JP5816748B2 - Moving means discriminating system, moving means discriminating apparatus, and moving means discriminating program - Google Patents

Description

  The present invention relates to a system for determining a moving means of a terminal.

  In recent years, with the spread of smartphones, tablet terminals, etc., using a sensor mounted on a mobile terminal, a method for estimating the means of movement of a person holding the terminal, and an attempt to provide a service using the result Has been made.

  For example, Patent Document 1 describes a mobile terminal device that provides support according to a moving means by switching an application program corresponding to the moving means even when a plurality of moving means are used. Specifically, using an azimuth sensor, temperature sensor, barometric sensor, tilt sensor, gyro sensor, GPS (Global Positioning System) signal receiver, map database, etc., a vehicle, walking, bicycle, motorcycle, train, etc. A method for discrimination is described.

  Also, Non-Patent Document 1 uses a machine learning technique, and based on the movement of a mobile terminal equipped with an acceleration sensor, the movement status of the terminal holder is “stationary”, “walking”, “running”, A method for estimating which of the four states of “bus / train” is described.

JP 2007-303989 A

Naoki Ikeya, Satoshi Kikuchi, Kenta Nagata, Masanori Hattori, “A movement estimation method using a three-axis acceleration sensor”, IEICE technical report. USN Ubiquitous Sensor Network, Vol.108, Num.138, pp.75-80, July 10, 2008

  However, neither Patent Document 1 nor Non-Patent Document 1 takes into consideration the factors of data variation in each moving means. Values output by acceleration sensors, gyro sensors, bearing sensors, barometric sensors, tilt sensors, etc. are strongly affected by differences in road conditions, for example, so it is important to consider this in order to perform accurate estimation It is.

  For example, on a well-maintained road, the moving body (for example, a vehicle or a person) has small up / down / left / right / front / back shaking, and thus the shaking transmitted to the terminal is also small. The variation is considered to be small. On the other hand, for example, on a gravel road or a road that is unevenly maintained even on asphalt, the moving body is greatly shaken up and down, left and right, and back and forth, so that the vibration transmitted to the terminal is also large. For this reason, it is considered that there is a large variation in the measurement values of various sensors due to this shaking, and it is necessary to set a criterion based on the road conditions. In particular, for vehicles, motorcycles, buses, and the like that are considered to be easily affected by road conditions, a discrimination method considering the road conditions is required.

  Further, for example, a driver's driving habit or the like can affect the sensor data. A discriminating method considering this factor is also important.

  The present invention has been made in view of such circumstances, and for example, a method for determining a moving means of a terminal in consideration of factors that affect sensor values such as road conditions, and a system for realizing the method The purpose is to provide.

  Of the inventions disclosed in this application, the outline of typical ones will be briefly described as follows.

  A first sensor, a granting unit for giving environmental information about the environment from which the data is acquired to data acquired by the first sensor, and a terminal on which the first sensor is mounted in association with each of the plurality of environmental information A storage unit that stores a discrimination reference value for discriminating the moving means, and a predetermined discrimination reference value is selected from the storage unit according to the given environment information, and the terminal and the terminal using the data and the predetermined discrimination reference value A moving means discriminating system having a moving means discriminating unit for discriminating moving means.

  According to the present invention, it is possible to accurately determine the moving means of the terminal.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is an example of the block diagram of a moving means discrimination | determination system. It is an example of the block diagram of the portable terminal for performing a moving means discrimination | determination. It is an example of the block diagram of the computer for performing a moving means discrimination | determination. It is an example of the flowchart explaining the process of a movement means discrimination | determination. It is an example of the collection data to which the label was given. It is an example of the data stored in the discrimination | determination reference | standard database classified by factor. It is an example of the data stored in the discrimination | determination reference | standard database classified by factor. It is an example of a block diagram of the moving means discrimination | determination system with a learning function. It is an example of the data stored in the database for determination criterion determination. It is an example of the system block diagram at the time of collecting learning data via a network. It is an example of the interface for data upload. It is an example of a block diagram of the movement means discrimination | determination system with a correction function. 10 is an example of a flowchart illustrating a procedure for performing moving means determination using a time-series determination information storage unit 1008. It is an example of the data stored in a time series discrimination information storage part. It is an example of a block diagram of the moving means discrimination | determination system with a stationary travel determination function. It is an example of the flowchart explaining the method of performing 2-class classification by k-means method. It is an example of the data stored in the discrimination | determination reference | standard database classified by factor. It is an example of the data stored in the database for determination criterion determination. It is an example of the flowchart explaining the procedure which performs stationary driving | running | working determination from GPS information. It is an example of a block diagram of the moving means discrimination | determination system with a walk exclusion function. It is an example of the acceleration data at the time of walking. It is an example of the acceleration data at the time of walking. It is an example of the acceleration data at the time of walking. It is an example of the acceleration data at the time of walking. It is an example of the flowchart explaining the process of the walk detection part 1710. It is an example of a block diagram of the moving means discrimination | determination system with a characteristic moving means exclusion function. 10 is an example of a flowchart for explaining processing of a characteristic moving means detection unit 2011. It is an example of the graph showing the characteristic of the speed at the time of train travel. It is an example of a block diagram of the movement means discrimination | determination system with a multiple sensor integrated function. It is an example of the data stored in the discrimination | determination reference | standard database classified by factor. It is an example of the data stored in the database for determination criterion determination. It is an example of a block diagram of the moving means discrimination | determination system with surrounding terminal information utilization function. It is an example of a movement means discrimination | determination result database. It is an example of a block diagram of the moving means discrimination | determination system with a road condition determination function. It is an example of the discrimination | determination reference | standard database classified by factor. It is an example of the database for determination criteria determination. It is an example of a block diagram of the moving means discrimination | determination system with an unlabeled data learning function. It is an example of the flowchart explaining the process which correct | amends a discrimination | determination reference value using unlabeled data. It is an example of the data stored in the database for non-label discrimination criterion correction. It is an example of the discrimination | determination result of unlabeled data. It is an example of correction | amendment of the discrimination | determination reference | standard database classified by factor. It is an example of a block diagram of the moving means discrimination | determination system with a data transmission control function. It is an example of a block diagram of the moving means discrimination | determination system with a sensor On / Off switching function. It is an example of a block diagram of a road congestion condition estimation system. It is an example of the data stored in the database for road congestion condition estimation. It is an example of a block diagram of a map utilization road congestion condition estimation system. It is an example of the block diagram of the road congestion condition estimation system with the congestion condition display function according to a moving means. It is an example of a display of the road congestion condition according to a moving means.

  Hereinafter, examples will be described with reference to the drawings.

  In the present embodiment, an example of a system 100 (hereinafter referred to as a “moving means discriminating system”) 100 that determines the moving means of a terminal holder using an acceleration sensor will be described.

  In the following, a method for discriminating between cars and motorcycles will be described as an example of a moving means. However, auto-rickshaws that are popular in buses, trains, bicycles, emerging countries, etc. can be discriminated by the same method. is there.

  Further, as factors that affect the sensor, environmental information related to the environment from which the data was acquired, specifically, information related to road conditions such as road attributes and regions is used as an example. In addition, for example, driver-specific information such as driving habits and other factors can be handled in the same way as environmental information, and it is also possible to switch the discrimination criteria for multiple factors among them It is.

  FIG. 1 is an example of a configuration diagram of a moving means determination system of the present embodiment. The moving means discriminating system 100 includes an acceleration sensor 101, a moving means discriminating unit 102, a factor-specific discriminant reference database 103, and a factor label attaching unit 104.

  The acceleration sensor 101 measures data at a predetermined sampling rate. The factor label assigning unit 104 assigns a label relating to the road condition to the measurement value obtained from the acceleration sensor 101. Based on the label, the moving means discriminating unit 102 reads out the discriminant reference value for each road situation stored in the factor-specific discriminant reference database 103, and compares the discriminant reference value with the collected data to move. Determine the means. The determination method will be described later.

  Here, the moving means determination system 100 can be realized by a single terminal having an acceleration sensor such as a smartphone, a calculation unit, and a storage unit. In addition to a terminal such as a smartphone having the acceleration sensor 100, a computer that performs arithmetic processing may be prepared, and the moving means determination unit 102 and the factor-specific determination reference database 103 may be included in the computer. In addition, what is necessary is just to implement | achieve the factor label provision part 104 with a terminal or a computer.

  When the moving means determination system 100 is realized by a single terminal, for example, a portable terminal 200 as shown in FIG. 2 can be used. In the portable terminal 200, the central processing unit 203 compares the measurement value obtained from the acceleration sensor 202 with the discrimination reference value for each road situation stored in the discrimination criteria database classified by factor stored in the storage device 204. Then, it is determined whether the moving means is a car or a motorcycle. Data is transmitted / received using the bus 206, for example. In addition, it is desirable to include an input control device 201 that accepts user input and a screen display device 205 for displaying determination results and the like.

  When the functions described in the embodiments are realized by a mobile terminal, the central processing unit 203 reads and executes various programs recorded in the storage device 204 to realize various functions. For example, the processing performed by the moving means determination unit 102 is realized by the central processing unit 203 reading and executing a moving means determination program recorded in the storage device 204. The same applies to other processes.

  In the case where a terminal having the acceleration sensor 100 and a computer having the moving means discriminating unit 101 and the factor-based discriminant reference database 103 are prepared separately, for example, a computer as shown in FIG. 3 can be used. In this case, the measurement value obtained from the acceleration sensor 100 may be transmitted to the computer 300 by connecting the acceleration sensor 100 and the computer 300 using a USB (Universal Serial Bus) cable, for example, or via a network. Alternatively, the data may be written once on a medium such as a CD or a DVD and then read by the computer 300. The measured values of the acceleration sensor 100 thus obtained in some way are read by the central processing unit 302 to the main storage device 303 and stored in the factor-specific discrimination reference database 103 held in the auxiliary storage device 304. By comparing with the determination reference value for each situation, it is determined whether the moving means is a car or a motorcycle. Data is transmitted and received using the bus 306, for example. Here, as the main memory 303, for example, a dynamic random access memory (DRAM) or a static random access memory (SRAM) can be used. As the auxiliary storage device 304, for example, a hard disk, a flash memory, a flexible disk, or the like can be used. In addition, in order to receive a user input, for example, the input control device 301 that processes input from the input device 310 such as a mouse or a keyboard, the output device 320 such as a display for displaying a discrimination result, and the output are controlled. It is desirable to include an output control device 305 or the like.

  Further, when the functions described in the embodiments are realized by a computer, the central processing unit 302 realizes various functions by reading and executing various programs recorded in the auxiliary storage device 304. For example, the processing performed by the moving means determination unit 102 is realized by the central processing unit 302 reading and executing a moving means determination program recorded in the auxiliary storage device 304. The same applies to other processes.

  Next, an example of the feature amount used for the moving unit determination process will be described. Here, a method for calculating a variance value of the norm of acceleration obtained from a triaxial acceleration sensor at regular intervals and using the median when the values are collected for a certain amount of time will be described as an example. .

Here, the norm of acceleration _satisfies a = (a _x ² + a _y ² + a _z ² ) ^1/2 when the measured values of triaxial acceleration are a _x , a _y , and a _z , respectively. It is a value. The reason for using this is that the acceleration measured in each axis direction depends greatly on the orientation of the terminal, so when you hold it in various ways, such as putting it in a pants pocket, putting it in a breast pocket, or putting it in a bag. , Because it is not possible to obtain a stable value with only the acceleration in each axis direction, but using the norm can handle the magnitude of acceleration regardless of the orientation, so it is considered that a stable value independent of the terminal posture can be obtained. It is.

  The reason for using the variance value is that the variance value is considered to be a value that well reflects the vibration inherent in the moving means. For example, simply using the absolute value of acceleration to detect the difference between acceleration and deceleration at start and stop, the moving means cannot be determined when driving at a constant speed, but when it is running at a constant speed. By paying attention to the difference in characteristics, it is possible to determine the moving means regardless of whether it is stationary, accelerated / decelerated, or traveling at a constant speed.

  Here, the reason why the vibration inherent in the moving means occurs will be described. For example, cars and motorcycles are equipped with a drive system such as an engine, and vibrations due to the drive system occur when the vehicle is operated. For example, in the case of a car, the drive system is often installed in the hood at the front of the vehicle. In some cases, the drive system is often provided just below the seat, and the vibration tends to be transmitted to the passenger. Or, the car is better than the bike because the suspension function that absorbs the vibration caused by the vertical movement of the vehicle, etc., and prevents it from being transmitted to the rider is better than the bike. However, there is also a tendency that vibration caused by road surface unevenness or the like is not easily transmitted to the passenger. In this way, there is a certain tendency in how vibration is transmitted depending on the moving means, which is the reason why the vibration inherent in the moving means is generated.

  The reason for using the median is to improve the stability of discrimination. For example, when a determination is made based on only one variance value in a short time segment such as 10 seconds (this is called a small segment), acceleration / deceleration was frequently repeated in that segment, especially on a road with poor pavement conditions. An erroneous determination may be made due to various reasons such as accidental large noise in the sensor. On the other hand, for example, 900 seconds, that is, 90 small segments, by using the median value when the dispersion values are collected for a certain period of time (this is called a large segment), the above noise can be prevented. Robust discrimination can be made.

  Hereinafter, a specific procedure of the moving means determination process will be described with reference to FIG.

  First, at 401, the measurement value of the acceleration sensor is read. Next, in 402, the factor label assigning unit 104 assigns a label for each factor that affects the measurement value of the sensor. For example, in the present embodiment, a label relating to road conditions is given. Here, the labels may be given to all the measured values at the same interval as the sensor sampling rate. Alternatively, in order to reduce the amount of data, it is possible to designate a start point and an end point, and assume that all the data between them is the same label, and attach only one label to each set of start point and end point.

  As a method for giving a label, it is possible to give information on road conditions such as a region and road attributes using GPS information. A specific method will be described later. In addition, for example, to make a note when collecting the discrimination target data, for example, in order to determine which discrimination criterion value should be read from the discrimination criteria database such as road attribute and area in the discrimination target data You may leave a record and give it manually based on it. FIG. 5A shows an example of data with labels.

Next, at 403, the data is divided into small segments of appropriate length, such as 10 seconds. Next, at 404, a norm variance σ is calculated for each small segment. Next, at 405, the median value medσ _j when the dispersion value σ _i is collected for a fixed time (large segment) such as 900 seconds, that is, 90 segments is calculated. Next, at 406, the road condition determination reference value corresponding to the label assigned at 402 is read from the factor-specific determination reference database 103 in which a threshold value θ _k set in advance for each road condition is stored. For example, a factor label as shown in FIG. 5B and a threshold value for each road attribute (asphalt pavement, asphalt pavement (many unevenness), gravel road,...) Are stored in the judgment standard database. ing. For the discrimination reference value, not only for the threshold value, but also for the discrimination function such as a conversion formula for discriminating by positive / negative of the result of the conversion, a likelihood function for outputting a certain moving object likeness from the input value, etc. Anything may be stored as long as it is used for the above. In addition, roads are well developed in the urban centers of developed countries, but roads are not paved in emerging countries, and even if there are many irregularities, the average road maintenance situation tends to vary by region. It is believed that there is. Therefore, the design may be such that the discrimination reference value is determined for each region as shown in FIG. In any case, it is important to design the database so that the discrimination reference value matches the road condition. Thereby, it is possible to select a discrimination reference value for discriminating between a car and a motorcycle according to the road condition, and the discrimination accuracy can be improved.

After reading the discrimination reference value, medσ _j and θ _k are compared in 407. If medσ _j is smaller than the discrimination reference value θ _k , the car label is output at 408, and if medσ _j is greater than the discrimination reference value θ _k , the motorcycle label is output at 409. This is because the above-described inherent vibration is considered to be larger in the motorcycle. This completes the process for one section of the large segment. Actually, this is repeated for the number of large segments, and for each large segment, it is determined whether the moving means is a car or a motorcycle.

  Note that the label application timing does not necessarily have to be immediately after data reading as described in the present embodiment. For example, the label may be applied to each large segment after being divided into large segments. Further, when the terminal and the computer are prepared separately as described above, the terminal can acquire the acceleration data and give it using the GPS information before transmitting it to the computer. In any case, as long as it is possible to determine which determination reference value should be read when determining the moving means of the large segment, the timing of labeling is not limited.

  In this embodiment, an example is described in which an acceleration sensor is used as a sensor for detecting vibrations inherent in the moving means. However, any gyro sensor, magnetic orientation sensor, or the like can be used as long as the sensor can detect the inherent vibrations. You may implement using another sensor and you may implement using these two or more.

  In addition, in this example, in order to capture the inherent vibration, the variance value of the norm was calculated at regular time intervals, and an example using the median value when the values were collected for a fixed time period was introduced. Any other index value may be used as long as it can represent a specific vibration, and the average value or quartile may be used instead of the median value.

  In this embodiment, a description will be given of an example of a system that automatically determines a discrimination reference value for discrimination of moving means from discrimination criteria determination data collected in advance (hereinafter referred to as learning data).

  In order to perform accurate movement means determination, it is necessary to appropriately determine the determination reference value stored in the determination reference database 103 described in the first embodiment. However, there is a problem that it is not always obvious what discrimination reference value should be used.

  On the other hand, an appropriate discrimination reference value based on actual data can be determined by collecting data on cars and motorcycles for each factor that affects the sensor, such as road conditions, and constructing a database in advance. By using the system described in the present embodiment, an appropriate discrimination reference value can be determined from learning data collected in advance, and the discrimination accuracy can be improved.

  FIG. 6 is an example of a configuration diagram illustrating a moving function determination system 600 with a learning function according to the second embodiment.

  In addition to the configuration of the moving means discriminating system 100 described in the first embodiment, the moving means discriminating system 600 with a learning function stores a sensor 606 for collecting learning data and the collected learning data. And a discrimination criterion determining unit 604 that determines a discrimination criterion for each factor affecting the sensor from the data. As the hardware configuration, for example, as shown in FIG. 6, the moving means discriminating unit 102, the factor-specific judgment criterion database 103, the factor label assigning unit 104, the discrimination criterion determining unit 604, and the discrimination criterion determining database 605 are realized on a computer. To do. The sensor 606 uses the same type of sensor as the sensor 101 used for moving means discrimination. As the sensor 606 for collecting learning data, the sensor 101 mounted on the terminal where the moving means is to be determined may be used, the acceleration sensor of another terminal may be used, or both of them may be used. A plurality of sensors may be used. In this embodiment, an example using an acceleration sensor will be described as an example.

  In addition to this configuration, the entire configuration shown in FIG. 6 may be realized on a single terminal using, for example, a smartphone having both sensor and computer functions, or the acceleration sensor 101 and the moving means determination unit. 102, the factor-specific discrimination standard database 103 and the factor label assigning unit 104 are realized by a terminal such as a single smartphone, the discrimination standard determination unit 604, the discrimination standard determination database 605 is a computer, and the acceleration sensor 605 is another smart phone. It may be realized by a terminal. In any case, as long as the function shown in FIG. 6 can be realized, the way of combining hardware is not limited. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted.

  FIG. 7 is an example of data stored in the discrimination criterion determination database 605. When constructing a database for determining discrimination criteria, in addition to manually labeling the collected data as car data or motorcycle data, for example, taking notes when collecting learning data, etc. In some form, a label is provided to indicate what kind of road conditions, such as road attributes and regions, the learning data is. Later, a method for automatically assigning a road condition label by adding another configuration will be described.

  In addition, the median of the variance values described when the processing procedure of FIG. 4 is described in the first embodiment is stored in the determination criterion determination database. Since the method of calculating the median of the variance values is the same as that described in the first embodiment, the description thereof is omitted here. Similar to the description in the first embodiment, the norm dispersion value is calculated every fixed time, and in addition to the median value when the values are collected for a fixed time, specific vibrations such as standard deviation and amplitude are calculated. Other index values may be used as long as they can be expressed, or an average value or a quartile may be used instead of the median value. In addition, as with the factor-specific discrimination reference database 103, the road status information can be labeled with, for example, road attributes and regions, but in this embodiment, only examples using regions are used. It is described.

  A method for determining the discrimination reference value will be described below. First, learning data is collected using the acceleration sensor 606 and stored in the discrimination criterion determination database 605. Then, the discrimination criterion determination unit 604 calculates the discrimination criterion value for each road condition with reference to the learning data stored in the discrimination criterion determination database 605 and stores the value in the factor-specific discrimination criterion database 103. . For example, SVM (Support Vector Machine) can be used for calculation of the discrimination reference value. SVM is a technique for estimating a hyperplane that best separates labeled learning data by solving a convex optimization problem.

Hereinafter, a method for applying SVM to this embodiment will be described. First, the hyperplane that separates car data (hereinafter referred to as car class data) and bike data (hereinafter referred to as bike class data) (in fact, the median of variances handled here is a one-dimensional value) Therefore, an equation of a straight line) is set as y (σ _n ) = w ₁ σ _n + w ₀ . Here, σ _n is the median value of the variance of the nth data. The purpose is to obtain the optimum values of the parameters w ₁ and w ₀ in this equation. Also, a value t _n is introduced such that +1 when the nth data is a car class and -1 when it is a motorcycle class. Here, normally, the car class data and the bike class data are rarely completely separated, and are considered to have a certain overlap. Therefore, in order to deal with such problems, we introduce a concept called soft margin, which has the effect of relaxing the constraints in SVM. Therefore, a slack variable ξ _n is introduced. ξ _n is a value defined as ξ _n = 0 when the data is correctly classified, and ξ _n = | t _n -y (σ _n ) | otherwise. Based on the above settings, based on the SVM theory, the parameters w ₀ and w ₁ of the hyperplane equation that separates the car class and the motorcycle class are the constrained optimization problems given by (Equation 1). It is obtained by solving for w ₀ and w ₁ . Here, since the optimization problem of (Equation 1) is an optimization problem having a quadratic objective function and linear constraints, a solution can be obtained without falling into a local optimization problem. Therefore, it may be solved using any existing algorithm such as the steepest descent method or Newton method. As for the parameter C, an appropriate value may be determined while confirming the discrimination accuracy when the value of C is variously changed using learning data by using an intersection confirmation method or the like.

The values of the parameters w ₁ and w ₀ obtained in this way are stored in the factor-specific discrimination reference database 103 as they are, and y (σ _n ) = w ₁ σ _n + w ₀ is obtained at the time of discrimination. When the value of σ _n ) is positive, the vehicle class may be determined, and when the value is negative, the vehicle class may be determined. Alternatively, a threshold value that separates the positive and negative values of y (σ _n ), such as σ = −w ₀ / w ₁ , may be calculated and stored in the factor-specific discrimination reference database 103. In addition to this, as to how to calculate the discrimination reference value, for example, a linear discrimination method may be used, logistic regression may be used, or a perceptron may be used. In any case, it is only necessary to be able to configure a classifier that receives the median of the variance of cars and motorcycles for each road condition and outputs a class label. Since the determination reference value is determined in this way and the value is stored in the factor-specific determination reference database 103, the part for performing the moving means determination is the same as the procedure described in FIG. Description is omitted in the examples.

  Note that when learning data is collected, data may be collected via a network. In that case, a transmission unit 807 for transmitting data via the network as shown in FIG. 8 may be added to the configuration. For example, the transmission unit 807 may be provided on a computer separately from a terminal on which the acceleration sensor 606 is mounted, and may be realized on the computer, or may be configured on the same terminal as the acceleration sensor 606. Also good. Further, the raw data of the sensor can be transmitted as it is, and the determination of the road condition and the calculation of the median value of the variance value can be performed on the discrimination criterion determination database side. However, with such a configuration, a large amount of data is transmitted via the network, and the load on the network increases. Therefore, the calculation is performed on the learning data collection terminal side including the sensor 606. It is preferable to use a configuration in which only the calculation result is transmitted. Further, as a method of transmitting data via a network, for example, data may be sent in the form of e-mail. In addition, for example, uploading may be performed using a data uploading interface as illustrated in FIG. 9. For example, in the case of discrimination of a car / bike, the data acquisition time can be easily determined by distinguishing the car / bike. Any interface may be used as long as transmission is possible.

  In the present embodiment, an example of a system that improves the accuracy of moving means discrimination by using time-series discrimination results will be described.

  As described in Example 1, the median value in the large segment can be used to increase the stability of the discrimination, but still when the special driving behavior continues intermittently in the large segment. There is a possibility of misclassification.

  On the other hand, it is difficult to think of changing cars and motorcycles in a short time period when actually living. So, for example, if a car continues as a series of discrimination results and there is only one section that is determined to be a motorcycle, that section is erroneously a motorcycle even though it was actually in a car. It can be determined that it is determined that. Therefore, the result determined to be a motorcycle can be corrected to the car.

  By using the configuration of the present embodiment, even if the determination result includes a small number of errors due to noise factors, it becomes possible to output a stable determination result by performing correction.

  FIG. 10 is an example of a configuration diagram illustrating a moving function determination system 1000 with a correction function according to the third embodiment.

  The moving means discrimination system with correction function 1000 includes a time series discrimination information storage unit 1008 in addition to the configuration of the movement means discrimination system 100 described in the first embodiment. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As a hardware configuration, for example, as illustrated in FIG. 10, a moving unit determination unit 102, a factor-specific determination reference database 103, a factor label assignment unit 104, and a time-series determination information storage unit 1008 are realized on a computer. In addition to this configuration, the entire configuration illustrated in FIG. 10 may be realized on a single terminal using, for example, a smartphone having both sensor and computer functions. In any case, as long as the function shown in FIG. 10 can be realized, the way of combining hardware is not limited.

  FIG. 11 is an example of a flowchart illustrating a procedure in which the moving unit determination unit 102 performs moving unit determination using the time series determination information storage unit 1008. First, at 1101, moving means discrimination is performed by the method described in the first or second embodiment. Next, at 1102, the temporary discrimination result obtained at 1101 is stored in the time series discrimination information storage unit 1008. Then, at 1103, the information is used to correct the discrimination result, and the label of the car or motorcycle is output and the process ends.

  FIG. 12 is an example of data stored in the time series discrimination information storage unit 1008. By the method described in the first or second embodiment, the determination is made for each large segment, and is stored in the time series determination information storage unit 1008 together with the time information.

As a specific correction processing procedure, for example, if there are a lot of results that are determined to be cars in a total of five segments, that is, a large segment of interest and two large segments before and after the large segment, If there are many results determined to be a car and a motorcycle, it may be corrected such that the discrimination result of the segment of interest is a motorcycle. For example, when there is data as shown in FIG. 12, the determination result of t _i is corrected to the car. In addition to this, of course, the number of segments in consideration of the preceding and following information may be changed, or processing such as correction may be performed only when the same determination result continues for a certain number or more. Alternatively, as described in the second embodiment, the determination reference value is obtained by learning, and when the determination reference value is determined, the likelihood of the determination result is output. When the likelihood is equal to or less than a certain value, Only the correction process may be performed, and the correction method does not matter.

  In this embodiment, a description will be given of a system that improves the accuracy of discrimination by determining whether a vehicle is stationary or running and thereby changing the selection of a discrimination reference value to be used.

  When the vehicle is stationary and when it is running, the characteristics of vibration transmitted to the terminal of the person riding on the vehicle are different. As described in the first embodiment, the difference can be absorbed to some extent by using the variance value, but the discrimination accuracy is further improved by using different discrimination reference values at the time of stationary and running. Can do.

  In general, the vibration transmitted to the terminal is larger when traveling than when stationary. In addition, if the conditions such as the terminal, vehicle, how to hold or install the terminal, road conditions, etc. are the same, the difference between the magnitude of vibration when stationary and the magnitude of vibration when traveling is almost constant. . Therefore, if a section with a small vibration and a section with a large vibration are separated in continuous data, it can be determined that each corresponds to a stationary state and a traveling state.

  Therefore, by using the configuration described in the present embodiment, it is possible to determine the moving means with higher accuracy by determining whether the vehicle is stationary and changing the determination reference value according to the determination result.

  FIG. 13 is an example of a configuration diagram of a moving means determination system with a stationary travel determination function in the fourth embodiment. In addition to the configuration of the moving means determination system 100 described in the first embodiment, the moving means determination system 1300 with a stationary travel determination function includes a stationary travel determination unit 1309. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As a hardware configuration, for example, as illustrated in FIG. 13, the moving means determination unit 102, the factor-specific determination reference database 103, the factor label assignment unit 104, and the stationary travel determination unit 1309 are realized on a computer. In addition to this configuration, the entire configuration shown in FIG. 13 may be realized on a single terminal using, for example, a smartphone having both sensor and computer functions. In any case, as long as the function shown in FIG. 13 can be realized, the way of combining hardware is not limited.

Below, the method of performing a stationary running determination from the measured value of a sensor is demonstrated. First, by the method described in the first embodiment, continuous data is divided into several large segments, and the median σ _{i of the} variance is calculated for each large segment. A set of median variances obtained here is S _σ . Then, a value belonging to this S _sigma, for example, classified into two classes by a method unsupervised learning.

Here, an example in which two-class classification is performed by the k-means method will be described with reference to FIG. First, at 1401, label 1 or 2 is randomly assigned to each σ _i belonging to S _σ . Next, at 1402, the average values m ₁ and m ₂ are calculated for each label according to (Equation 2).

Here, L _i represents the set of data attached labels i, N _i denotes the number of data attached label i. Then, in 1403, for each sigma _i, the distance between the average value m ₁ of the Class 1 | compare, | σ _i -m ₂ | distance and the average m ₂ for class 2 | σ _i -m ₁ If the distance from m ₂ is closer, the label of σ _i is set to 2 at 1405, otherwise the label of σ _i is set to 1 at 1404. Next, in 1406, it is checked whether or not there is a label change in σ _i . If there is no label change in any data, the process is terminated. If any data has a label change, the process returns to 1402. And continue processing.

  In this way, of the two obtained classes, the class having the larger value may be determined as “running”, and the class having the smaller value may be determined as “still”. In addition, although the method of performing stationary running determination from sensor data by the k-means method has been described here, the method is not limited to the k-means method, and a method such as hierarchical clustering or a self-organizing map is used. May be used instead.

  When the stationary travel determination is completed in this way, when the factor-specific discrimination reference database 103 is constructed, the stationary travel determination is performed before data storage, and the data is stored with the label of stationary or traveling. To. Further, in the case where the determination reference value is determined by learning using the configuration as described in the second embodiment, the learning data is also described in the present embodiment when the determination reference determination database 605 is constructed. Use the same method as the above method to determine stationary travel, store data after classifying stationary and travel in addition to different means of transportation and road conditions, and when determining criteria Determines the criterion value for each of these factors.

  FIG. 15A shows an example of data held in the factor-specific discrimination reference database 103 in this embodiment. FIG. 15B shows an example of data held in the discrimination criterion determination database 605 when constructing a moving means discrimination system with a learning function as described in the second embodiment. The database described in the embodiments so far is provided with other information on stationary travel.

  Further, since the position information of the terminal can be obtained from the GPS information, the moving speed of the vehicle can be calculated by using the terminal position information, thereby making it possible to determine whether the vehicle is stationary or running. In the following, a method of performing stationary running determination from GPS information when a GPS receiver is added to the configuration will be described using FIG.

First, in 1601, GPS information is read. From the GPS information, the position information of the terminal and the time information acquisition information can be obtained. Assuming that the (latitude, longitude) in decimal notation received at times t ₁ and t ₂ is (lat ₁ , lon ₁ ) (lat ₂ , lon ₂ ), the speed between them is 1602 ( It can be calculated as shown in Equation 3). Here, r in (Expression 3) is a value representing the radius of the earth.

Further, in 1603, the average value meanv _j of the speed is calculated for each large segment. Then, in 1604, the average speed meanv _j is compared with a threshold value φ for determination of stationary travel. If meanv _j is equal to or smaller than φ, it is determined that the vehicle is stationary in 1605, and if meanv _j is larger than φ, it is determined that the vehicle travels in 1606. To do. As the threshold value φ, for example, a value such as 5 km / h can be used.

  When an acceleration sensor is used as the sensor, the speed may be calculated by integrating the acceleration, and the stationary / running determination may be performed from the speed information by the method described in the present embodiment, for example.

  In the present embodiment, an example of a system capable of detecting a walking section by detecting a characteristic peculiar to walking will be described.

  When determining the moving means, if it is attempted to determine data using various moving means mixed using only the same reference, the accuracy of the determination may be reduced.

  On the other hand, when a person walks, it usually has a constant rhythm, such as walking about two steps per second, so walking can be detected by capturing the periodic rhythm that appears on the sensor. It is thought that.

  According to the configuration of the present embodiment, the discrimination means described in the above-described embodiments is applied to the section in which walking is detected and the portion is excluded, so that the moving means for data including walking can be discriminated. It can be performed with high accuracy. For this reason, for example, when generating traffic information after making a moving means determination, by excluding walking, it is erroneously recognized as a vehicle traveling on a road such as a car or a motorcycle. Traffic information can be generated with high accuracy.

  FIG. 17 is an example of a configuration diagram illustrating a moving means determination system 1700 with a walking exclusion function in the fifth embodiment. In addition to the configuration of the moving means determination system 100 described in the first embodiment, the moving means determination system 1700 with a walking exclusion function includes a walking detection unit 1710. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As a hardware configuration, for example, as illustrated in FIG. 17, the moving unit determination unit 102, the factor-specific determination reference database 103, the factor label assignment unit 104, and the walking detection unit 1710 are realized on a computer. In addition to this configuration, the entire configuration shown in FIG. 17 may be realized on a single terminal using, for example, a smartphone having both sensor and computer functions. In any case, as long as the function shown in FIG. 17 can be realized, the way of combining hardware is not limited.

  Below, the characteristic which appears at the time of a walk is demonstrated using FIG. FIG. 18A and FIG. 18B are examples of walking data when the horizontal axis represents time and the vertical axis represents the norm value of the acceleration sensor. One scale on the horizontal axis corresponds to 2 seconds. From FIG. 18A and FIG. 18B, it can be seen that there are peaks in the sensor data about twice a second. This can be considered as a feature that appears because a person walks about two steps per second. It can be considered that the peaks of different sizes appearing alternately in FIG. 18 (a) are the difference between when the right foot is stepped on and when the left foot is stepped on. Depending on how the terminal is held, the size of the mountain may differ depending on whether the left or right foot is stepped on, or may be almost the same as shown in FIG. 18B. In any case, as a cycle, a mountain appears about twice a second.

  FIG. 18 (c) is obtained by subjecting the acceleration data of FIG. 18 (a) to Fourier transform, and FIG. 18 (d) is obtained by subjecting the acceleration data of FIG. 18 (b) to Fourier transform. The horizontal axis represents frequency (unit: Hz), and the vertical axis represents power. It can be seen from FIGS. 18C and 18D that a strong peak appears in the frequency band of about 2 Hz. This can be interpreted as a result reflecting the walking of about two steps per second as described above. That is, it is possible to detect walking by capturing this feature in the frequency domain.

  Hereinafter, based on the above characteristics, the operation of the walking detector 1710 will be described with reference to FIG.

  First, in 1901, the obtained sensor data is divided into appropriate time intervals (for example, about 10 seconds). Next, in 1902, Fourier transform is applied to the data in the section. Next, in 1903, it is determined whether a peak around 2 Hz has been detected. If it can be detected, it is determined in 1904 that this section is walking, a walking label is output, and the process is terminated. If it is not detected, it is determined in 1905 that it is not walking, and the process proceeds to the subsequent movement means determination unit 102.

  In addition, although the example which detects a walk using an acceleration sensor was described here, since the same feature appears also at the time of other sensors, such as a gyro sensor and a magnetic direction sensor, walk detection using other sensors. The same method can be used for.

  In the present embodiment, an example of a system that can detect not only the detection of walking but also other moving means based on characteristics unique to the moving means will be described.

  As described in the fifth embodiment, when determining the moving means, if the data containing various moving means is determined using only the same reference, the determination accuracy may be lowered. On the other hand, some moving means have characteristics unique to the moving means, and it is considered that the moving means can be detected by capturing the features.

  According to the configuration of the present embodiment, such a moving means is detected, and the discrimination means described in the above-described embodiments is applied to the section excluding that part, thereby accurately determining the moving means. Can be done well. As a result, for example, when generating traffic information after determining the means of transportation, by excluding trains, airplanes, boats, bicycles, etc., it will accidentally run on roads such as cars and motorcycles. It is possible to prevent erroneous recognition as a vehicle and to generate accurate traffic information.

  FIG. 20 is an example of a configuration diagram showing a moving means discrimination system 2000 with a characteristic moving means excluding function.

  The moving means discrimination system 2000 with the characteristic moving means exclusion function includes a characteristic moving means detection unit 2011 and a GPS receiver 2012 in addition to the configuration of the moving means discrimination system 100 described in the first embodiment. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As a hardware configuration, for example, as shown in FIG. 20, a moving means discriminating section 102, a factor-specific judgment reference database 103, a factor label assigning section 104, and a characteristic moving means detecting section 2011 are realized on a computer, and the acceleration sensor 101 The GPS receiver 2012 is realized by another terminal. In addition to this configuration, the entire configuration shown in FIG. 20 may be realized on one terminal by using, for example, a smartphone having both sensor and computer functions. For example, only the GPS receiver 2012 may be used. It may be realized by the terminal. In any case, as long as the function shown in FIG. 20 can be realized, the way of combining hardware is not limited.

  Hereinafter, the operation of the characteristic moving means detection unit 2011 will be described with reference to FIG. Here, trains, bicycles, airplanes, and ships are described as examples of characteristic moving means.

  First, a method for detecting an airplane considered to be relatively easy to detect will be described. One of the characteristics of an airplane when compared with other vehicles is that it has a high moving speed. For example, there are few examples of vehicles other than airplanes that exceed 500 km / h. Therefore, a speed threshold value is set, and when a speed faster than this is detected in 2101, the moving means in that section is determined to be an airplane, an airplane label is output, and the process is terminated. Since the speed can be calculated using GPS information as described in the fourth embodiment, the description thereof is omitted.

  Next, a method for detecting a ship will be described. A characteristic of a ship when compared to other vehicles is that it moves over the sea or lake. Airplanes may also move between these locations, but since plane data is excluded in 2101, it may be considered here that only ships move in these locations. Since position information is obtained from the GPS receiver 2012, when it is detected in 2102 that these places are moving, it is determined that the moving means in the section is a ship, and the processing is terminated.

Next, a method for detecting a bicycle will be described. A characteristic of a bicycle when compared with other vehicles is that a periodic rhythm appears in the data. Such a rhythm is also detected during walking, but the walking section can be detected by the method described in Example 5 separately, so if walking is detected first, there is no need to consider it. Using a method similar to the method described in the fifth embodiment, frequency conversion is performed to detect the periodic rhythm. However, unlike the case of walking, for variations in the speed of operating the bicycle is large, if a strong peak at a certain threshold value theta _f following frequency bands appeared at 2103, to determine the moving means of the section with a bicycle The process is terminated. The value of the threshold value θ _f can be determined by a method of learning from separately collected bicycle data, for example.

Next, a method for detecting a train will be described. The characteristics of a train when compared with other vehicles are that it often moves linearly and changes in speed. Unlike cars and motorcycles traveling on the road, trains rarely turn nearly 90 degrees at intersections, and are hardly stopped by signals while traveling between stations. Therefore, the movement path is likely to be linear, and the speed is likely to have characteristics as shown in FIG. In FIG. 22, t ₁ , t ₅ , t ₉ are sections stopping at the station, t ₂ , t ₆ are sections leaving the station and accelerating, and t ₃ , t ₇ are constant speeds between the stations. The running sections, t ₄ and t ₈ are sections decelerating before arrival at the station. The position and speed can be calculated by the method described in the previous examples. If such a feature can be detected in 2104, it is determined that the moving means in that section is a train, and the process is terminated.

  If no characteristic moving means is detected in any of the above steps, it is determined in 2105 that it is not a characteristic moving means, and the process proceeds to the subsequent moving means determining unit 102.

  Here, trains, bicycles, airplanes, and ships are described as examples of characteristic moving means. However, other moving means can be similarly detected by capturing characteristics unique to the moving means.

  In this embodiment, an example of a system that can improve the discrimination accuracy by using a plurality of sensors in an integrated manner will be described.

  If it is attempted to discriminate the moving means with only one sensor, noise unique to that sensor may often be added, which may reduce the discrimination accuracy. For example, in the case of an acceleration sensor, acceleration / deceleration is repeated intermittently, in the case of a gyro sensor, when traveling on a road with many curves, in the case of a magnetic direction sensor, a facility such as a power plant that emits strong electromagnetics. An example is when driving nearby. On the other hand, it is considered that it is possible to cope with this by using a plurality of sensors. However, simply using a plurality of sensors makes it impossible to construct a discriminating system that is robust against the individual problems of the sensors described above. By changing the sensor according to the situation where noise inherent to the sensor is likely to be picked up and using the plurality of sensors in an integrated manner, the influence of noise inherent in the sensor can be reduced.

  According to the configuration of the present embodiment, by using a plurality of sensors in an integrated manner, even in a situation where misidentification occurs due to the influence of noise with only one sensor, information from other sensors is used. A stable discrimination result can be obtained.

  FIG. 23 is an example of a configuration diagram illustrating a moving unit determination system 2300 with a multi-sensor integration function according to the seventh embodiment.

  In addition to the configuration of the movement means determination system 100 described in the first embodiment, the movement means determination system 2300 with a multi-sensor integration function includes other sensors 2313 and an integrated movement means determination section 2315. Instead, a sensor-specific moving means determination unit 2314 is provided. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As a hardware configuration, for example, as shown in FIG. 23, a factor-specific judgment reference database 103, a factor label assigning unit 104, a sensor-specific moving unit discriminating unit 2314, and an integrated moving unit discriminating unit 2315 are realized on a computer. The sensor 101 and the other sensors 2313 are realized by different terminals. In addition to this configuration, the entire configuration shown in FIG. 20 may be realized on a single terminal using, for example, a smartphone having both sensor and computer functions. It may be realized by a terminal. In any case, as long as the function shown in FIG. 23 can be realized, the way of combining hardware is not limited.

  Hereinafter, a method for performing moving means determination using a plurality of sensors in an integrated manner will be described. First, the moving means discriminating method described in the above embodiments is applied to each sensor, and moving means discrimination for each sensor is performed. That is, first, the data is divided into small segments, the norm variance value is calculated, and then the median value of a large segment obtained by collecting a certain number of them is calculated. Then, for example, the criterion value for each factor such as the road condition is read from the factor-specific criterion database 103 and compared with it to determine the moving means for the segment of interest.

  Here, the discrimination reference value is provided for each sensor as shown in FIG. The discrimination reference value for each sensor may be determined using learning data by the method described in the second embodiment, for example. In that case, after learning data is collected for each sensor, the discrimination reference value is learned. The discrimination criterion determination database is as shown in FIG. In addition, by appropriately adding the configuration described in the above embodiments, walking detection, stationary running determination, automatic determination of road conditions, correction of determination results using time-series determination results, and the like are performed. It's okay.

  Next, the final movement means determination result for each sensor is sent to the integrated movement means determination unit 2315. When the discrimination results from the respective sensors are obtained in this way, the integrated moving means discriminating unit 2315 uses these results in an integrated manner and outputs a final moving means discrimination result.

Specifically, for example, when discriminating between a car and a motorcycle, it can be discriminated that the vehicle is discriminated when the inequality represented by (Equation 4) is established, and that the vehicle is discriminated when it is not. Here, s _i in (Equation 4) is the type of sensor, S is a set of sensors, w _i is a weight to be given to the sensor s _i , C _i is 1 when the determination result of the sensor s _i is a car, numbers when a was the motorcycle becomes 0, B _i is 0 when the determination result of the sensor s _i was a car, is a value of 1 when was motorcycle.

Highly accurate discrimination can be performed by appropriately setting the weight w _i given to each sensor. It is possible to appropriately set a weight by detecting a time section in which noise is larger than that of other sensors and reducing the weight of the sensor for the time section. For example, if the acceleration / deceleration is intermittently continued, the acceleration sensor is reduced in weight. Further, the weight of the gyro sensor is reduced in a section with many curves. Alternatively, the weight of the magnetic azimuth sensor is reduced in a section traveling near a place where electromagnetic disturbance is large.

  The integration method described in the present embodiment is an example, and weighting can be performed by defining the reliability for each sensor in advance. In addition, there is no limitation on the integration method, such as assigning reliability to the discrimination result based on the degree of deviation from the discrimination reference value and using the value as a weight when discriminating each sensor.

  In the present embodiment, an example of a system capable of correcting determination of moving means after collecting the determination results of a plurality of moving means determining devices will be described.

  If the moving means is determined using only one portable terminal, the accuracy of the determination may be reduced due to noise caused by some factor, such as noise on the sensor value when the terminal has heat. Can happen.

  On the other hand, if the information of the terminals existing in the vicinity can be used, stable discrimination can be performed even in the presence of such noise. For example, if there are 5 or more terminals within a few meters, they take almost the same movement trajectory, and the characteristics of sensor data are similar, they move on the same vehicle. Can be determined. Assuming that only one of these terminals outputs a different discrimination result, the discrimination result can be corrected so as to be consistent with other terminal information.

  According to the configuration of the present embodiment, it is possible to determine the moving means with high accuracy by performing the correction processing of the moving means determination using the information of surrounding terminals on the server side.

  FIG. 25 is an example of a configuration diagram illustrating a moving means determination system with a peripheral terminal information use function according to the eighth embodiment.

  In addition to the configuration of the moving means determination system 100 described in the first embodiment, the moving means determination system 2500 with surrounding terminal information utilization function includes a GPS receiver 2516, a moving means determination result database 2517, and a determination result correction unit 2518. ing. In the moving means discriminating system 100 in FIG. 1, the description of the components having the same functions as those already described with reference to FIG. 1 is omitted. As the hardware configuration, for example, as shown in FIG. 25, the moving means discriminating unit 102, the factor-specific judgment reference database 103, and the factor label assigning unit 104 are realized on a computer, the moving means discriminating result database 2517, the discrimination result correcting unit 2518 is realized on a server, and the acceleration sensor 101 and the GPS receiver 2516 are realized on different terminals. In addition to this configuration, for example, a smartphone having both sensor and computer functions may be used, so that parts other than the server may be realized on one terminal, and the functions of the computer and server may be realized by one computer. It may be realized. In any case, as long as the function shown in FIG. 25 can be realized, the way of combining hardware is not limited.

  In the following, a method for correcting the discrimination result of the moving means using the peripheral terminal information will be described. First, the moving means determination is performed by the method described in the embodiments so far, and the result is stored in the moving means determination result database 2517. At this time, the ID of the terminal, the latitude / longitude information of the terminal obtained by using the GPS receiver 2516, and the date / time when the GPS information was received are also stored together for use in searching for surrounding terminals. . FIG. 26 shows an example of data stored in the moving means determination result database.

Next, the discrimination result correction unit 2518 searches for peripheral terminals of a certain terminal based on the data stored in the moving means discrimination result database 2517. Specifically, for example, a terminal ID whose position and time are within a certain value is searched from the moving means determination result database 2517. Assuming that (latitude, longitude) in decimal notation is (lat ₁ , lon ₁ ) (lat ₂ , lon ₂ ), the distance d between these two points can be calculated by (Equation 5). Here, r in (Expression 5) is a value representing the radius of the earth.

If the time when the two data were obtained was t ₁ and t ₂ , the time difference can be calculated as | t ₂ −t ₁ |. When d <θ _d and | t ₂ −t ₁ | <θ _t are satisfied by using these and appropriate threshold values θ _d and θ _t , it is assumed that the two terminals transmitting these data exist in the vicinity. judge. For the values of θ _d and θ _t , for example, 10 m, 1 second, etc. can be used. When such a condition continues for a certain time or longer, for example, continuously for 10 minutes or longer, it can be determined that these terminals exist on the same vehicle.

  In this way, when a plurality of terminals existing on the same vehicle are found, the moving means discrimination result is corrected. Specifically, for example, in the discrimination between a car and a motorcycle, when there are N terminals that are determined to exist on the same vehicle, if the determination result at half of the terminals exceeding N / 2 is a car, the same vehicle If the discrimination result of all terminals judged to be on the vehicle is a car, and the discrimination result on a terminal exceeding N / 2 is a motorcycle, the discrimination result of all terminals judged to be on the same vehicle is a motorcycle, If the result is divided into the same number of N / 2, it is sufficient not to perform correction. Of course, the method of correction is not limited to this, and the determination result is based on the degree of deviation from the determination reference value when determining the reliability for each terminal in which the reliability is pre-defined and weighted. Various correction methods are possible, for example, by assigning a reliability to and using the value as a weight.

  In this embodiment, an example of a system capable of automatically assigning a label about road conditions by estimating a region or the like from GPS information will be described.

  For example, when preparing a large amount of learning data, it takes a lot of work to manually provide labels related to road conditions such as road attributes and regions. On the other hand, if GPS information is used, the position information of the GPS receiver can be known. Since road conditions tend to differ depending on the location, it is thought that road conditions can be estimated from GPS information. Therefore, by using the configuration described in the present embodiment, this label can be automatically assigned from the GPS information, and the labor for labeling can be reduced.

  FIG. 27 is an example of a configuration diagram illustrating a moving means determination system 2700 with an automatic road condition determination function in the ninth embodiment.

  The moving means discriminating system 2700 with road condition automatic determination function includes GPS receivers 2719 and 2721 and road condition determining units 2720 and 2722 in addition to the configuration of the moving means discriminating system 600 with learning function described in the second embodiment. Yes. Note that the processing of the factor label assigning unit 104 is realized by the road condition determining units 2720 and 2722. In the moving means discrimination system 600 with learning function in FIG. 6, the description of the components having the same functions as those already described with reference to FIG. 6 is omitted.

  As the hardware configuration, for example, as shown in FIG. 27, the moving means discriminating unit 102, the factor-specific judgment criterion database 103, the discrimination criterion determining unit 604, and the discrimination criterion determining database 605 are realized on a computer, and the acceleration sensor 101, The GPS receiver 2719 and the road condition automatic determination unit 2720 are realized by a terminal such as a smartphone (referred to as a discrimination terminal), and the acceleration sensor 606, the GPS receiver 2721, and the road condition automatic determination unit 2722 are terminals such as a smartphone ( This is called a collection terminal). In addition to this configuration, for example, the discrimination terminal and the collection terminal may be realized by the same terminal, or the discrimination terminal and the computer part are used by using a smartphone having both the sensor and the computer functions. May be realized on one terminal, or all the configurations shown in FIG. 27 may be realized on one terminal. In any case, as long as the function shown in FIG. 27 can be realized, the way of combining hardware is not limited.

  Next, a procedure in which the road condition determination unit generates a road condition label from GPS information will be described. Here, an example will be described in which a region ID that is uniquely determined is generated using latitude and longitude information, and is used as a road status label when this is stored in the factor-specific discrimination criterion database 103 or the discrimination criterion determination database 605.

  The GPS information includes the latitude / longitude information of the GPS receiver. Here, it is assumed that these latitudes and longitudes are expressed in decimal notation. For example, a region ID that is uniquely determined can be generated by dividing each of the regions by 0.1 degrees and giving a unique ID to each region. In addition, latitude 0.1 degree corresponds to about 11km, longitude 0.1 degree is less than that (about 11km on the equator, the distance decreases with increasing latitude), so this is about 11km square or less Corresponds to the area. In such a region, it can be used as an ID indicating the road condition because it is unlikely that the road condition will greatly change in the region except for some exceptions. Of course, if necessary, it is possible to increase the uniformity of road conditions in the region by dividing it into smaller regions.

As a specific ID generation method, for example, when latitude is lat (degree, -90 ≦ lat ≦ 90) and longitude is lon (degree, -180 <lon <180), i = [lat * 10] (where [] indicates a Gaussian symbol and [X] means the largest integer not exceeding X), j = [lon * 10], and the ID of the point is A _{i, j} do it. According to this example, for example, when the latitude is 35.68 degrees and the longitude is 139.76 degrees, the generated ID is A _356,1397 .

  Using the area ID obtained in this way as a label representing road conditions, the factor-specific discrimination criterion database 103 and the discrimination criterion determination database 605 are constructed. Specifically, the area ID may be used in place of the road attributes in FIG. 5B and the areas in FIGS. 5C and 7. It is also possible to search for a region or road attribute corresponding to the region ID using map information or the like and use them as a label representing the road condition. FIG. 28A shows an example of a factor-specific discrimination standard database in this embodiment, and FIG. 28B shows an example of a discrimination standard determination database. After that, the method for determining the moving means is the same as the method described in the first embodiment, and thus will be omitted in this embodiment.

  In addition to the method described in the present embodiment, in order to generate a road condition label from GPS information, for example, a database for road condition determination is separately prepared, and information on latitude and longitude is previously stored therein. And the area ID may be stored, and the area ID (that is, the average road condition of the area) may be obtained by referring to the information. Alternatively, the coordinates of the representative point of the area may be determined, and the area ID may be given if the distance from the representative point is within a certain value. Further, for example, an unsupervised learning method such as the k-means method may be applied to divide the learning data into k classes based on latitude and longitude, and k road status IDs may be generated accordingly. Further, when determining which ID the discrimination data belongs to, a method such as a k-nearest neighbor method may be used, and means for determining a road condition label from GPS data is not limited.

  In addition to the configuration of this embodiment, the database for storing map information can be used to distinguish roads such as “highways”, “general roads” and “farm roads” by using existing map matching techniques. You may make it use as a label to represent.

  In the present embodiment, an example of a system that can utilize data without a moving means label as learning data will be described.

  As described in the embodiments so far, it is desirable that there is a large amount of learning data in order to perform accurate discrimination. However, when collecting data for learning, for example, it takes a certain amount of time to keep a record of whether it is a car or a motorcycle and give a label and upload it. This work is not preferred for obtaining.

  On the other hand, even for data without a label, the label may be estimated with high accuracy from time series information. Therefore, if such data is used effectively, a large amount of data can be collected without taking time and effort for labeling, and a discrimination reference value based on the data can be set.

  According to the configuration of the present embodiment, a small amount of data with a label of a moving means and a large amount of data without a label of a moving means are used, and data without a label is also used as learning data. By using it, it is possible to collect a large amount of learning data while reducing labor and to set a highly accurate discrimination reference value.

  FIG. 29 is an example of a configuration diagram illustrating a moving means determination system 2900 with an unlabeled data learning function according to the tenth embodiment.

  An unlabeled data learning function-equipped moving means discrimination system 2900 includes a labelless discrimination standard correction database 2925 and a discrimination standard correction unit 2926 in addition to the configuration of the learning function-equipped moving means discrimination system 600 described in the second embodiment. . In the moving means determination system 600 with a learning function in FIG. 6, the description of the components having the same functions as those already described in FIG. 6 is omitted.

  For example, as shown in FIG. 29, the hardware configuration includes a moving means determination unit 102, a factor-specific determination criterion database 103, a factor label assignment unit 104, a determination criterion determination unit 604, a determination criterion determination database 605, and an unlabeled determination criterion. The correction database 2925 and the discrimination reference correction unit 2926 are realized on a computer, and the acceleration sensor 101 and the acceleration sensor 606 are realized on different terminals. In addition to this configuration, for example, the same acceleration sensor 101 and acceleration sensor 606 may be used, or the entire configuration shown in FIG. You may implement | achieve on a terminal. In any case, as long as the function shown in FIG. 29 can be realized, the way of combining hardware is not limited.

  Hereinafter, a method for correcting the discrimination criterion will be described with reference to FIG. The other data collection part and movement means determination part are the same as the methods described in the above embodiments, and thus the description thereof is omitted here. Further, an example of discriminating between a car and a motorcycle will be described as the moving means.

  First, in 3001, learning data with a label of car or motorcycle is read from the discrimination criterion determination database 605, and the discrimination criterion value is determined according to the method described in the second embodiment. Next, at 3002, the discrimination reference correction unit 2926 reads data that is not labeled as a car or a motorcycle from the unlabeled discrimination reference correction database 2925, and applies the moving means discrimination method described in the second embodiment.

Here, the unlabeled discrimination reference correction database 2925 stores data that is not labeled as a car or a motorcycle as shown in FIG. As a result of applying the moving means discrimination method described in the second embodiment, for example, a result as shown in FIG. 31B is obtained. Here, when the front and rear information is taken into consideration, it is determined that the determination result of t _i-1 is likely to be a car. Therefore, next, at 3003, the number _Nj of data determined to require such correction is counted for each road condition. Here, whether or not correction is necessary is determined when, for example, the determination result differs only in the data of interest among the total of 5 data of the data of interest and the two data before and after the data of interest. It may be determined that correction is necessary.

Next, at 3004, for each data to be corrected, the difference between the median σ _{k, j of} the data and the corresponding discrimination reference value θ _j , Δθ _{k, j} = σ _{k, j} −θ _j is calculated. To do. Next, in 3005, a value of Δθ _j is calculated by summing up N _j values obtained by multiplying Δθ _{k, j} by a correction factor α _k , represented by (Expression 6).

Here, as the correction rate α _k , for example, 1 / N _j or the like may be used uniformly (this corresponds to taking the average of the differences in the data to be corrected) and should be corrected. Weighting may be performed based on the certainty degree to be determined. Specifically, for example, in the case where the discrimination result is different only in the data of interest among the total of 7 data of the data of interest and the three data before and after that, α _k = 1 / N _j , interest and if also another data other than data which the determination result is different α _k = 1 / 2N _j, if the focused also two data in addition to data that the determination result is different alpha _k = 1 / 3N _j , etc. Using [Delta] [theta] _j obtained in this way, finally, in 3006, the updated _{θ j ← θ j + Δθ j} . As a result, for example, as shown in FIG. 31C, the determination reference value can be corrected so that the determination result is likely to be a car.

  In this embodiment, an example of a moving means discrimination system that transmits data only when a GPS signal can be received and collects data will be described.

  When a large amount of learning data is collected via a network using a large number of terminals, there may be a problem that the load on the network line for data transmission increases and the communication speed decreases.

  On the other hand, since there is data that cannot be used for learning in the learning data, it is not always necessary to transmit all collected data. For example, in the moving means discriminating system with a road situation determination function described in the ninth embodiment, the GPS information is used to determine the road situation. Therefore, the data when the GPS data cannot be acquired is used as learning data. Cannot be used. Alternatively, even if it can be obtained, it is not appropriate to use it if the accuracy is extremely poor. Positioning using GPS information is impossible in principle unless radio waves can be received from four or more GPS satellites. For example, when it is hidden in a tall building or when traveling in a tunnel, it often happens that radio waves cannot be received from four or more GPS satellites and the accuracy is lowered. As described above, when the GPS signal is not received or the accuracy is remarkably low even if the GPS signal is received, the data on the road condition is manually labeled as described in the first embodiment. If it is not transmitted, it cannot be used as learning data. For this reason, even if data without a label is transmitted, uselessness occurs.

  Therefore, only necessary data is sent by adding a part to control whether data is transmitted depending on whether GPS information can be received and whether the accuracy is sufficient if it can be received. be able to. As a result, a large amount of learning data can be collected while reducing the load on the network.

  FIG. 32 is an example of a configuration diagram illustrating a moving means determination system 3200 with a data transmission control function according to the eleventh embodiment.

  The moving means discriminating system with data transmission control function 3200 has a transmission unit 807 described in the second embodiment and a transmission having a new configuration in addition to the configuration of the moving means discriminating system with road condition determining function 2700 described in the ninth embodiment. A control unit 3227 is provided. The transmission control unit 3227 controls whether or not to transmit data depending on whether or not the GPS data can be acquired and whether or not the GPS data is sufficiently accurate. In addition, in the moving means determination system 2700 of FIG. 27, description is abbreviate | omitted about the part which has the same code | symbol shown by already demonstrated FIG. 27, and the part which has the same function.

  As the hardware configuration, for example, as shown in FIG. 32, the moving means discriminating unit 102, the factor-specific judgment criterion database 103, the discrimination criterion determining unit 604, and the discrimination criterion determining database 605 are realized on a computer, and the acceleration sensor 101, The GPS receiver 2719 and the road condition automatic determination unit 2720 are realized by a terminal such as a smartphone (referred to as a determination terminal), and include an acceleration sensor 606, a GPS receiver 2721, a road condition automatic determination unit 2722, a transmission control unit 3227, The transmission unit 807 is realized by a terminal such as a smartphone (referred to as a collection terminal). In addition to this configuration, for example, the discrimination terminal and the collection terminal may be realized by the same terminal, or the discrimination terminal and the computer part are used by using a smartphone having both the sensor and the computer functions. May be realized on a single terminal, or all the configurations shown in FIG. 32 may be realized on a single terminal. In any case, as long as the function shown in FIG. 32 can be realized, the way of combining hardware is not limited.

  In the present embodiment, an example of a moving means discriminating system that can be controlled to perform measurement by an acceleration sensor only when a GPS signal can be received will be described.

  In order to measure acceleration using an acceleration sensor, relatively large power is consumed. Further, if measurement is continued at a high sampling rate, memory consumption for storing the data also increases.

  On the other hand, as described in the eleventh embodiment, in the moving means discriminating system with a road condition determination function described in the ninth embodiment, when the GPS signal cannot be received or is received, the road condition is extremely low. Cannot be determined. Therefore, even if only the learning data collection sensor can be measured, it cannot be used as learning data, and even if the data is transmitted, waste is generated. Therefore, it can be said that it is not necessary to measure the acceleration sensor in the first place when the GPS information is not acquired or when the accuracy is remarkably low.

  Therefore, the power consumption and the memory consumption can be reduced by performing the measurement by the acceleration sensor only when the GPS information can be acquired and the accuracy is sufficient.

  FIG. 33 is an example of a configuration diagram illustrating a moving means determination system 3300 with a sensor On / Off switching function according to the twelfth embodiment.

  The moving means discriminating system 3300 with a sensor On / Off switching function includes a measurement On / Off switching unit 3328 in addition to the configuration of the moving means discriminating system 2700 with a road condition determining function described in the ninth embodiment. The measurement On / Off switching unit 3328 controls whether the acceleration sensor is turned on or turned off depending on whether GPS data is acquired and whether the accuracy is sufficient. In addition, in the moving means determination system 2700 of FIG. 27, description is abbreviate | omitted about the part which has the same code | symbol shown by already demonstrated FIG. 27, and the part which has the same function.

  As the hardware configuration, for example, as shown in FIG. 33, the moving means discriminating unit 102, the factor-specific judgment criterion database 103, the discrimination criterion determining unit 604, and the discrimination criterion determining database 605 are realized on a computer, and the acceleration sensor 101, The GPS receiver 2719 and the road condition automatic determination unit 2720 are realized by a terminal such as a smartphone (referred to as a determination terminal), and the acceleration sensor 606, the GPS receiver 2721, the road condition automatic determination unit 2722, and the measurement on / off switching. The unit 606 is realized by a terminal such as a smartphone (referred to as a collection terminal). In addition to this configuration, for example, the discrimination terminal and the collection terminal may be realized by the same terminal, or the discrimination terminal and the computer part are used by using a smartphone having both the sensor and the computer functions. May be realized on one terminal, or all the configurations shown in FIG. 33 may be realized on one terminal. In any case, as long as the function shown in FIG. 33 can be realized, the way of combining hardware is not limited.

  In the present embodiment, an example of a road congestion situation estimation system that estimates the road congestion situation and usage situation using the result of the moving means determination will be described.

  When making a road maintenance plan, there is a demand for grasping the current congestion and usage of roads. On the other hand, conventionally, a method of collecting traffic information using an in-vehicle device or a method of collecting traffic information by installing a radio beacon on the road side has been used. However, particularly in emerging countries, there is a problem that the installation cost of in-vehicle devices and beacons is high, and their diffusion has not progressed. On the other hand, mobile terminals such as smartphones equipped with a GPS function are spreading, and there is an expectation for a road congestion state estimation technique using this. By calculating the position and vehicle speed from the GPS information, it is possible to estimate road congestion and usage. However, unlike Japan's traffic situation where the majority of roads are cars, motorcycles that are cheaper than cars are widespread, especially in emerging countries. And running. Since cars and motorcycles have different body sizes, it can be said that even if the road is congested for a car, the bike can pass through between the vehicles, so the bike is not so crowded. For this reason, if the road congestion situation is estimated without distinguishing between a car and a motorcycle, there is a problem that the estimation is not accurate and does not represent the actual situation.

  According to the configuration of the present embodiment, it is possible to estimate road congestion and usage conditions after discriminating between a car and a motorcycle by the method described in the previous embodiments, so that an estimation result closer to the actual situation is obtained. be able to.

  In the following, a method for estimating the road congestion state will be mainly described. However, since the road use state can also be estimated by a similar method, the description is omitted.

  FIG. 34 is an example of a configuration diagram illustrating a road congestion situation estimation system 3400 according to the thirteenth embodiment.

  A road congestion situation estimation system 3400 includes the moving means determination system 100 described in the first embodiment, a congestion situation estimation database 3401, a congestion situation estimation unit 3402, and a GPS receiver 3403. Here, the set of the moving means discriminating system 100 and the GPS receiver is not limited to one, and a plurality of moving means discriminating systems and GPS receivers may be used. Further, the moving means discriminating system 100 may be replaced by using a moving means discriminating system provided with other functions described in the second embodiment or later.

  As a hardware configuration, for example, as shown in FIG. 34, a congestion state estimation database 3401 and a congestion state estimation unit 3402 are realized by a single computer. As described in the first embodiment, the configuration of the moving unit determination system 100 may be realized by separately preparing a sensor and a computer, or may be realized on one terminal using a smartphone or the like. . In addition to this configuration, for example, the congestion state estimation database 3401 and the congestion state estimation unit 3402 may be realized by using a computer used for moving means determination. In any case, as long as the function shown in FIG. 34 can be realized, the way of combining hardware is not limited.

  Below, operation | movement of the road congestion condition estimation system 3400 is demonstrated. First, each moving means determination system 100 performs moving means determination. The result, the terminal ID, the position information (latitude and longitude) obtained from the GPS receiver 3403, the speed that can be calculated by the method described in the fourth embodiment from the GPS information, and the data acquisition date and time are transmitted together. As a result, the data shown in FIG. 35 is stored in the congestion state estimation database 3401, for example. For example, if the length of a large segment, which is a unit for determining the moving means, is 900 seconds, for example, and data for estimating road congestion status is generated every second, 900 lines of data shown in FIG. Are sent together. At this time, the determination results of the moving means are the same for the 900 rows of data.

When these data are obtained from a plurality of portable terminals, the congestion state estimation unit 3402 estimates the road congestion state. Specifically, for example, the region ID is determined from the latitude and longitude information by the method described in the ninth embodiment, and the average speed is calculated for each car and motorcycle within the region. If the value is below a certain value, it is congested, and if it is above a certain value, it is judged that the vehicle is not congested separately. Thereby, the congestion situation of the car and motorcycle of an applicable area can be estimated. Of course, in addition to these two stages, it can be divided into multiple stages according to the average speed, such as vacant, somewhat vacant, normal, somewhat congested, and congested. Further, instead of discretization in this way, for example, the average speed may be converted by some function to calculate a continuous road congestion index. For example, the road congestion index I can be calculated as (Equation 7). In (Expression 7), v _Max is the maximum speed in the area or road such as legal speed, v _Min is the minimum speed such as 0, and v _mean is the average speed calculated as described above.

  In (Equation 7), when the average speed is the same as the maximum speed, the road congestion index is 0, and when the average speed is the same as the minimum speed, the road congestion index is 100. According to this, a value between 0 and 100 is taken.

  In addition, although the example which estimates a road congestion condition by a simple method was described here, since there are many techniques for estimating a road congestion condition from GPS information, the road congestion condition is applied by applying these existing technologies. May be estimated. In any case, any means for estimation may be used as long as the congestion state estimation data is collected and the congestion state with respect to each of the car and the motorcycle can be estimated after the moving means is determined.

  In the present embodiment, an example of a road congestion situation estimation system capable of performing detailed road congestion situation estimation by using map information will be described.

  In Example 13, the example of the road congestion condition estimation system which does not require road map information was described. By dividing the area that can be limited from GPS information, it is possible to estimate the road congestion in a certain area. However, in order to estimate more detailed information, for example, the congestion situation for each road and link it to an actual road, a road map is required. Therefore, in this embodiment, a system that can estimate a detailed road congestion state associated with an actual road by using a map database will be described.

  FIG. 36 is an example of a configuration diagram illustrating a map-use road congestion situation estimation system according to the fourteenth embodiment.

  In addition to the configuration of the road congestion situation estimation system 3400 described in the thirteenth embodiment, the map use road congestion situation estimation system 3600 includes a map database 3605 and a map utilization congestion situation estimation unit 3604 instead of the congestion situation estimation unit 3402. Yes. In the road congestion situation estimation system 3400 in FIG. 34, the description of the components having the same functions as those already described with reference to FIG. 34 is omitted.

  As a hardware configuration, for example, as shown in FIG. 36, a congestion state estimation database 3401, a map use congestion state estimation unit 3604, and a map database 3605 are realized by a single computer. As described in the first embodiment, the configuration of the moving unit determination system 100 may be realized by separately preparing a sensor and a computer, or may be realized on one terminal using a smartphone or the like. . In addition to this configuration, for example, the congestion state estimation database 3401 and the congestion state estimation unit 3402 may be realized by using the computer used for moving means determination, and the map utilization database is realized on another computer. You may do it. In any case, as long as the function shown in FIG. 36 can be realized, the way of combining hardware is not limited.

  Below, operation | movement of the map utilization congestion condition estimation part 3604 is demonstrated. The map information utilization congestion state estimation unit 3604 reads map information from the map database 3605 and obtains the road on which the moving means determination system 100 existed at the time of data transmission by an arbitrary existing method such as map matching. As a result of the map matching, if it can be determined which road is running, as described in the thirteenth embodiment, the speeds of the cars and motorcycles are totaled for each road, and the average value of the speeds is calculated. If the value is below a certain value, the road is congested, and if it is above a certain value, it is determined that the road is not congested separately for each car and motorcycle. Thereby, the congestion situation of the car and motorcycle of an applicable area can be estimated.

  In the present embodiment, an example of a road congestion situation estimation system capable of displaying the road congestion situation for each movement means using the results of movement means discrimination and congestion situation estimation will be described.

  Even if the estimation result of the road congestion situation for each moving means is obtained by the method described in the embodiment 13 and the embodiment 14, any road is congested only by holding it as text information. It is difficult to intuitively understand which road is not crowded. In addition, it is difficult to compare the congestion status of each moving means on the same road.

  On the other hand, for example, when the road congestion state is displayed by overlapping with a map for each moving means, or the display method is changed for each moving means, the visibility is improved, for example, for each moving means in a certain area. The traffic flow can be grasped intuitively. Using this road congestion information for each means of transportation, for example, when making a new road construction plan or setting new traffic regulations, for example, the number of motorcycle lanes, car lanes, and bus lanes should be set appropriately. It is possible to make various plans according to the local circumstances, such as by restricting traffic on certain roads so that only buses can pass during rush hour.

  FIG. 37 is an example of a configuration diagram illustrating a road congestion state estimation system with a congestion state display function by moving means according to the fifteenth embodiment.

  In addition to the configuration of the road congestion situation estimation system 3400 described in the thirteenth embodiment, the road congestion situation estimation system 3700 with a congestion situation display function by movement means includes a congestion situation display unit 3706 by movement means. In the road congestion situation estimation system 3400 in FIG. 34, the description of the components having the same functions as those already described with reference to FIG. 34 is omitted.

  As a hardware configuration, for example, as shown in FIG. 37, a congestion state estimation database 3401, a congestion state estimation unit 3402, and a road state display unit 3706 for each moving means are realized by a single computer. As described in the first embodiment, the configuration of the moving unit determination system 100 may be realized by separately preparing a sensor and a computer, or may be realized on one terminal using a smartphone or the like. . In addition to this configuration, for example, the congestion state estimation database 3401 and the congestion state estimation unit 3402 may be realized by using the computer used for the movement unit determination, and the movement state-specific congestion state display unit 3706 is different. It may be realized on a computer. In any case, as long as the function shown in FIG. 37 can be realized, the way of combining hardware is not limited.

  Hereinafter, the operation of the congestion status display unit 3706 for each moving means will be described. Here, in addition to the above-described minimum configuration, a method of displaying the congestion status according to the map using the map database 3605 described in the embodiment 14 will be described. Thereby, it is possible to grasp a more detailed road congestion situation. Of course, even if map information is not available, the same method can be realized if the method described in the present embodiment, which is applied to each road, is obtained for each region ID using the method described in the thirteenth embodiment and applied to each region ID. .

  First, the congestion status display unit 3706 for each moving means receives the congestion status estimation result for each moving means for each road from the congestion status estimation unit 3402. According to the result, for example, as shown in FIG. 38, the color is classified for each moving means, and the average speed is represented by the length of the arrow, and the road congestion state is displayed on the display device. Of course, other display methods are possible. For example, the average speed may be represented by shades of color, or an animation function may be added to display an icon representing a moving means or a color-coded mark for each moving means. The vehicle may be moved on the road, and the speed of the movement may be set according to the average speed calculated by the road congestion state estimation unit. At this time, the visibility can be further improved by adding a device such as highlighting a busy intersection. In addition, it is expected that road congestion will vary from time to time. Therefore, for example, the above-described display can be divided and displayed every hour to make it easy to grasp the transition of the traffic volume for each time zone. Furthermore, by adding a storage unit as necessary and accumulating such information, for example, when a certain traffic regulation is performed, how the traffic flow is changed by the current traffic It can also be grasped by comparing the road congestion situation and the past road congestion situation. In addition, by combining such information with existing simulation technology, for example, it may be possible to display how the traffic flow is expected to change when a certain traffic regulation is introduced, for example, by moving means. In any case, any means may be used as long as it is a method for improving the visibility at the time of display by changing the display method according to the moving means or changing the display method such as color coding or shape.

  Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications can be made. It is possible to appropriately combine the above-described embodiments. It will be understood by the contractor.

DESCRIPTION OF SYMBOLS 100 Moving means discrimination | determination system 101 Acceleration sensor 102 Moving means discrimination | determination part 103 Factor discrimination | determination discrimination | determination reference | standard database 200 Portable terminal 201 Input control apparatus 202 Acceleration sensor 203 Central processing unit 204 Storage apparatus 205 Screen display apparatus 206 Bus 300 Computer 301 Input control apparatus 302 Central processing unit 303 Main storage unit 304 Auxiliary storage unit 305 Output control unit 306 Bus 310 Input unit 320 Output unit 600 Moving means discrimination system with learning function 604 Discrimination criterion determination unit 605 Discrimination criterion determination database 606 Learning data collection acceleration sensor 807 Transmission unit 1000 Moving means discriminating system with correction function 1008 Time series discriminating information storage unit 1300 Moving means discriminating system with still running judgment function 1309 Still running judgment unit 1700 With walking exclusion function Moving means discriminating system 1710 Walking detecting section 2000 Moving means discriminating system with characteristic moving means excluding function 2011 Characteristic moving means detecting section 2012 GPS receiver 2300 Moving means discriminating system with multiple sensor integration function 2313 Other sensors 2314 Moving means discriminating by sensor Unit 2315 integrated moving unit discriminating unit 2500 moving unit discriminating system 2516 with surrounding terminal information utilization function GPS receiver 2517 moving unit discriminant result database 2518 discrimination result correcting unit 2700 moving unit discriminating system with road condition determining function 2719 GPS receiver 2720 road Situation determining unit 2721 GPS receiver 2722 Road condition determining unit 2900 Moving means discriminating system with unlabeled data learning function 2925 Unlabeled discriminant reference correction database 2926 Discriminant reference correcting unit 3200 Data transmission Moving means discriminating system with control function 3227 Transmission control unit 3300 Moving means discriminating system with sensor On / Off switching function 3328 Measurement On / Off switching unit 3400 Road congestion state estimation system 3401 Congestion state estimation database 3402 Congestion state estimation unit 3403 GPS reception 3600 Map usage road congestion status estimation system 3604 Map usage congestion status estimation unit 3605 Map database 3700 Road congestion status estimation system 3706 with movement status display function by movement means Congestion status display section by movement means

Claims (15)

A first sensor;
An assigning unit that assigns environmental information regarding the environment from which the data is acquired to the data acquired by the first sensor;
A storage unit for storing a discrimination reference value for discriminating a moving means of a terminal on which the first sensor is mounted in association with each of a plurality of environment information;
A moving unit determining unit that selects a predetermined determination reference value from the storage unit according to the given environment information, and determines the moving unit of the terminal using the data and the predetermined determination reference value; A moving means discriminating system. It is a moving means discrimination | determination system of Claim 1, Comprising:
A second sensor of the same type as the first sensor;
A moving unit discriminating system further comprising: a discriminant reference determining unit that determines the discriminant reference value stored in the storage unit using learning data acquired by the second sensor. It is a moving means discrimination | determination system of Claim 1, Comprising:
Using the data acquired by the first sensor, further comprising a stationary travel determination unit that determines whether the terminal is stationary or traveling;
The moving means discriminating system, wherein the moving means discriminating unit changes a discrimination reference value for discriminating the moving means of the terminal depending on whether the moving state is the stationary state or the traveling state. It is a moving means discrimination | determination system of Claim 1, Comprising:
A GPS receiver that obtains location information of the terminal;
A determination result correction unit that searches for peripheral terminals located around a predetermined terminal using the position information of a plurality of terminals and the time at which the position information is acquired; and
The discrimination result correction unit is a moving means discriminating system that corrects the moving means discrimination result of the predetermined terminal by using the moving means discrimination result of the peripheral terminal. It is a moving means discrimination | determination system of Claim 1, Comprising:
A GPS receiver that obtains location information of the terminal;
A road condition determination unit that determines a region or a road attribute when the data is acquired from the position information;
The assigning unit is a moving means determination system that assigns the region or the road attribute as the environment information. It is a moving means discrimination | determination system of Claim 2, Comprising:
A GPS receiver that obtains GPS information;
And a switching unit that controls whether to perform measurement by the second sensor according to the accuracy of the GPS information. It is a moving means discrimination | determination system of Claim 1, Comprising:
The moving means discriminating system, wherein the first sensor is an acceleration sensor for detecting vibration. A road congestion situation estimation system having the moving means discrimination system according to claim 1,
A GPS receiver that obtains location information of the terminal;
In each of a plurality of terminals, using the position information, the moving speed calculated from the position information, the acquisition time of the data, and the determination result of the moving means, a congestion situation estimation unit that estimates the congestion situation of the road, Road congestion situation estimation system. The road congestion situation estimation system according to claim 8,
A road congestion status estimation system having a congestion status display section for each moving means for displaying the congestion status of the road for each moving means. A moving means discriminating apparatus for discriminating moving means of a terminal equipped with the first sensor using data acquired by a first sensor,
Environmental information regarding the environment from which the data was acquired is given to the data,
A storage unit that stores a determination reference value for determining the moving means of the terminal in association with each of a plurality of environment information;
A moving unit discriminating unit that selects a predetermined discriminant reference value from the storage unit according to environmental information given to the data, and discriminates a moving unit of the terminal using the data and the predetermined discriminant reference value. And a moving means discriminating apparatus. The movement means determination device according to claim 10,
A moving means discriminating apparatus further comprising a discriminant reference determining unit for determining the discriminant reference value stored in the storage unit using learning data acquired by a second sensor of the same type as the first sensor. The movement means determination device according to claim 10,
Using the data acquired by the first sensor, further comprising a stationary travel determination unit that determines whether the terminal is stationary or traveling;
The moving means determining device is configured to change a determination reference value for determining the moving means of the terminal according to whether the moving means is in the stationary state or the running state. The movement means determination device according to claim 10,
Using a positional information of a plurality of the terminals acquired by the GPS receiver and a time when the positional information is acquired, and further including a determination result correction unit that searches for peripheral terminals located around a predetermined terminal;
The discrimination result correction unit is a moving means discriminating apparatus that corrects the moving means discrimination result of the predetermined terminal using the moving means discrimination result of the peripheral terminal. A road congestion situation estimation system comprising the moving means determination device according to claim 10,
The location information of the terminal is acquired by the GPS receiver,
In each of a plurality of terminals, a congestion state estimation unit that estimates a road congestion state using the position information, the movement speed calculated from the position information, the data acquisition time, and the determination result of the moving means Road congestion estimation system. A moving unit determination program for causing a moving unit determination device to determine a moving unit of a terminal on which the first sensor is mounted using data acquired by a first sensor,
Environmental information regarding the environment from which the data was acquired is given to the data,
In correspondence with each of a plurality of environment information, a determination reference value for determining the moving means of the terminal is stored,
A predetermined discrimination reference value is selected from the stored discrimination reference value according to the environmental information given to the data, and the moving means of the terminal is discriminated using the data and the predetermined discrimination reference value. Moving means discrimination program.

Images