JP5704561B2 - Traveling direction estimation device, portable terminal, control program, computer-readable recording medium, and traveling direction estimation method - Google Patents

Description

  The present invention relates to a traveling direction estimation device for estimating a traveling direction of a moving body.

  A person as a pedestrian using a built-in sensor (acceleration, gyroscope, magnetism, temperature, barometric pressure sensor, etc.) built into a terminal worn on a person's waist or toe, or a terminal held by a person in a free posture. In particular, a technique for measuring the position / orientation is called pedestrian dead reckoning (PDR = Pedestrian Dead Reckoning), and has been well studied in the past (Non-patent Document 1).

  In most cases, pedestrians are subject to the constraint of walking and moving on a plane. In the PDR, the position / orientation of a pedestrian is measured in consideration of such constraints. For this reason, PDR roughly includes the following two elemental technologies.

  That is, there are two techniques: (i) a technique for estimating the movement azimuth angle of a pedestrian on a plane and (ii) a technique for estimating the stride (walking speed). The present invention relates to the technique (i) for estimating the moving azimuth angle. Note that the technique (ii) for estimating the stride (walking speed) is disclosed in Patent Document 1, for example.

In (i) above, the azimuth of the pedestrian on the plane is estimated by estimating (tracking) the three vectors shown in FIG. That is,
(1) Gravity orientation vector 501 in the sensor coordinate system
(2) Horizontal reference direction vector 502 (for example, true north direction) in the sensor coordinate system
(3) Traveling direction vector 503 in the sensor coordinate system
It is three.

  By estimating (tracking) the vectors (1) to (3) above, an angle formed by the traveling azimuth (= azimuth angle 504) with respect to the horizontal reference azimuth is estimated, and the movement azimuth angle of the pedestrian on the plane is estimated. Estimation can be performed.

  A method for estimating (tracking) the vectors (1) to (3) will be described as follows.

  First, the gravity direction vector 501 of the above (1) can be estimated as follows. The gravity direction vector 501 is basically in the same direction as the gravity acceleration vector. Therefore, tracking of the gravitational azimuth can be realized using a framework such as a Kalman filter by combining the outputs of the acceleration sensor and the angular velocity sensor and the signal processing thereof.

  For example, Non-Patent Document 1 discloses a specific method for tracking the gravity direction vector 501. A technique using magnetism is also known for tracking the gravity direction vector 501.

  Next, for the horizontal reference orientation vector 502 in (2) above, the true north orientation can be estimated in principle by correcting the declination of the horizontal component of the geomagnetic vector observable by the magnetic sensor. is there. Such tracking can be realized by combining the output of acceleration, angular velocity, and magnetic sensor with signal processing. For example, Non-Patent Document 1 shows a specific method for tracking the horizontal reference orientation vector 502.

  Next, regarding the traveling direction vector 503 in (3) above, when sensing means (for example, acceleration, angular velocity, temperature, magnetic sensor) are fixedly attached to the pedestrian's waist, etc. There are different things to consider.

  When the relationship between the posture of the sensing means and the pedestrian is fixed, the direction of travel is generally known because the relationship between the posture of the sensing means and the pedestrian is fixed. Therefore, in this case, the traveling direction vector 503 need not be tracked.

  On the other hand, when the relationship between the posture of the sensing means and the pedestrian is not fixed, it is impossible to handle the direction of travel with the pedestrian as a known parameter, so obtaining the direction of travel of the pedestrian is a technical problem. . As an example of such a case, when a smartphone or mobile terminal equipped with sensing means is held in an arbitrary posture / position by the user, or carried in a pocket or bag And the like.

  Therefore, various proposals have been made for the above technical problems (Patent Documents 2 to 4).

  For example, Patent Document 2 discloses a technique for measuring the advancing direction of a pedestrian from acceleration when walking with a posture fixed or when a terminal is swung.

JP 2005-114537 A (published April 28, 2005) International Publication No. 2006/104140 (October 5, 2006) JP 2003-302419 A (released on October 24, 2003) JP 2008-116315 A (published May 22, 2008)

Masakatsu Oki, Takashi Otsuki, Takeshi Kurata, "Indoor positioning system using self-contained sensor module for pedestrian navigation and its evaluation", Proceedings of Symposium "Mobile 08", 2008, pp151-156

  However, in the above prior art, there is a case where the optimal traveling direction cannot be estimated when the relative position between the moving body and the apparatus changes.

  The present invention has been made in view of the above problems, and its purpose is to make it possible to accurately estimate the traveling direction from the moving operation of the moving body even if the relative position between the moving body and the apparatus changes. The purpose is to realize a direction estimation device and the like.

  In order to solve the above problems, a traveling direction estimation apparatus according to the present invention estimates a traveling direction when a moving body moves by a constant moving operation using measurement data of a sensor that measures a motion state of the moving body. In the traveling direction estimation apparatus, a data acquisition unit that acquires measurement data that is measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor that is the sensor and is represented by a predetermined coordinate system, and the data acquisition unit acquires The measured data is decomposed into a vertical component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, and a provisional lateral direction component orthogonal to the provisional traveling direction and the vertical direction. Component decomposing means, and patterns appearing in the provisional traveling direction component and the provisional transverse direction component obtained by the component decomposing means. The pattern comparison means for comparing with a predetermined pattern representing the periodicity of the acceleration or angular velocity in the traveling direction and the lateral direction of the moving body at the time of the moving operation, and the pattern as a result of comparison by the pattern comparison means And output means for outputting the provisional traveling direction as the estimated traveling direction when the above conditions are met.

  In order to solve the above problem, the traveling direction estimation method according to the present invention uses the measurement data of a sensor that measures the motion state of the moving body, and the traveling direction when the moving body moves by a constant moving operation. In the traveling direction estimation method for estimating the data, a data acquisition step for acquiring measurement data measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor as the sensor and represented by a predetermined coordinate system; and the data acquisition step The measurement data acquired in step 1 is converted into a vertical direction component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, and a provisional lateral direction component orthogonal to the provisional traveling direction and the vertical direction. The component decomposition step for component decomposition, and the provisional traveling direction component and the provisional lateral direction component obtained by the component decomposition step For each of the patterns, the pattern comparison step for comparing with the predetermined pattern representing the periodicity of the moving direction and lateral acceleration or angular velocity at the time of the moving operation, and the result of the comparison in the pattern comparison step, An output step of outputting the temporary traveling direction as the estimated traveling direction when the pattern matches a predetermined level or more.

  Before describing the operations and effects derived from the above configuration, a general concept framework applied to the present invention will be described.

(Theoretical framework)
There are cases where reproducibility and periodicity appear in the movement pattern of the moving motion of the moving body. The moving operation refers to, for example, an operation when a person moves with both feet. As an example of an operation when a person moves, an operation such as running on a flat ground, walking, or going up a slope or stairs can be cited.

  In addition, human walking motion is highly controlled and highly reproducible exercise, and the pattern of motion applied to the center of gravity of a person (pedestrian) is highly reproducible and periodic. For this reason, when a person walks, a pattern with a high characteristic periodicity appears in the acceleration vector and the angular velocity vector applied to the center of gravity.

  Even if a sensor built into a mobile phone or smartphone is held in a person's hand or placed in a pocket of clothes or a handbag, the movement of its center of gravity is basically transmitted. Therefore, the same movement tends to be observed. That is, the motion of the center of gravity of the moving body is transmitted to a mobile phone or the like equipped with the traveling direction estimation device.

  Therefore, the traveling direction of the moving body can be predicted by analyzing the acceleration vector or the angular velocity vector indicating the acceleration or the angular velocity applied to the moving body from a pattern characteristic to the moving operation of the moving body.

  Hereinafter, characteristic patterns regarding the acceleration vector and the angular velocity vector during the walking motion will be described.

[Characteristic pattern of acceleration vector]
As for the acceleration vector, a pattern with one step as one cycle appears in the vertical component, and a pattern with one step as one cycle also appears in the component in the traveling direction. Also, a pattern with two steps as one cycle appears in the horizontal component.

  Such feature points also appear in the power spectrum obtained by Fourier transforming such a waveform.

[Characteristic pattern of angular velocity vector]
For the angular velocity vector, a pattern with two steps as one cycle appears in the angular velocity component around the yaw axis (ie, the vertical axis), and one step appears in the angular velocity component around the pitch axis (ie, the horizontal axis). A pattern with a period of 1 appears. A pattern with two steps as one cycle appears in the angular velocity component around the roll axis (that is, the axis in the traveling direction).

[Estimation of traveling direction]
It has been known in advance that the aforementioned features are observed for acceleration vectors or angular velocity vectors. For this reason, the traveling direction can be estimated by determining the axis so that this feature appears on each axis. The vertical component axis can be determined, for example, by obtaining a gravity direction vector. The gravity direction vector can be measured using sensor data of an acceleration sensor, a gyro sensor, a geomagnetic sensor, and the like.

  The above-mentioned features are also observed for the acceleration differential vector and the angular acceleration vector, which will be described in detail later.

(About the above configuration)
According to the above configuration, the component of the measurement data decomposed using the provisional traveling direction is compared with the predetermined pattern representing the characteristics of the movement operation. An arbitrary coordinate system can be adopted as the predetermined coordinate system, for example, a world coordinate system can be adopted. The world coordinate system is a sensor-specific coordinate system converted to a global coordinate system, and is sometimes referred to as a global geodetic system.

  In the component decomposition, the vertical direction component in the coordinate system is obtained in advance by obtaining a gravity direction vector. Therefore, the plane perpendicular to the vertical axis can be obtained based on the vertical component thus obtained.

  The temporary travel direction temporarily indicates the travel direction on the plane and can be set arbitrarily. And in the said structure, the axis | shaft of a temporary advance direction and the axis | shaft orthogonal to a temporary advance direction in the said plane, ie, the axis | shaft of a temporary horizontal direction, are determined.

  In this way, the measurement data is decomposed into a vertical direction component, a provisional traveling direction component, and a provisional transverse direction component.

  The pattern representing the periodicity of the moving operation is a characteristic pattern showing the periodicity that appears when the acceleration or the angular velocity actually generated in the traveling direction and the lateral direction is actually measured in the moving operation exemplified above. . Examples of the pattern representing the periodicity of the moving operation include the characteristic pattern in the walking operation. As this characteristic pattern, a pattern that is measured and stored in advance by measuring the acceleration or angular velocity in the traveling direction and the lateral direction in the moving operation may be used.

  Pattern comparison is to collate waveform patterns having such periodicity or to compare with characteristic points of the power spectrum.

  And according to the said structure, when the pattern is matched more than a predetermined degree as a result of the comparison of a pattern, the said temporary moving direction is output as an estimated moving direction.

  That is, at least one of the pattern matching between the provisional traveling direction component and the traveling direction component of the characteristic pattern, and the pattern matching between the provisional lateral direction component and the lateral direction component of the characteristic pattern is not less than a predetermined level. If there is, it may be determined that the pattern matches a predetermined degree or more.

  As described above, according to the above configuration, the provisional traveling direction is changed depending on whether the pattern of the provisional traveling direction component and the provisional lateral direction component matches the actual characteristic pattern or more. , And can be output as the estimated traveling direction. If the pattern matches above a predetermined level, it can be said that the provisional travel direction is closer to the actual travel direction.

  As a result, even if the relative position between the moving body and the apparatus changes, the traveling direction can be accurately estimated from the moving operation of the moving body.

  Note that the above pattern comparison may be performed for the vertical direction component.

  In the traveling direction estimation apparatus according to the present invention, there is provided time differentiation means for differentiating the measurement data represented by the coordinate system with respect to time, and the data acquisition means acquires the measurement data that has been time differentiated by the time differentiation means. It is preferable.

  In recent years, a MEMS (MicroElectroMechanical System) type acceleration sensor often used has a zero G offset (that is, an offset output in a state where no acceleration is applied) at a level that cannot be ignored. For this reason, when using the output of the acceleration sensor, it is necessary to correct appropriately, but the procedure is complicated.

  Furthermore, the characteristics of the offset output output from the acceleration sensor change due to external factors such as temperature. For this reason, if the raw output of the acceleration sensor is used as a feature quantity to be analyzed, it is difficult to correct the influence of the offset output or to cope with a characteristic change due to an external factor.

  By the way, with respect to the output of the acceleration sensor sampled at a sufficiently high speed, when the change in posture is small and can be ignored, the acceleration differential (so-called jerk) is obtained from the difference data of the acceleration sensor sample at time t and time t + Δt. Can be obtained.

  Here, when the samples are sampled at a very short interval, the zero G offset outputs included in the outputs of the adjacent acceleration sensors (for example, at time t and time t + Δt) can be regarded as the same. For this reason, in the difference data, the zero G offset output is canceled and is not affected by the offset output.

  In addition, a pattern having the same periodic property as the above-described acceleration pattern appears in the time differentiation of acceleration (however, the phase is shifted by 90 degrees).

  Therefore, there is an advantage that the differential amount of acceleration is not affected by the offset output of the acceleration sensor as the characteristic amount to be analyzed.

  According to the above configuration, since the differentiated measurement data is converted to obtain post-conversion data, the influence of the zero G offset of the acceleration sensor can be reduced. The same is true for angular velocity.

  In the traveling direction estimation apparatus according to the present invention, the traveling direction estimation device includes a power spectrum acquisition unit that acquires a power spectrum for the temporary traveling direction component and the temporary lateral direction component obtained by the component decomposition unit, and the pattern comparison unit includes It is preferable to compare the power spectrum acquired by the power spectrum acquisition means with a power spectrum pattern representing periodicity of acceleration or angular velocity in the traveling direction and the lateral direction of the moving object during the moving operation.

  According to the above configuration, by comparing with the power spectrum pattern representing the periodicity of the moving operation, it is possible to relatively easily compare with the pattern representing the periodicity of the moving operation. The power spectrum can be obtained by applying a one-dimensional fast Fourier transform to each axis component.

  In the traveling direction estimation apparatus according to the present invention, the component decomposition unit performs the component decomposition in the temporary traveling direction set for each direction, and the pattern comparison unit includes the respective directions obtained by the component decomposition unit. The provisional traveling direction component and the provisional lateral direction component are compared with the predetermined pattern, and the output means, based on the result of the comparison by the pattern comparison means, is the best among the preliminary traveling directions in each direction. It is preferable to output the provisional travel direction that matches this pattern as the estimated travel direction.

  According to the above configuration, the temporary value can be updated in order from a predetermined direction in all directions on the plane, and the best match with the characteristic pattern described above can be selected as the traveling direction. The direction can be set with an arbitrary resolution.

  The traveling direction estimation apparatus according to the present invention includes low-pass means for applying a low-pass filter to the measurement data represented by the coordinate system, and the data acquisition means includes a low-pass filter by the low-pass means. It is preferable to acquire the measurement data to which is applied.

  The acceleration component or the angular velocity component may include a high frequency noise component. Examples of the high frequency noise component include a component caused by vibration generated when the foot is landed.

  According to the above configuration, the high-frequency noise component can be removed in advance by applying a low-pass filter to the converted data prior to component decomposition. Thereby, the accuracy of the estimation of the traveling direction can be improved.

  The traveling direction estimation device according to the present invention is preferably installed in a mobile terminal. In addition, since the traveling direction estimation device according to the present invention can accurately estimate the traveling direction even if the relative position with the person constantly changes, it can be preferably applied to a portable terminal used without being fixed to the human body. .

  Examples of portable terminals include mobile phones, smartphones, portable music players, and the like.

  The traveling direction estimation device may be realized by a computer, and in this case, a control program for causing the computer to operate the traveling direction estimation device by operating the computer as each of the above means and a computer recording the same A readable recording medium falls within the scope of the present invention.

  The traveling direction estimation apparatus according to the present invention includes a data acquisition unit that acquires measurement data that is measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor and is represented by a predetermined coordinate system, and the data acquisition unit acquires The measurement data is decomposed into a vertical direction component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, and a provisional lateral direction component orthogonal to the provisional traveling direction and the vertical direction. The component decomposition means and the period of acceleration or angular velocity in the traveling direction and lateral direction of the moving body during the moving operation for each of the patterns appearing in the provisional traveling direction component and the provisional lateral direction component obtained by the component decomposing means Pattern comparison means for comparing with a predetermined pattern representing the characteristics, and the result of comparison by the pattern comparison means When There are matches the above predetermined degree, the configuration and output means for outputting the provisional traveling direction, as the traveling direction estimated.

  The traveling direction estimation method according to the present invention includes a data acquisition step of acquiring measurement data measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor and represented by a predetermined coordinate system, and the data acquisition step. The acquired measurement data is divided into a vertical direction component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, and a provisional lateral direction component orthogonal to the provisional traveling direction and the vertical direction. The component decomposition step to be decomposed, and the acceleration and angular velocity in the traveling direction and the lateral direction of the moving body during the moving operation for each of the patterns appearing in the temporary traveling direction component and the temporary lateral direction component obtained by the component decomposing step A pattern comparison step for comparing with a predetermined pattern representing the periodicity of the pattern, and the pattern Compare the comparison in step a result, when the pattern meets the above predetermined order, the method comprising an output step of outputting the provisional traveling direction, as the traveling direction estimated.

  Therefore, even if the relative position between the moving body and the apparatus changes, the traveling direction can be accurately estimated from the moving operation of the moving body.

It is a functional block diagram which shows the structural example of the advancing direction estimation apparatus which concerns on one Embodiment of this invention. It is a figure shown about the structural example of the acceleration differential feature-value calculation part with which the said advancing direction estimation apparatus is provided. It is a graph which shows the example of the characteristic pattern of acceleration data, (a), (b), and (c) are the vertical direction, the advancing direction, and the side which can be observed when a person moves by walking motion, respectively. It represents the acceleration in the direction. It is a graph which shows the characteristic pattern of the acceleration differential data which differentiated acceleration, (a), (b), and (c) represent the feature-value of the acceleration differential of a perpendicular direction, a traveling direction, and a horizontal direction, respectively. ing. It is a graph which shows the power spectrum obtained by carrying out one-dimensional fast Fourier transform (FFT) of acceleration differential data. It is a graph which shows the example of a power spectrum in case a temporary advancing direction does not correspond with an actual advancing direction. It is a flowchart which illustrates about the flow of the pre-processing in the said acceleration differential feature-value calculation part. It is a flowchart which illustrates about the flow of the estimation process of the advancing direction in the said advancing direction estimation apparatus. It is a graph which shows the feature-value of the power spectrum obtained by applying FFT to acceleration data. It is a functional block diagram which shows the structural example of the advancing direction estimation apparatus which concerns on other embodiment of this invention. It is a graph which shows the example of the characteristic pattern of angular velocity data, (a), (b) and (c) have shown about the component of the yaw angular velocity, the roll angular velocity, and the pitch angular velocity, respectively. It is a graph which shows the power spectrum obtained by carrying out one-dimensional fast Fourier transform (FFT) of angular velocity data. It is a figure shown about the structural example of the angular velocity feature-value calculation part with which the said advancing direction estimation apparatus is provided. It is a figure shown about three vectors estimated (tracking) in order to estimate the movement azimuth of a pedestrian on a plane.

Embodiment 1
An embodiment of the present invention will be described with reference to FIGS. First, a schematic configuration of the traveling direction estimation apparatus 10 will be described with reference to FIG. FIG. 1 is a functional block diagram illustrating a configuration example of the traveling direction estimation apparatus 10.

  As illustrated in FIG. 1, the traveling direction estimation device 10 includes a measurement unit 50, a control unit 100, and a storage unit 200.

  The traveling direction estimation device 10 is a device that is attached or held to a moving body (a person or an object) and outputs a parameter indicating the traveling direction of the moving body. Specifically, the traveling direction estimation device 10 estimates the traveling direction of movement of the moving body using measurement data measured by the measuring unit 50 including a sensor that detects a movement state, posture change, and the like of the moving body. And output.

  More specifically, the measurement unit 50 includes an angular velocity sensor (three-dimensional angular velocity sensor) 51 and an acceleration sensor (three-dimensional acceleration sensor) 52.

As the angular velocity sensor 51, a triaxial angular velocity sensor is used. The angular velocity sensor 51 outputs the measured angular velocity as angular velocity data ω _S that is vector data. Note that the angular velocity data ω _S output from the angular velocity sensor 51 is represented by a sensor coordinate system. The angular velocity sensor 51 is illustratively a MEMS type. However, the present invention is not limited to this, and the angular velocity sensor 51 may be a crystal type sensor.

As the acceleration sensor 52, a triaxial acceleration sensor is used. The acceleration sensor 52 outputs an acceleration measured as the acceleration data A _S is a vector data. The acceleration data A _S of the acceleration sensor 52 is output is represented by the sensor coordinate system. The acceleration sensor 52 is illustratively a MEMS type.

  It is assumed that the angular velocity sensor 51 and the acceleration sensor 52 are sampling sufficiently fast. Hereinafter, the sampling interval of the output is represented by Δt. Moreover, the structure of the measurement part 50 is not restrict | limited to what was shown above. For example, the measurement unit 50 may be configured to further include a geomagnetic sensor.

Thereafter, collectively acceleration data A _S of the angular velocity data omega _S and the acceleration sensor 52 the angular velocity sensor 51 outputs the output, also referred to as measurement data.

  The control unit 100 performs overall control of the traveling direction estimation device 10 and performs control for estimating and outputting the traveling direction of the moving body using the measurement data output from the measuring unit 50.

  The storage unit 200 stores various data used by the traveling direction estimation device 10.

(About feature data)
Here, characteristic patterns of various data relating to acceleration obtained when a person moves by a walking motion will be described with reference to FIGS. 3, 4, and 5.

[Characteristic pattern of acceleration data]
FIG. 3 is a graph showing an example of a characteristic pattern of acceleration data. (A), (b), and (c) of the figure are graphs representing accelerations in the vertical direction, the traveling direction, and the lateral direction, respectively, that can be observed when a person moves by walking.

  The following characteristic pattern appears for the acceleration vector obtained by measuring the acceleration. First, a pattern having one step as one cycle appears in the vertical component shown in FIG. 3A, and one step is also set as one cycle in the component in the traveling direction shown in FIG. 3B. A pattern appears. And the pattern which makes 2 steps 1 period appears in the horizontal direction (left-right direction) shown in (c) of FIG.

  In this example, a high-frequency noise component included in the acceleration component is removed in advance using a low-pass filter.

[Characteristic pattern of acceleration differential data]
FIG. 4 is a graph showing a characteristic pattern of acceleration differential data obtained by differentiating acceleration. (A), (b), and (c) of the same figure show acceleration differential data obtained by time-differentiating the acceleration data shown in (a), (b), and (c) of FIG. 3, respectively. It is a represented graph.

  In the acceleration differential data obtained by time differentiation of the acceleration data, as shown in FIGS. 4A to 4C, a characteristic pattern having the same periodicity as the characteristic pattern of the acceleration data described above appears. . However, the graphs shown in (a) to (c) of FIG. 4 are 90 degrees out of phase with the graphs shown in (a) to (c) of FIG.

  In this example as well, a high-frequency noise component included in the acceleration differential component is removed in advance using a low-pass filter.

  In addition, since the differential amount of acceleration is amplified even if an error or noise component included in the acceleration data is small, it is preferable to apply a low-pass filter to the differential amount of acceleration.

[Acceleration differential power spectrum]
FIG. 5 is a graph showing a power spectrum obtained by performing one-dimensional fast Fourier transform (FFT) on acceleration differential data.

  As shown in FIG. 5, the following characteristic pattern appears in the power spectrum of acceleration differentiation. First, in the power spectrum in the vertical direction indicated by a dotted line, a frequency peak PK11 appears in the vicinity of 2 Hz. This indicates that 2 Hz is a basic frequency of walking motion. In addition, regarding the power spectrum in the traveling direction indicated by the solid line, a frequency peak PK12 appears in the vicinity of 2 Hz. Thus, a peak similar to the vertical direction appears in the traveling direction.

  In addition, regarding the power spectrum in the horizontal direction indicated by the broken line, a frequency peak PK13 appears in the vicinity of 1 Hz. In the horizontal direction, the frequency at which the peak appears is ½ of the frequency at which the peak appears in the vertical direction (or the traveling direction).

  Thus, a characteristic pattern appears in the power spectrum of acceleration differentiation when a person moves by walking motion.

  Therefore, in order to estimate the traveling direction, the provisional traveling direction is arbitrarily determined, and the peak of the frequency of the power spectrum of acceleration differentiation is examined to determine whether the characteristic pattern appears. In other words, a feature amount in the frequency domain of acceleration differentiation may be examined to determine whether it matches a characteristic pattern in walking motion.

  More specifically, it may be determined whether or not frequency peaks appear in the vicinity of 1 Hz and 2 Hz as described above.

  If the tentative travel direction matches the actual travel direction, the characteristic pattern appears. In other words, the axis may be determined so that the characteristic pattern appears on each axis in the vertical direction, the traveling direction, and the lateral direction orthogonal to the traveling direction.

  In the power spectrum in the left-right direction shown in FIG. 5, there is a mountain near 3 Hz, but this mountain may be ignored.

  On the other hand, such a characteristic pattern does not appear when the provisional traveling direction does not match the actual traveling direction. An example in which the provisional traveling direction does not match the actual traveling direction will be described with reference to FIG. FIG. 6 is a graph showing an example of a power spectrum when the provisional traveling direction does not match the actual traveling direction.

  As shown in FIG. 6, first, in the power spectrum in the vertical direction indicated by the dotted line, a frequency peak PK11 appears in the vicinity of 2 Hz as seen in the example of FIG.

  In the power spectrum in the tentative travel direction indicated by the solid line in FIG. 6, the peak PK122 having the highest frequency appears in the vicinity of 2 Hz, but the peak PK121 having a slightly higher frequency appears in the vicinity of 1 Hz.

  In addition, regarding the power spectrum in the temporary horizontal direction indicated by the broken line in FIG. 6, the peak PK131 having the largest frequency appears near 1 Hz, but the peak PK132 having a slightly larger frequency appears near 2 Hz.

  In this way, in the power spectrum in the provisional traveling direction and provisional lateral direction shown in FIG. 6, unlike FIG. 5, two peaks appear. This is because the tentative traveling direction does not match the actual traveling direction, so that the traveling direction peak and the lateral peak appear mixed.

  The power spectrum in the vertical direction is not so different between FIG. 5 and FIG. 6 because the vertical direction can be estimated with a considerably high accuracy by acceleration, angular velocity, and the like. Because it can be treated as a known one.

(Memory part)
The storage unit 200 includes a provisional value storage unit 201, a pattern storage unit 202, and a coincidence degree storage unit 203.

  The temporary value storage unit 201 stores a temporary value indicating an orientation indicating a temporary traveling direction. The provisional value is represented by a value indicating an orientation divided at a predetermined resolution. The azimuth is illustratively divided in increments of 0.5 degrees. However, the present invention is not limited to this, and the direction only needs to be divided at a resolution required for estimating the traveling direction with a desired accuracy.

  The pattern storage unit 202 stores a characteristic pattern of the peak frequency of the power spectrum obtained for the differential feature quantity of acceleration measured when a person moves by a walking motion. Specifically, the characteristic pattern of the peak frequency includes the vertical direction on the world coordinates in the human walking motion, the direction of movement of the person on the horizontal plane perpendicular to the vertical direction, and the horizontal direction orthogonal to the direction of movement on the horizontal plane. And is remembered for each.

  The pattern storage unit 202 stores, for example, a characteristic pattern of the peak frequency of the power spectrum as shown in FIG.

  What is necessary is just to memorize | store the characteristic pattern of this peak frequency obtained based on the acceleration measured in advance in the pattern memory | storage part 202. FIG.

  The coincidence degree storage unit 203 stores a provisional value in the traveling direction and a coincidence degree obtained by pattern comparison in association with each other.

(Control part)
The control unit 100 includes an acceleration differential feature amount calculation unit (data acquisition unit) 110, a vertical / horizontal plane decomposition unit (component decomposition unit) 130, a one-dimensional discrete Fourier transform unit (power spectrum acquisition unit) 140, a pattern comparison unit (pattern comparison). Means) 150, provisional value update unit 160, and estimated traveling direction output unit (output unit) 170.

The acceleration differential feature value calculation unit 110 calculates an acceleration differential feature value a _W from the measurement data output by the measurement unit 50. Here, the details of the acceleration differential feature quantity calculation unit 110 will be described with reference to FIG. FIG. 2 is a diagram illustrating a more detailed configuration example of the acceleration differential feature quantity calculation unit 110.

  As shown in FIG. 2, the acceleration differential feature quantity calculation unit 110 includes an attitude angle tracking unit 111, a coordinate system conversion unit 112, a time differentiation unit (time differentiation unit) 113, and a low-pass filter (low-pass unit) 114. It has.

The posture angle tracking unit 111 tracks posture angles. Attitude angle tracking unit 111, specifically, by using the angular velocity data omega _S outputted from the angular velocity sensor 51, the acceleration data A _S outputted from the acceleration sensor 52, and calculates the attitude angle. Then, the attitude angle tracking unit 111 generates and outputs a rotation matrix R that converts the coordinate system unique to the sensor into the world coordinate system based on the calculated attitude angle.

  Here, examples of the world coordinate system include an NED (North-East-Down) coordinate system. The coordinate system to be converted is not limited to the world coordinate system, and may be a coordinate system set with an arbitrary point as a reference. For example, the coordinate system unique to the sensor may be converted into a coordinate system based on the point where the measurement unit 50 starts measurement. In this case, conversion to a coordinate system including an axis in the gravity direction, an axis in the true north direction, and an axis orthogonal to these axes may be performed. As shown in FIG. 14, the true north direction can be detected by a geomagnetic sensor or the like.

The coordinate system converting unit 112 applies the rotation matrix R output from the attitude angle tracking unit 111 to the acceleration differential data A _S ′ output from the time differentiating unit 113 to convert the sensor-specific coordinate system into the world coordinate system. The coordinate system conversion unit 112 outputs acceleration differential data A _W ′ converted into the world coordinate system.

Time differential unit 113 outputs the acceleration differential data A _{S 'obtained} by differentiating the acceleration data A _S outputted from the acceleration sensor 52 times. Specifically, the time differentiating unit 113 can obtain acceleration differential data A _S ′ (jerk acceleration) by calculating a difference between acceleration data at time t and acceleration data at time (t + Δt). it can.

The low-pass filter 114 outputs an acceleration differential feature amount a _W ′ obtained by removing high-frequency components of the acceleration differential data A _W ′ converted into the world coordinate system. The low-pass filter 114 can be realized by, for example, a FIR (Finite Impulse Response) type filter.

  Returning to FIG. 1 again, the remaining parts will be described.

The vertical / horizontal plane decomposition unit 130 decomposes the acceleration differential feature quantity a _W ′ into a vertical direction component, a provisional traveling direction component, and a provisional lateral direction component. Specifically, the vertical / horizontal plane decomposition unit 130 performs the above-described component decomposition as follows.

First, the vertical / horizontal plane decomposition unit 130 calculates the gravity azimuth vector using the acceleration data _AS and angular velocity data ω _S output from the measurement unit 50. Then, the vertical / horizontal plane decomposition unit 130 determines a vertical direction axis using the gravity direction vector, and obtains a vertical direction component by decomposing the acceleration differential feature quantity a _W ′ according to the vertical direction axis.

In addition, the vertical / horizontal plane decomposition unit 130 acquires a temporary value from the temporary value storage unit 201, and further performs component decomposition on a horizontal plane orthogonal to the vertical direction of the acceleration differential feature value a _W ′ based on the acquired temporary value. Do.

In other words, the vertical / horizontal plane decomposition unit 130 decomposes the acceleration differential feature quantity a _W ′ into a temporary travel direction component indicated by a temporary value and a temporary lateral direction component orthogonal to the temporary travel direction component with respect to the horizontal plane.

In other words, the vertical / horizontal plane decomposition unit 130 determines a provisional traveling direction axis orthogonal to the vertical direction axis and a provisional lateral direction axis orthogonal to the vertical axis and the provisional traveling direction axis, and accelerates the acceleration. It can be said that the differential feature amount a _W ′ is mapped to the temporary travel direction axis and the temporary lateral direction axis.

In this way, the vertical / horizontal plane decomposition unit 130 determines the vertical direction axis (Z axis), the provisional traveling direction axis (X axis), and the provisional lateral direction axis (Y axis), and in accordance with each axis, acceleration differentiation The feature quantity a _W ′ is divided into a vertical component a _z ′, a provisional traveling direction component a _x ′, and a provisional transverse direction component a _y ′, and each is output.

  The one-dimensional discrete Fourier transform unit 140 applies FFT to each decomposed axis. More specifically, the one-dimensional discrete Fourier transform unit 140 includes a vertical axis processing unit 141, a horizontal plane X-axis processing unit 142, and a horizontal plane Y-axis processing unit 143.

The vertical axis processing unit 141, the horizontal plane X-axis processing unit 142, and the horizontal plane Y-axis processing unit 143 are respectively used for the vertical direction component a _z ′, the provisional traveling direction component a _x ′, and the provisional lateral direction component a _y ′. Apply FFT.

  Note that the length of the one-dimensional FFT time window in the one-dimensional discrete Fourier transform unit 140 is preferably within a range in which the walking motion is within about 4 or 5 steps, for example. Therefore, the length of the time window is preferably set to about 5.12 seconds, for example.

As a result of applying FFT, the vertical axis processing unit 141, the horizontal plane X-axis processing unit 142, and the horizontal plane Y-axis processing unit 143 respectively apply the vertical power spectrum P (a for the Z axis, the horizontal X axis, and the horizontal Y axis. _z ′), provisional traveling direction power spectrum P (a _x ′), and provisional lateral direction power spectrum P (a _y ′) are output.

  The pattern comparison unit 150 compares the power spectrum in each direction with a characteristic pattern having a predetermined peak frequency.

More specifically, the pattern comparison unit 150 displays the vertical power spectrum P (a _z ′), the provisional traveling direction power spectrum P (a _x ′), and the provisional lateral direction power spectrum P (a _y ′) Comparison is made with the characteristic pattern of the peak frequency of the power spectrum stored in the storage unit 202.

  Further, as a result of the comparison, the pattern comparison unit 150 calculates the degree of coincidence with the characteristic pattern of the peak frequency of the power spectrum for each direction component. Details of this calculation will be described later.

  Then, the pattern comparison unit 150 stores the calculated matching degree in the matching degree storage unit 203 in association with the provisional value in the traveling direction.

  The pattern comparison unit 150 performs the above comparison for 360 azimuths from 0 degrees to 0.5 degrees in increments of 0.5 degrees.

  The provisional value update unit 160 corrects the provisional value in the traveling direction stored in the provisional value storage unit 201 and performs control so as to estimate the traveling direction based on the provisional value after the correction. That is, when the comparison in the pattern comparison unit 150 ends, the temporary value update unit 160 adds 0.5 degrees to the current temporary value, and then instructs the vertical / horizontal plane decomposition unit 130 to perform component decomposition again.

  The provisional value updating unit 160 illustratively updates the provisional value from 0 degrees to 180 degrees in 0.5 degree increments. That is, the provisional value update unit 160 sets a provisional traveling direction in order for 360 directions.

  The estimated traveling direction output unit 170 refers to each azimuth stored in the coincidence degree storage unit 203 and the degree of coincidence when the comparison is completed for all the azimuths, and determines the direction with the highest degree of coincidence. Output as the estimated direction of travel.

(Process flow)
Hereinafter, the flow of processing executed by the traveling direction estimation device 10 will be described with reference to FIGS. 7 and 8.

  First, the flow of processing in the acceleration differential feature quantity calculation unit 110 will be described with reference to FIG. FIG. 7 is a flowchart illustrating the flow of preprocessing in the acceleration differential feature value calculation unit 110.

  As shown in FIG. 7, first, the acceleration differential feature quantity calculation unit 110 acquires sensor data (acceleration data) from the acceleration sensor (S11).

  Subsequently, the time differentiation unit 113 calculates the time differentiation of the sensor data (S12). Subsequently, the coordinate system conversion unit 112 converts the time differentiation of the sensor data into the world coordinate system (S13). Subsequently, the low pass filter 114 applies the low pass filter to the sensor data whose coordinates have been converted (S14).

  The acceleration differential feature quantity calculation unit 110 outputs the data after applying the low-pass filter thus obtained (S15).

  Next, the flow of the traveling direction estimation process in the traveling direction estimation apparatus 10 will be described with reference to FIG. FIG. 8 is a flowchart illustrating the flow of the traveling direction estimation process in the traveling direction estimation apparatus 10.

  As shown in FIG. 8, first, the vertical / horizontal plane decomposition unit 130 acquires an acceleration differential feature amount (S21). Subsequently, the vertical / horizontal plane decomposition unit 130 decomposes the acceleration differential feature quantity into a vertical axis and a horizontal plane (two axes) based on the provisional value in the traveling direction (S22).

  Subsequently, the one-dimensional discrete Fourier transform unit 140 applies a one-dimensional discrete Fourier transform to each axis component (S23).

  Subsequently, the pattern comparison unit 150 compares the peak frequency with the characteristic pattern of the walking motion and calculates the degree of match (S24).

  Here, when the provisional value has not been updated to the end (NO in S25), the provisional value update unit 160 updates the provisional value in the traveling direction (S26). On the other hand, when the provisional value has been updated to the end (YES in S25), the estimated traveling direction output unit 170 outputs the traveling direction in which the provisional value having the highest degree of matching is estimated (S27).

(Action / Effect)
As described above, the traveling direction estimation apparatus 10 estimates the traveling direction when the own apparatus moves with the walking motion of the person using the measurement data of the sensor held by the own apparatus. An acceleration differential feature value that is measured by the acceleration sensor 52, which is the sensor, is time-differentiated, converted to the world coordinate system, and outputs an acceleration differential feature value a _W ′ to which a low-pass filter is applied. The calculation unit 110 and the acceleration differential feature value a _W ′ are orthogonal to the vertical direction component in the world coordinate system, the provisional traveling direction component set in a direction orthogonal to the vertical direction, the provisional traveling direction, and the vertical direction. The vertical / horizontal plane decomposition unit 130 that decomposes components into the provisional transverse direction component and the patterns that appear in the provisional traveling direction component and the provisional transverse direction component are transferred during the walking motion. As a result of updating and comparing the provisional value with the pattern comparison unit 150 that compares with the characteristic pattern representing the periodicity of the acceleration or angular velocity in the traveling direction and the lateral direction of the body, the provisional value with the highest degree of matching is estimated. It is the structure provided with the estimated traveling direction output part 170 which outputs as a traveling direction.

  Even if the relative position of a person and the own apparatus changes by the said structure, there exists an effect that the advancing direction in the said world coordinate system can be estimated with a sufficient precision from a person's walking motion. Further, as shown in FIG. 14, by detecting the horizontal reference azimuth (true north direction) with a geomagnetic sensor or the like, the azimuth angle for the traveling azimuth with respect to the horizontal reference azimuth can be suitably obtained.

  Further, since the time differentiating unit 113 differentiates the acceleration data of the sensor coordinate system instead of the acceleration data of the world coordinate system, it is effective in reducing noise.

  Further, since the acceleration sensor 52 employs the MEMS type, the output has a level at which the influence of the zero G offset output cannot be ignored.

Here, the time differentiating unit 113 outputs acceleration differential data A _S ′ by calculating a difference between acceleration data at time t and acceleration data at time (t + Δt). Since Δt is a very short interval, the zero G offset output included in the output of the acceleration sensor 52 can be regarded as the same. Therefore, since the zero G offset output is canceled out in the acceleration differential data A _S ′ that is the difference data, the influence is reduced.

(Modification)
Below, the modification of the advancing direction estimation apparatus 10 is demonstrated.

[Omitted time differentiation]
In the above description, the acceleration differential feature quantity calculation unit 110 includes the time differentiation unit 113. However, the time differentiation unit 113 may be omitted.

That is, the acceleration data A _S of the acceleration sensor 52 is outputted to the coordinate transformation by the coordinate system converter 112 may further apply filtering by the low pass filter 114.

  Then, it is possible to adopt a configuration in which component decomposition by the vertical / horizontal plane decomposition unit 130 and FFT by the one-dimensional discrete Fourier transform unit 140 are performed on the acceleration feature value obtained in this way.

  FIG. 9 shows the characteristic amount of the power spectrum obtained by applying FFT to the acceleration data. As shown in FIG. 9, a characteristic pattern similar to that shown in FIG. 5 appears in the power spectrum of acceleration.

  More specifically, first, in the power spectrum in the vertical direction indicated by the dotted line, a frequency peak PK21 appears in the vicinity of 2 Hz. In addition, regarding the power spectrum in the traveling direction indicated by the solid line, a frequency peak PK22 appears in the vicinity of 2 Hz.

  And about the power spectrum about the horizontal direction shown with a broken line, the frequency peak PK23 appears in 1 Hz vicinity.

[About low-pass filter]
In the above description, the acceleration differential feature quantity calculation unit 110 includes the low-pass filter 114, but the low-pass filter 114 may be omitted.

  For example, such a configuration can be considered when the measurement accuracy of the acceleration sensor 52 is sufficient and it is not necessary to remove high-frequency components, or when the acceleration sensor 52 includes a low-pass filter. .

  Further, the low-pass filter 114 may be provided between the time differentiation unit 113 and the coordinate system conversion unit 112.

  Note that it is also possible to employ both the above-described modification of “omission of time differentiation unit” and the modification of “about low-pass filter” simultaneously.

[About provisional value update]
The provisional value update unit 160 is configured to set provisional values for the total 360 azimuths from 0 degree to 180 degrees in 0.5 degree increments, but is not limited thereto. For example, the initial value of the temporary value set by the temporary value update unit 160 can be arbitrarily set. Further, if the traveling direction estimation device 10 is mounted on a mobile phone and is held in a human hand, the traveling direction candidates can be roughly predicted. Therefore, the provisional value update unit 160 may update the provisional value for the candidate for the traveling direction predicted to some extent.

[Pattern comparison]
The pattern comparison unit 150 is configured to compare the power spectrum pattern of each axis after FFT with the characteristic pattern stored in the pattern storage unit 202. However, the present invention is not limited to this, and pattern determination may be performed as follows, for example. That is, the pattern comparison unit 150 may determine whether or not the ratio between the intensity value (Power) at 1 Hz and the intensity value at 2 Hz is equal to or greater than a predetermined threshold. Further, the pattern comparison unit 150 may determine whether or not the difference between the intensity value at 1 Hz and the intensity value at 2 Hz is equal to or greater than a predetermined threshold value.

[Embodiment 2]
Another embodiment of the present invention will be described with reference to FIGS. For convenience of explanation, members having the same functions as those in the drawings described in the first embodiment are given the same reference numerals, and descriptions thereof are omitted.

  A traveling direction estimation apparatus 10A shown in FIG. 10 estimates the traveling direction based on the angular velocity feature amount.

(About feature data)
Here, before describing the specific configuration of the traveling direction estimation apparatus 10A, the characteristic patterns of various data relating to acceleration obtained when a person moves by walking motion will be described with reference to FIGS. 11 and 12. It is as follows.

[Characteristic pattern of angular velocity data]
FIG. 11 is a graph showing an example of a characteristic pattern of angular velocity data. (A), (b), and (c) of FIG. 6 are respectively observed around a yaw axis (that is, a vertical axis) and a roll axis (that is, an axis in the traveling direction) that can be observed when a person moves by walking motion. ) And angular velocity around the pitch axis (ie, the lateral axis).

  The following characteristic pattern appears for the angular velocity vector obtained by measuring the angular velocity. First, a pattern with two steps as one cycle appears in the yaw angular velocity component shown in FIG. In addition, a pattern having two steps as one cycle appears in the component of the roll angular velocity shown in FIG. A pattern having one step as one cycle appears in the pitch angular velocity component shown in FIG.

  In this example, a high-frequency noise component included in the angular velocity component is removed in advance using a low-pass filter.

[Power spectrum of angular velocity]
FIG. 12 is a graph showing a power spectrum obtained by performing one-dimensional fast Fourier transform (FFT) on angular velocity data.

  As shown in FIG. 12, the following characteristic pattern appears in the power spectrum of angular velocity. First, in the power spectrum for the yaw angular velocity indicated by the dotted line, a frequency peak PK31 appears in the vicinity of 1 Hz.

  In addition, regarding the power spectrum for the roll angular velocity indicated by the solid line, a frequency peak PK32 appears in the vicinity of 1 Hz.

  Further, for the power spectrum for the pitch angular velocity indicated by the broken line, a frequency peak PK33 appears in the vicinity of 2 Hz.

  Thus, a characteristic pattern appears in the power spectrum of the angular velocity when a person moves by walking motion.

  Therefore, in order to estimate the traveling direction, a provisional traveling direction is arbitrarily determined, and the frequency peak of the angular velocity differential power spectrum is examined to determine whether the characteristic pattern appears. This is the same as the case of the power spectrum of acceleration differentiation described above.

(Configuration example)
Next, the configuration of the traveling direction estimation device 10A will be described with reference to FIG. FIG. 10 is a functional block diagram illustrating a configuration example of the traveling direction estimation apparatus 10A.

  Differences between the traveling direction estimation device 10 shown in FIG. 1 and the traveling direction estimation device 10A shown in FIG. 10 will be described as follows.

  The traveling direction estimation device 10A includes an acceleration differential feature quantity calculation unit 110, a vertical / horizontal plane decomposition unit 130, a one-dimensional discrete Fourier transform unit 140, and a pattern comparison unit 150 in the traveling direction estimation device 10 shown in FIG. , Angular velocity feature quantity calculation unit (data acquisition unit) 120, vertical / horizontal plane decomposition unit (component decomposition unit) 130A, one-dimensional discrete Fourier transform unit (power spectrum acquisition unit) 140A, and pattern comparison unit (pattern comparison unit) ) The configuration is changed to 150A.

(About differences)
The difference will be described as follows.

The angular velocity feature amount calculation unit 120 calculates the angular velocity feature amount ω _Wb from the measurement data output by the measurement unit 50. Here, the details of the angular velocity feature value calculation unit 120 will be described with reference to FIG. FIG. 13 is a diagram illustrating a more detailed configuration example of the angular velocity feature quantity calculation unit 120.

  As shown in FIG. 13, the angular velocity feature quantity calculation unit 120 changes the coordinate system conversion unit 112 to the coordinate system conversion unit 121 in the acceleration differential feature quantity calculation unit 110 shown in FIG. It has been removed.

The coordinate system conversion unit 121 applies the rotation matrix R output from the attitude angle tracking unit 111 to the angular velocity data ω _S output from the angular velocity sensor 51, and converts the coordinate system unique to the sensor into a world coordinate system. The coordinate system conversion unit 121 outputs angular velocity data ω _Wa converted into the world coordinate system.

The low-pass filter has been redesigned so as to output angular velocity feature value ω _{Wb from} which high-frequency components of angular velocity data ω _Wa have been removed.

  Returning to FIG. 10 again, the remaining parts will be described.

The vertical / horizontal plane decomposition unit 130A acquires a temporary value from the temporary value storage unit 201, sets the acquired temporary value as a temporary roll axis, and converts the angular velocity feature value ω _Wb into the yaw axis, temporary roll axis, and temporary pitch axis. Map.

Then, the vertical / horizontal plane decomposition unit 130A determines a yaw axis (Z axis), a temporary roll axis (X axis), and a temporary pitch axis (Y axis), and calculates an acceleration differential feature quantity a _W ′ according to each axis. , The yaw angular velocity component ω _z , the temporary roll angular velocity component ω _x , and the temporary pitch angular velocity component ω _y are each decomposed and output.

  The one-dimensional discrete Fourier transform unit 140A applies FFT to each decomposed axis. More specifically, the one-dimensional discrete Fourier transform unit 140A includes a yaw axis processing unit 141A, a roll axis processing unit 142A, and a pitch axis processing unit 143A.

The yaw axis processing unit 141A, the roll axis processing unit 142A, and the pitch axis processing unit 143A apply FFT to the yaw angular velocity component ω _z , the temporary roll angular velocity component ω _x , and the temporary pitch angular velocity component ω _y , respectively.

Then, as a result of applying the FFT, the yaw axis processing unit 141A, the roll axis processing unit 142A, and the pitch axis processing unit 143A respectively obtain the yaw angular velocity component power spectrum P () for the Z axis, the horizontal X axis, and the horizontal Y axis. ω _z ), provisional roll angular velocity component power spectrum P (ω _x ), and provisional pitch angular velocity component power spectrum P (ω _y ) are output.

  The pattern comparison unit 150A compares the power spectrum of each angular velocity component with the predetermined peak frequency characteristic pattern.

(Modification)
Hereinafter, a modified example of the traveling direction estimation apparatus 10A will be described.

[Addition of time derivative part]
In the angular velocity feature amount calculation unit 120, a time differentiation unit may be provided between the angular velocity sensor 51 and the coordinate system conversion unit 121.

  That is, as in the case of the acceleration described above, the temporal differentiation (angular acceleration) of the angular velocity may be used as the feature amount. In this case, it should be noted that the phase is shifted by 90 degrees with respect to the angular velocity data.

[Add output correction unit]
When a MEMS type or crystal type sensor is adopted as the angular velocity sensor 51, a null offset output is present in the output in the same manner as the output of the acceleration sensor 52. However, in the first place, in the posture angle estimation process in the PDR, in order to improve the accuracy of posture angle estimation, it is assumed that the offset of the angular velocity sensor is sufficiently corrected and removed in advance. That is, it is conceivable to provide an output correction unit (not shown) that corrects angular velocity data of the angular velocity sensor.

  For this reason, if the corrected data obtained by correcting the output of the angular velocity sensor by the output correcting unit is used, there is a possibility that the influence of the offset is not so much even if time differentiation is not performed. When the output correction unit is provided, a configuration in which a temperature sensor is built in or externally attached and the output correction unit estimates and corrects the null offset output based on the temperature is conceivable.

(Other configuration examples)
Finally, each block of the above-described traveling direction estimation devices 10 and 10A may be realized by hardware by a logic circuit formed on an integrated circuit (IC chip), or a CPU (Central Processing Unit) is used. It may be realized by software.

  In the latter case, each device includes a CPU that executes instructions of a program that realizes each function, a ROM (Read Only Memory) that stores the program, a RAM (Random Access Memory) that expands the program, the program, and various types A storage device (recording medium) such as a memory for storing data is provided. An object of the present invention is to provide a recording medium in which a program code (execution format program, intermediate code program, source program) of a control program for each of the above devices, which is software that realizes the above-described functions, is recorded in a computer-readable manner. This can also be achieved by supplying each of the above devices and reading and executing the program code recorded on the recording medium by the computer (or CPU or MPU).

  Examples of the recording medium include tapes such as magnetic tape and cassette tape, magnetic disks such as floppy (registered trademark) disks / hard disks, and CD-ROM / MO / MD / DVD / CD-R / Blu-ray disks (registered trademarks). ) And other optical disks, IC cards (including memory cards) / optical cards, semiconductor memories such as mask ROM / EPROM / EEPROM / flash ROM, PLD (Programmable Logic Device) and FPGA (FPGA) Logic circuits such as Field Programmable Gate Array can be used.

  Further, each of the above devices may be configured to be connectable to a communication network, and the program code may be supplied via the communication network. The communication network is not particularly limited as long as it can transmit the program code. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network, telephone line network, mobile communication network, satellite communication network, and the like can be used. The transmission medium constituting the communication network may be any medium that can transmit the program code, and is not limited to a specific configuration or type. For example, even with wired lines such as IEEE 1394, USB, power line carrier, cable TV line, telephone line, ADSL (Asymmetric Digital Subscriber Line) line, infrared rays such as IrDA and remote control, Bluetooth (registered trademark), IEEE 802.11 wireless, HDR ( It can also be used by radio such as High Data Rate (NFC), Near Field Communication (NFC), Digital Living Network Alliance (DLNA), mobile phone network, satellite line, and digital terrestrial network. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave in which the program code is embodied by electronic transmission.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  For example, the traveling direction estimation devices 10 and 10A may be mounted on a mobile terminal such as a mobile phone. Even in the case where the relative position between the person and the terminal may change constantly as in a portable terminal, the traveling direction can be accurately estimated, and the azimuth angle with respect to the traveling direction can be suitably obtained.

  In addition, it is possible to store time series data of the estimated traveling direction and take a history of movement of a person by a walking motion. For example, when applied to an ordering terminal such as GOT (Graphical Order Terminal), it is possible to grasp the movements and leads of employees working in a store by taking a history of estimated directions.

  According to the present invention, since the moving direction of a person who moves by walking can be estimated with high accuracy, the present invention can be suitably applied to a navigation device or the like that displays a person's current location on a map.

10, 10A Traveling direction estimation device 50 Measuring unit 51 Angular velocity sensor (three-dimensional angular velocity sensor)
52 Acceleration sensor (3D acceleration sensor)
100 Control Unit 110 Acceleration Differential Feature Value Calculation Unit (Data Acquisition Unit)
111 posture angle tracking unit 112 coordinate system conversion unit 113 time differentiation unit (time differentiation unit)
114 Low-pass filter (low-pass means)
120 Angular velocity feature quantity calculation unit (data acquisition means)
121 Coordinate system conversion unit 130, 130A Vertical / horizontal plane decomposition unit (component decomposition means)
140 One-dimensional discrete Fourier transform unit (power spectrum acquisition means)
141 Vertical axis processing unit 142 Horizontal plane X-axis processing unit 143 Horizontal plane Y-axis processing unit 140A One-dimensional discrete Fourier transform unit (power spectrum acquisition means)
141A Yaw axis processing unit 142A Roll axis processing unit 143A Pitch axis processing unit 150, 150A Pattern comparison unit (pattern comparison unit)
160 Temporary Value Update Unit 170 Estimated Travel Direction Output Unit (Output Unit)
200 Storage Unit 201 Temporary Value Storage Unit 202 Pattern Storage Unit 203 Matching Level Storage Unit

Claims (9)

In the traveling direction estimation device that estimates the traveling direction when the moving body moves by a constant movement operation using the measurement data of the sensor that measures the motion state of the moving body,
Data acquisition means for acquiring measurement data measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor, which is the sensor, and represented by a predetermined coordinate system;
The measurement data acquired by the data acquisition means includes a vertical component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, a provisional lateral direction orthogonal to the provisional traveling direction and the vertical direction. Component decomposing means for decomposing into directional components;
For each of the patterns appearing in the provisional traveling direction component and the provisional lateral direction component obtained by the component decomposing means, a predetermined direction representing the periodicity of the traveling direction and lateral acceleration or angular velocity of the moving body at the time of the moving operation. Pattern comparison means for comparing with a pattern;
As a result of the comparison by the pattern comparison means , when any of the patterns appearing in the provisional traveling direction component and the provisional lateral direction component matches a predetermined level or more, the provisional traveling direction is output as the estimated traveling direction. And a traveling direction estimation device. A time differentiating means for differentiating the measurement data represented by the coordinate system with respect to time;
2. The traveling direction estimation apparatus according to claim 1, wherein the data acquisition unit acquires the measurement data time-differentiated by the time differentiation unit. For the provisional traveling direction component and the provisional transverse direction component obtained by the component decomposing means, comprising a power spectrum acquisition means for acquiring a power spectrum,
The pattern comparison means compares the power spectrum acquired by the power spectrum acquisition means with a power spectrum pattern representing periodicity of acceleration or angular velocity in the traveling direction and the lateral direction of the moving object during the moving operation. The traveling direction estimation apparatus according to claim 1 or 2, wherein The component decomposition means performs the component decomposition in the provisional travel direction set for each direction,
The pattern comparison means compares the provisional traveling direction component and provisional transverse direction component of each direction obtained by the component decomposition means with the predetermined pattern,
The output means outputs, as an estimated traveling direction, the provisional traveling direction that best matches the predetermined pattern among the provisional traveling directions in the respective directions from the result of comparison by the pattern comparison means. The traveling direction estimation device according to any one of claims 1 to 3. Comprising low-pass means for applying a low-pass filter to the measurement data represented by the coordinate system,
5. The traveling direction estimation apparatus according to claim 1, wherein the data acquisition unit acquires the measurement data to which a low-pass filter is applied by the low-pass unit.   A portable terminal provided with the advancing direction estimation apparatus of any one of Claim 1 to 5.   A control program for operating the traveling direction estimation apparatus according to any one of claims 1 to 5, wherein the control program causes a computer to function as each of the means.   A computer-readable recording medium on which the control program according to claim 7 is recorded. In the traveling direction estimation method for estimating the traveling direction when the moving body moves by a constant moving motion using the measurement data of the sensor that measures the motion state of the moving body,
A data acquisition step of acquiring measurement data measured by a three-dimensional acceleration sensor or a three-dimensional angular velocity sensor, which is the sensor, and represented by a predetermined coordinate system;
The measurement data acquired in the data acquisition step includes a vertical component in the coordinate system, a provisional traveling direction component set in a direction orthogonal to the vertical direction, a provisional lateral direction orthogonal to the provisional traveling direction and the vertical direction. A component decomposition step that decomposes the component into direction components;
For each of the patterns appearing in the provisional traveling direction component and the provisional lateral direction component obtained by the component decomposition step, a predetermined direction that represents the periodicity of the traveling direction and lateral acceleration or angular velocity of the moving body during the moving operation. A pattern comparison step for comparing with the pattern;
As a result of the comparison in the pattern comparison step , when any of the patterns appearing in the provisional traveling direction component and the provisional lateral direction component matches a predetermined level or more, the provisional traveling direction is output as the estimated traveling direction. A traveling direction estimation method comprising: an output step.

Images