JP7333063B2 - Sensing data compression device, sensing data restoration device, sensing data transmission device, sensing data compression program, sensing data restoration program, and sensing data transmission program - Google Patents

Description

The present invention relates to a sensing data compression device, a sensing data restoration device, a sensing data transmission device, a sensing data compression program, a sensing data restoration program, and a sensing data transmission program for detecting spatially available frequency resources.

By attaching a wireless function to a sensor that recognizes the state of an object, a technology that records sensor information on the Internet, monitors the state of an object, controls equipment, collects billing information, etc. has recently been used in Machine to Machine (M2M) Communications It is called the Internet of Things (IoT), and its application in a wide range of fields is under study. In particular, wireless communication for collecting sensor information is called a wireless sensor network (WSN).

With the spread of such IoT technology, the frequency of use of the so-called unlicensed band near the 2.4 GHz band tends to increase. The unlicensed band has no administrator and does not require acquisition of a radio station license, allowing users to freely share frequency resources. On the other hand, however, the bandwidth is very limited (around 100 MHz). Therefore, research and development are being conducted on techniques for searching for spatially available frequency resources and for effectively using frequency resources.

For example, based on the reception power of radio waves received in each frequency band of a receiving device fixed outdoors, map information that estimates the range where devices that use radio waves of each frequency band exist and displays the specified area. is displayed and used to search for available frequency resources (Patent Document 1).

In addition, in LPWA, when the number of terminals sharing a frequency increases, the frequency of packet collisions increases. Studies have also been conducted to keep the packet loss rate during communication below an allowable value (Non-Patent Document 1).

JP 2019-118078 A

Yuta Yamazaki, Takeo Fujii, Osamu Takyu, Mai Ota, Koichi Adachi, Examination of frequency sharing method for LPWA using spectrum database, IEICE Technical Report, vol. 119, no. 62, SR2019-10, pp. 63-68, May 2019

However, when data such as radio wave intensity measured by a group of sensors in an area is collected and transmitted to the base station in real time as area map information or the like, frequency resources will be consumed to some extent. Moreover, as real-time performance is demanded, in other words, as delay reduction is demanded, the frequency of transmission of area map information increases, causing a contradiction in that limited frequency resources are further squeezed. Therefore, how to reduce the amount of information in the area map is important.

A sensing data compression device according to an aspect of the present disclosure transmits spatial distribution information of physical quantities detected by a plurality of sensors installed in a predetermined area from a signal of a wireless terminal transmitted from an arbitrary position in the predetermined area. A data compression device, comprising: a unit for creating weight set information based on a base set pattern set from the spatial distribution information; and a unit for transmitting all or part of the elements of the weight set information instead of the spatial distribution information. Further, the base set pattern set is a main spatial distribution information set based on a data set obtained by detecting physical quantities of signals of existing wireless stations transmitted from a plurality of positions in the area in advance by the plurality of sensors. It is characterized in that it is obtained in advance by a unit that performs component analysis.

The unit for performing the principal component analysis includes a unit for centering the spatial distribution information set, a unit for generating a covariance matrix from the centered spatial distribution information set, and a unit for eigenvalue decomposition of the covariance matrix. may be provided.

The unit for transmitting the elements of the weight set information evaluates the absolute value of each element constituting the weight set information, selects and transmits the elements in descending order of the absolute values, and the elements to be transmitted may be in the range of the top 25% to 50% of all elements .

The physical quantity may be radio wave intensity.

A sensing data restoration device according to an aspect of the present disclosure is a sensing data restoration device that restores the spatial distribution information from the weight set information transmitted from the sensing data compression device, It is characterized by comprising a unit for weighting the base set pattern set using the weight set information.

A sensing data transmission device according to an aspect of the present disclosure may include the sensing data compression device and the sensing data restoration device.

A sensing data compression program according to one aspect of the present disclosure is a method for compressing spatial distribution information based on data detected by a plurality of sensors installed in a predetermined area from a signal of a wireless terminal transmitted from an arbitrary position in the predetermined area. A sensing data compression program for generating information, comprising: creating weight set information based on a base set pattern set from the spatial distribution information; and replacing all or part of the weight set information with the spatial distribution information. Further, the base set pattern set is a space based on a data set obtained by the plurality of sensors detecting physical quantities of signals of existing wireless stations transmitted in advance from a plurality of locations in the area. It is characterized in that it is obtained in advance by the step of executing principal component analysis of the distribution information set.

The program for executing the principal component analysis includes steps of centering the spatial distribution information set, generating a covariance matrix from the centered spatial distribution information set, and performing eigenvalue decomposition of the covariance matrix. may include

The step of transmitting the elements of the weight set information evaluates the absolute value of each element constituting the weight set information, selects and transmits the elements in descending order of the absolute values, and the elements to be transmitted may be in the range of the top 25% to 50% of all elements .

The physical quantity may be radio wave intensity.

A sensing data restoration program according to an aspect of the present disclosure is a sensing data restoration program for restoring the spatial distribution information from the weight set information generated by the sensing data compression program described above, wherein the area held in advance using the weight set information.

A sensing data transmission program according to an aspect of the present disclosure includes the data compression program and the sensing data restoration program.

According to one aspect of the present disclosure, area map information (spatial distribution information) in an arbitrary area is subjected to eigenvalue decomposition using a base set pattern set that reflects the terrain, buildings, and climate peculiar to the area, and all of the eigenvalues Alternatively, by transmitting part of it, the frequency resources required for transmitting area map information can be greatly reduced. As a result, real-time transmission of area map information that changes from moment to moment becomes possible.

1 is an overall conceptual diagram (transmission of sensing information) of an embodiment of the present disclosure; FIG. 4 is a flowchart (generation and transmission of weight set information w) according to an embodiment of the present disclosure; FIG. 4 is a conceptual diagram (generation and transmission of base set pattern U) of an embodiment of the present disclosure; 4 is a flowchart (generation and transmission of base set pattern U) according to an embodiment of the present disclosure; 4 is a flowchart (reconstruction of sensing information) according to an embodiment of the present disclosure; FIG. 3 is a building layout diagram for simulation in an embodiment of the present disclosure; FIG. 3 is a layout diagram of a simulated transmission and reception point in an embodiment of the present disclosure; FIG. 2 is a conceptual diagram of simulation in one embodiment of the present disclosure; 4 is a simulation result (radio field intensity distribution) in an embodiment of the present disclosure; FIG. 10 is a simulation result (narrowing down of weight set elements) according to an embodiment of the present disclosure; FIG.

An embodiment (hereinafter referred to as the present embodiment) according to one aspect of the present disclosure will be described below. FIG. 1 shows a conceptual diagram of this embodiment. In FIG. 1, it is assumed that a wireless terminal 4 (such as a smart phone) occupies a predetermined frequency as an existing wireless station at an arbitrary position in a predetermined area. A plurality of sensors 5a to 5f are installed in the same area. Each of the sensors 5a to 5f sends the strength of the detected signal transmitted from the wireless terminal 4 to the relay station 3 as data.

(sensing data compression)
The sensing information compression device 1 arranges the data of the sensors 5a to 5f collected by the relay station 3 to create spatial distribution information (vector) x of radio wave intensity, and further creates weight set information (vector) w from this spatial distribution information. . The sensing information compression device 1 may be a part of the function of the relay station 3, or may be connected to the relay station 3 via an interface. In this embodiment, the sensing information compression device 1 further has a processor 100, and the processor 100 executes a program represented by the flowchart of FIG. A detailed description will be given below.

First, data is received from each of the sensors 5a to 5f via the relay station 3 (S101). Next, these data are arranged to generate spatial distribution information x (S102). Note that the spatial distribution information x is represented by a vector (N×1 matrix) such as (x ₁ , x ₂ , . . . x _N ) ^T . Here, N is the total number of signals from the sensors 5a to 5f received by the relay station 3. N=70 in the examples described later. Next, weight set information w is created from this spatial distribution information x (S103). The weight set information w is transmitted to the base station or the like that performs channel allocation via the relay station 3 (S104).

The calculation of the weight set information w is performed using, for example, the following formula (1).


Here, A is a spatial distribution information set based on a data set obtained by sensors 5a to 5f (N pieces) detecting radio waves from existing wireless stations 40 transmitted from a plurality of (K locations) positions in the area in advance. and U represents a base set pattern set obtained by subjecting the spatial distribution information set A to principal component analysis. The spatial distribution information set A and the base set pattern set U will be described below with reference to FIGS. 3 and 4. FIG.

(Generation of Spatial Distribution Information Set A and Base Set Pattern Set U)
In FIG. 3, reference numeral 40 denotes an existing radio station for radio wave environment measurement, which transmits test radio waves of a predetermined frequency at a plurality of (K) transmission points within an area. The existing radio station 40 may circulate through K transmission points, or may be distributed in advance at K points and sequentially transmit test radio waves. K=70 in the embodiment described later. In the present embodiment, both the spatial distribution information set A and the base set pattern set U are generated by the processor 100 according to the program shown in the flowchart of FIG.

The sensing information compression device 1 receives the data of the sensors 5a to 5f via the relay station 3 for each transmission point (S140), arranges the data collections of the sensors 5a to 5f for each transmission point, and obtains the spatial distribution information set A is generated (S150). The spatial distribution information set A is represented by an N×K matrix ((a ₁₁ ,a ₁₂ ,...a _1K ),(a ₂₁ ,...),...,(a _N1 ,...a _NK )) ^T be able to. Here, N is the total number of signals from the sensors 5a to 5f received by the relay station 3, and K is the number of points from which the existing radio station 40 has transmitted test radio waves.

The sensing information compression device 1 then performs principal component analysis on the spatial distribution information set A (S160) and extracts a base set pattern set U. FIG. The principal component analysis includes three steps of centering the spatial distribution information set (matrix) A (S161), generating a covariance matrix (S162), and eigenvalue decomposition of the covariance matrix (S163). First, centering is performed using equation (2) below.


Here, mean ( ) is a vector (Σ _{k = 1, Ka 1k} , Σ _{k = 1, Ka 2k} , . . . , Σ _{k = 1, K} a _Nk )/K, and Equation (2) calculates the average value (Σ _{k = 1, K} a _{nk /} K) is subtracted. Next, a covariance matrix C is generated using equation (3) below.


Furthermore, this covariance matrix C can be decomposed as shown in the following equation (4) using a diagonal matrix D in which eigenvalues are arranged diagonally and a matrix V composed of eigenvector groups.

Next, a base set pattern set U is generated from the diagonal matrix D and the eigenvector matrix V thus obtained (S170). For example, a base set pattern set U is defined as in Equation (5) below.


Then, the centered spatial distribution information set A can be transformed as shown in Equation (6).

Furthermore, by defining the weight set as in the following equation (7),


The spatial distribution information set A can be expressed as in Equation (8) below.

The base set pattern set U obtained as described above is appropriately transmitted to the database 20 provided in the base station or the like (S180). You may transmit mean (A) collectively. At this time, real-time property is unnecessary. In other words, the base set pattern set U is determined by the topography or buildings of the area to which the sensors 5a to 5f belong, or by the weather. is considered sufficient. Alternatively, the database may form the pattern set U using sensing data previously collected from wireless terminals. The weight set W (matrix) obtained by the step of generating the base set pattern set U using the existing wireless station 40 does not have to be transmitted.

(Sensing data restoration)
When the sensors 5a to 5f detect radio waves emitted by the wireless terminal 4 at an arbitrary position within the area, the weight set information (vector) w (=(w ₁ , w ₂ ,· • Only w _N ) ^T ) needs to be transmitted to the base station in real time. A sensing data restoration device 2 is provided on the base station side, and the processor 200 in the sensing data restoration device 2 executes the program represented by the flowchart in _FIG . Restore x ₂ , . . . x _N ) ^T ).

First, the weight set information w is received from the relay station 3 (S201), then the corresponding base set pattern set U is read out from the database 20 (S202), and finally the weighting operation is executed (S203). A weighting operation is performed, for example, as shown in Equation (9) below.

Note that the weight set information w is _, in other words, the basis set pattern vectors (u ₁ , u ₂ , . This information indicates whether or not the spatial distribution information by the terminal 4 can be restored. Therefore, among the N elements, there are elements that are zero or close to zero. You may

For example, the absolute value of each element of the weight set information w may be evaluated and transmitted in descending order of the absolute value. Even if it is narrowed down and transmitted, the error rate is within a little over 10%. Increasing the number of transmitted elements out of N elements reduces the error rate, but due to the trade-off between the effect of reducing transmitted information and the effect of improving the error rate, the upper limit of the number of transmitted elements is about the top 50%. It is preferable to keep

However, when the elements of the weight set w are rearranged, the rearrangement information must also be notified to the base station side, resulting in some overhead. For example, when sending some elements of a set of w, it is necessary to include in the header or the like information about which element of w was sent. However, the increase is only a few bytes at most, which is considered to be sufficiently small compared to the overall reduction amount.

The present embodiment has been described above. In this embodiment, compression and decompression of sensing data are performed by programs incorporated in processor 100 and processor 200, respectively. may be

Examples of the present disclosure will be described below. In this embodiment, a ray tracing simulation using an urban area model (FIG. 6) and its results are shown. Table 1 shows the simulation conditions.

There are 70 existing radio stations (FPU) and 70 sensors. For reference, the transmission positions of the existing wireless stations for testing and the layout of the sensors are shown in FIGS. 7(a) and 7(b), respectively. In FIG. 7(a), there are 71 transmission positions of existing radio stations for testing. It is a place.

FIG. 8 shows the distribution of radio field intensity when 70 sensors detect test radio waves transmitted from three locations while the existing radio station 40 changes its position in sequence, and a spatial distribution information set A is generated by collecting these. A state is shown conceptually. As described above, the spatial distribution information set A forms a matrix of N×K=70×70, and the outputs of 70 sensors at arbitrary positions of existing radio stations are arranged on one column. FIG. 9 shows an example of calculation of the radio wave intensity in the same city model using ray tracing when the wireless terminal 4 uses a frequency of 2350 MHz.

FIG. 10 shows a graph in which the recovery rate of the spatial distribution information x is calculated when the elements of the weight set information (vector) w are narrowed down from the one with the largest absolute value. The horizontal axis is the error rate, and the vertical axis is the cumulative distribution (CDF). Numerical values (displayed as percentages) in the graph indicate to what percentage of the elements of w the absolute values were narrowed down (whether other elements were cut off). For example, even when up to 33% of w are transmitted in descending order of absolute value (when the remainder is cut off), all spatial distribution information can be restored with an error rate of 10% or less.

The invention can be used, for example, to search for free frequency resources in unlicensed bands. Also, in the license band, it can be used when frequency sharing is performed.

1 sensing information compression device 2 sensing information restoration device 20 database 3 relay station 4 wireless terminal 40 existing wireless stations 5a to 5f sensor 100 processor 200 processor

Claims (12)

A sensing data compression device that transmits spatial distribution information of physical quantities detected by a plurality of sensors installed in a predetermined area from a wireless terminal signal transmitted from an arbitrary position in the predetermined area,
a unit for creating weight set information based on a set of basis set patterns from the spatial distribution information;
a unit that transmits all or part of the weight set information instead of the spatial distribution information;
Further, the base set pattern set is obtained by performing principal component analysis of a spatial distribution information set based on a data set obtained by detecting physical quantities of signals of existing wireless stations transmitted from a plurality of positions in the area in advance by the plurality of sensors. A sensing data compression device characterized in that the data is obtained in advance by a unit that executes the data. The unit for performing the principal component analysis includes a unit for centering the spatial distribution information set, a unit for generating a covariance matrix from the centered spatial distribution information set, and a unit for eigenvalue decomposition of the covariance matrix. The sensing data compression device according to claim 1, characterized by comprising: The unit for transmitting the elements of the weight set information evaluates the absolute value of each element constituting the weight set information, selects and transmits the elements in descending order of the absolute values, and the elements to be transmitted 3. The sensing data compression device according to claim 1, wherein the upper limit of the number of is in the range of the upper 25% to 50% of all elements . 4. The sensing data compression device according to claim 1, wherein said physical quantity is radio wave intensity. A sensing data restoration device for restoring the spatial distribution information from the weight set information transmitted from the sensing data compression device according to any one of claims 1 to 4,
A sensing data restoration apparatus, comprising: a unit for weighting a prestored base set pattern set related to the area using the weight set information. A sensing data transmission device comprising the sensing data compression device according to any one of claims 1 to 4 and the sensing data restoration device according to claim 5. A sensing data compression program for generating compressed information related to spatial distribution information based on data detected by a plurality of sensors installed in a predetermined area from a wireless terminal signal transmitted from an arbitrary position in the predetermined area,
creating weight set information based on a base set pattern set from the spatial distribution information;
transmitting all or some elements of the weight set information in place of the spatial distribution information;
Further, the base set pattern set is obtained by performing principal component analysis of a spatial distribution information set based on a data set obtained by detecting physical quantities of signals of existing wireless stations transmitted from a plurality of positions in the area in advance by the plurality of sensors. A sensing data compression program, characterized in that it is acquired in advance by the step to be executed. The program for executing the principal component analysis includes steps of centering the spatial distribution information set, generating a covariance matrix from the centered spatial distribution information set, and performing eigenvalue decomposition of the covariance matrix. 8. The sensing data compression program according to claim 7, comprising: The step of transmitting the elements of the weight set information evaluates the absolute value of each element constituting the weight set information, selects and transmits the elements in descending order of the absolute values, and the elements to be transmitted 9. The sensing data compression program according to claim 7, wherein the upper limit of the number of is in the range of the upper 25% to 50% of all elements . 10. The sensing data compression program according to any one of claims 7 to 9, wherein said physical quantity is radio wave intensity. A sensing data restoration program for restoring the spatial distribution information from the weight set information generated by the sensing data compression program according to any one of claims 7 to 10,
A sensing data restoration program, comprising: performing a weighting operation on a prestored base set pattern set related to the area using the weight set information. A sensing data transmission program comprising the sensing data compression program according to any one of claims 7 to 10 or the sensing data restoration program according to claim 11.