as a propagation model. Assuming that the wall of the building material is sufficiently larger than the wavelength, the transmission coefficient is incident on the medium 2 having a finite thickness d sandwiched between the medium 1 and the medium 3 as shown in FIG. Using the Fresnel transmission coefficient when obliquely incident at an angle, calculation is performed using equations (9) and (10). FIG. 10 is an explanatory diagram showing a transmission model.

Here, n ₁₂ is the relative complex refractive index of medium 1 and medium 2, λ ₁ is the wavelength in medium 1, ε ₁ is the dielectric constant in medium 1, ε ₂ is the dielectric constant in medium 2, and μ ₁ is the medium. 1, μ ₂ is the permeability in the medium 2, σ ₁ is the conductivity in the medium 1, σ ₂ is the conductivity in the medium 2, ω is the angular frequency, β is the thickness d and in the medium 1 Is a function of the wavelength λ. Further, t ₁₂ is a Fresnel transmission coefficient when an incident wave is transmitted from the medium 1 to the medium 2, and t ₂₃ is a Fresnel transmission coefficient when an incident wave is transmitted from the medium 2 to the medium 3. T _‖ (TE incidence) and t _⊥ (TM incidence) in the case of transmitting from the light to the medium 2 are expressed as follows.

(11) shows the Fresnel transmission coefficient _{t 12} at TE incidence, (12) shows a Fresnel transmission coefficient _{t 12} during TM incidence. For even t _23, it can be obtained by replacing the subscript. At this time, the relative complex refractive index of the medium 2 and the medium 3 is n ₂₃ , and the wavelength in the medium 2 is λ ₂ .

Similarly, the reflection coefficients r _‖ (TE incidence) and r _⊥ (TM incidence) when transmitting from the medium 1 to the medium 2 are expressed as follows.

Equation (13) shows the reflection coefficient r ₁₂ when TE is incident, and equation (14) shows the reflection coefficient r ₁₂ when TM is incident. For r ₂₃ also can be obtained by replacing the subscript.
By substituting Equations (11), (12), (13), and (14) into Equation (9), the transmission characteristics for each polarization in the transmission model of FIG. 10 can be calculated.

In this embodiment, the process of estimating the maximum likelihood solution of the parameters of the propagation model is different. Therefore, the operation of the estimation unit 35 will be described. The area characteristic determination unit 27 searches for a reference vector closest to each input vector, and assigns the same logical area information as the reference vector to the input vector. For each area, the use frequency, the transmission / reception terminal position, and the transmission attenuation amount are extracted, the incident angle θ and the transmission attenuation amount _TMeasure (dB) are input, and the relative permeability μ and the ratio satisfying the evaluation function of Expression (15) are satisfied. Among the combinations of dielectric constant ε, conductivity σ, and thickness d, the combination that minimizes the evaluation function is selected. I is the number of data used.

  The area estimation unit 28 calculates the relational expression between the absolute position of the access point, the distance from the access point, and the incident angle and attenuation obtained for each area by the area characteristic determination unit 27, and near the building material of the target wireless device Estimate the communication area. Calculate (9) from the relative permeability, relative permittivity, conductivity, and thickness calculated for each area, calculate the free propagation attenuation from the distance from the access point, and calculate the reception level map near the building material To do.

  In this way, the server 2 manages the position of the terminal 1 carried by the user, and when the terminal 1 is within the desired measurement range, instructs the terminal 1 to acquire data, and the terminal 1 follows this instruction. Data including the reception level information of the radio signal from the acquired access point and the position information of the terminal 1 is received, and the received data is stored in the 31 measured value database in association with the logical area information. The accumulated data is categorized according to the difference in feature quantity, accumulated in the categorized database 32, and the maximum likelihood estimation is performed on the parameters of the radio propagation model corresponding to each category based on the data accumulated in the categorized database 32. The wireless communication area was estimated based on the parameters.

  According to this configuration, the area can be subdivided according to the characteristics of the propagation environment, the parameters of the radio propagation model for each area can be estimated, and the accuracy of radio communication area estimation can be improved.

  As described above, by categorizing machine learning, it is possible to perform highly accurate communication area estimation by subdividing an area according to the characteristics of the propagation environment and estimating the parameters of the propagation model for each area. Furthermore, the server 2 and the user's terminal 1 are linked, and as a necessary information for machine learning, the server links logical area information to the information acquired by the terminal from the position information, thereby enabling a simple site survey. It can be carried out.

  According to this configuration, the server 2 and the user's terminal 1 are linked, and the server 2 associates logical area information with the information acquired by the terminal 1 as teacher information, so that categorized machine learning can be applied. It becomes like this. Furthermore, by categorizing machine learning, data that can be acquired as the topology of the radio propagation characteristics is mapped to the feature space around the axis, the area is subdivided according to the characteristics of the propagation environment, and the parameters of the propagation model are estimated, Since modeling for each area becomes possible, communication area estimation accuracy can be improved.

  You may make it implement | achieve the function of the terminal 1 in the embodiment mentioned above, and the server 2 with a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory inside a computer system serving as a server or a client in that case may be included and a program held for a certain period of time. Further, the program may be for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in the computer system. It may be realized using hardware such as PLD (Programmable Logic Device) or FPGA (Field Programmable Gate Array).

  As mentioned above, although embodiment of this invention has been described with reference to drawings, the said embodiment is only the illustration of this invention, and it is clear that this invention is not limited to the said embodiment. is there. Therefore, additions, omissions, substitutions, and other modifications of the components may be made without departing from the technical idea and scope of the present invention.

  By categorizing machine learning, the area can be subdivided according to the characteristics of the propagation environment, and by estimating the parameters of the propagation model for each area, it can be applied to applications where it is essential to perform highly accurate communication area estimation.

  DESCRIPTION OF SYMBOLS 1 ... Terminal, 11 ... Data measurement part, 12 ... Data transmission part, 2 ... Server, 21 ... Terminal position monitoring part, 22 ... Data acquisition instruction | indication part, 23 ... Data receiving unit, 24 ... feature space mapping unit, 25 ... initial value generation unit, 26 ... vector learning unit, 27 ... area characteristic determination unit, 28 ... area estimation unit, 29 ... Terminal location database, 30 ... logical area database, 31 ... actual value database, 32 ... categorize database, 33 ... acquisition unit, 34 ... learning unit, 35 ... estimation unit

Claims (6)

A communication area estimation system that includes a wireless terminal and a server device and estimates a communication area of the wireless terminal,
The wireless terminal is
In accordance with an instruction from the server device, while existing within a predetermined measurement range, for each area surrounded by an area boundary, reception level information, a data measurement unit that acquires its own position information,
A data transmission unit that transmits the reception level information and the position information to the server device;
The server device
A logical area database that stores area information logically divided in advance;
A measured value database for storing measured values;
A categorization database for storing information on the result of categorizing the acquired information;
Based on the position information transmitted from the wireless terminal, the position of the wireless terminal is managed, and when the wireless terminal enters a measurement range in which a communication area should be estimated, data acquisition is performed for the wireless terminal. A data acquisition instruction unit for transmitting
In response to the instruction, the reception level information transmitted from the wireless terminal is received, the logically divided area information stored in the logical area database and the reception level information are associated with each other, and the actual measurement value database A data receiving unit stored in
Using the information stored in the actual measurement value database, categorizing the reception level information according to the difference in the feature amount of the reception level information, and storing a categorization result in the categorization database;
An area characteristic determination unit that performs maximum likelihood estimation of a radio propagation model parameter assigned to each category based on information stored in the categorization database;
A communication area estimation system comprising: an area estimation unit that estimates a wireless communication area from a propagation model of each category based on the maximum likelihood estimated parameter. The learning unit
Mapping the received level information acquired on the feature space centered on the received level information resulting from the characteristics of the propagation model to be generated, and learning vector quantization using the logical area information associated with the received level information as teacher information Categorizing the received reception level information using the method, dividing the area with the similarity of the reference vector in the feature space, and storing the position information of the reference vector in the feature space in the categorization database The communication area estimation system according to claim 1, wherein reception level information is categorized according to a difference in amount. The server device in a communication area estimation system configured by a wireless terminal and a server device, and estimating a communication area of the wireless terminal,
A logical area database that stores area information logically divided in advance;
A measured value database for storing measured values;
A categorization database that stores information on the results of categorizing acquired data;
Based on the position information transmitted from the wireless terminal, the position of the wireless terminal is managed, and when the wireless terminal enters a measurement range in which a communication area should be estimated, data acquisition to the wireless terminal is performed. A data acquisition instruction unit for transmitting an instruction;
Data that receives reception level information transmitted from the wireless terminal in response to the instruction, and stores the area information logically divided into the logical area database and the reception level information in the measured value database A receiver,
Using the information stored in the actual measurement value database, categorizing the reception level information according to the difference in the feature amount of the reception level information, and storing a categorization result in the categorization database;
An area characteristic determination unit that performs maximum likelihood estimation of a radio propagation model parameter assigned to each category based on information stored in the categorization database;
A server apparatus comprising: an area estimation unit that estimates a wireless communication area from a propagation model of each category based on the maximum likelihood estimated parameter. A communication area estimation method comprising a wireless terminal and a server device, and performed by a communication area estimation system for estimating a communication area of the wireless terminal,
The server device
A logical area database that stores area information logically divided in advance;
A measured value database for storing measured values;
A categorization database that stores information on the results of categorizing acquired data;
The server device manages the position of the wireless terminal based on the position information transmitted from the wireless terminal, and when the wireless terminal enters a measurement range in which a communication area should be estimated, A data acquisition instruction step for transmitting a data acquisition instruction;
A data measurement step for acquiring reception level information for each area surrounded by an area boundary and its own position information while the wireless terminal is within a predetermined measurement range according to an instruction from the server device;
A data transmission step in which the wireless terminal transmits the reception level information and the position information to the server device;
In response to the instruction, the server device receives the reception level information transmitted from the wireless terminal and associates the logically divided area information stored in the logical area database with the reception level information. A data receiving unit for storing in the measured value database;
The server device uses the information stored in the measured value database to categorize the reception level information according to a difference in the feature quantity of the reception level information, and stores a categorization result in the categorization database;
An area characteristic determination step in which the server device performs maximum likelihood estimation of a radio propagation model parameter assigned to each category based on information stored in the categorization database;
An area estimation step in which the server device estimates a wireless communication area from a propagation model of each category based on the maximum likelihood estimated parameter. In the learning step,
Mapping the received level information acquired on the feature space centered on the received level information resulting from the characteristics of the propagation model to be generated, and learning vector quantization using the logical area information associated with the received level information as teacher information Categorizing the received reception level information using the method, dividing the area with the similarity of the reference vector in the feature space, and storing the position information of the reference vector in the feature space in the categorization database The communication area estimation method according to claim 4, wherein the reception level information is categorized according to a difference in amount.   A communication area estimation program for causing a computer to function as the server device according to claim 3.

Images