JP5256846B2 - Posture specifying device, moving direction specifying device, position specifying device, computer program, and posture specifying method - Google Patents

Description

  The present invention relates to position specifying technology, and in particular, includes a posture specifying device that can specify its posture even when a pedestrian carries it, a moving direction specifying device including the posture specifying device, and a moving direction specifying device. The present invention relates to a position specifying device, a computer program for realizing the posture specifying device, and a posture specifying method.

  Examples of position detection methods widely used in navigation for detecting the position of a moving body such as a vehicle include self-contained navigation, satellite navigation, map matching, and hybrid navigation. Self-contained navigation uses a distance sensor, an azimuth sensor, an angular velocity sensor, etc., for example, based on the azimuth angle of the vehicle traveling with respect to an orthogonal coordinate system based on the longitude-latitude coordinate system and the traveling distance per unit time. Although it is calculated, consistency with the road is not taken into consideration, and there is a problem that errors in the vehicle position accumulate as the travel distance increases.

  Satellite navigation uses GPS (Global Positioning System), and the detected position includes an error of about 10 to 20 m. Since GPS is used, an in-vehicle sensor such as a distance sensor, an azimuth sensor, or an angular velocity sensor is unnecessary. However, on roads under elevated roads, roads between buildings, mountain roads, roadside trees, etc., radio waves cannot be received from a predetermined number of GPS satellites, and the detection accuracy is greatly degraded. is there. In addition, there is also a problem that the traveling road is wrong in a narrow street with a narrow road interval.

  In addition, the map matching method detects the position of the vehicle in consideration of the consistency (matching) between the travel locus by the self-contained navigation and the road map (see Patent Document 1). That is, the most probable road is selected from a plurality of road candidates considered to be traveling while comparing the trajectory obtained by the self-contained navigation with the road map data. Then, when the number of candidate roads is limited to one, the traveling locus of the vehicle obtained by the self-contained navigation is matched with the road. However, if the limited road is wrong, there is a problem that position detection after that becomes impossible.

  Hybrid navigation is a combination of satellite navigation and map matching, and it rationally estimates the position of the vehicle and identifies the road on which it is traveling, taking into account the errors between autonomous navigation and satellite navigation. (See Patent Document 2). In hybrid navigation, for example, the position of a vehicle is detected using a map matching method in normal times. When the vehicle position cannot be detected by the map matching method, the vehicle position and direction are detected by satellite navigation to estimate the vehicle position, and the vehicle position is detected in consideration of consistency with the road map data. To do. With hybrid navigation, except for special cases, there is almost no possibility of mistaken roads on which vehicles are traveling, and the positional accuracy in the direction of the road is within an error range of about 10 m on average. In the target navigation, the accuracy level is almost no problem in practical use.

  On the other hand, in portable devices such as a mobile phone or a simple navigation device carried by a pedestrian, a method for detecting the current position of the pedestrian using communication with a GPS satellite or a base station has been put into practical use. When detecting the position by receiving radio waves from GPS satellites, if the GPS satellite reception is good, the position error is about 10-20m, but in the valleys of buildings such as buildings in the city center or roads under elevated In some cases, it becomes impossible to detect the position, or the position error becomes several hundred meters due to the influence of multipath or the like, and an accurate position cannot be obtained. In particular, in the case of pedestrians, unlike mobile objects such as vehicles, there is a tendency to walk near buildings along the buildings, so the reception level of radio waves from GPS satellites decreases.

  In addition, there is a conventional portable device that displays the position of the pedestrian on the map, but it simply displays the position of the pedestrian on the road on the map close to the position measured by the radio wave from the GPS satellite, No consideration is given to the walking trajectory of the pedestrian, the area where the pedestrian can walk other than the road, and the connectivity between the walking area and the road. For this reason, when walking on narrow streets with narrow road intervals, or when the radio wave reception level of GPS satellites deteriorates, the position of the pedestrian frequently flies over the road on the map over time. This happens.

  Also, unlike the case of vehicles, pedestrians frequently enter indoors as well as outdoors. When pedestrians enter buildings, they can receive almost all radio waves from GPS satellites. In such a case, if the communication with the base station is impossible, the position cannot be detected at all. Even if communication with the base station is possible, the accuracy of position detection is greatly reduced.

  Conventionally, in position detection using a portable device, if a general position is known, a pedestrian can also determine the final position, so that it can be used even if there is some error. However, in recent years, the necessity of crime prevention for mobile device owners or the prevention of traffic accidents with vehicles has increased, and more accurate position detection has become necessary from the viewpoint of safety or security. There is a demand for improvement of position accuracy in narrow streets with narrow road intervals, identification of buildings, underpasses, and the like.

  Under such circumstances, in position detection using a portable device, not only position detection by a GPS satellite but also one using a self-contained navigation method or one using a map matching method has been proposed.

For example, when GPS data is corrected or GPS data cannot be acquired using a device that calculates a distance from the pedestrian's stride (see Patent Documents 3 and 4), self-contained navigation using a geomagnetic orientation sensor, an angular velocity sensor, and an acceleration sensor A method using self-contained navigation (see Patent Document 5), storing the geomagnetic vertical component of the measurement point, calculating the tilt angle with respect to the horizontal plane from the data detected by the 3-axis geomagnetic sensor, and based on this Is disclosed (see Patent Document 6).
JP-A-63-148115 JP-A-2-275310 JP-A-8-68643 Japanese Patent Laid-Open No. 9-89584 JP 2004-233058 A JP-A-63-27711

  However, the conventional self-contained navigation by the portable device is based on the premise that the attitude of the sensor built in the portable device is unchanged, or the detection axis of the sensor is horizontal to the pedestrian coordinate system (for example, horizontal to the moving direction of the pedestrian). Assuming that the sensor posture satisfies a limited condition when it is rotated only around one of the x-axis, the pedestrian movement direction is the y-axis, and the vertical direction is the z-axis). It was necessary.

  In the case of a device installed in a vehicle, since the relative posture between the vehicle and the device is unchanged, the posture viewed from the vehicle coordinate system is unchanged. In addition, since the vibration of the vehicle is relatively minute, as long as the vehicle travels linearly on a horizontal plane, the posture variation seen from the road coordinate system is also minute.

  However, in the case of a pedestrian, it is possible to limit the posture of the portable device to some extent only when the portable device is held in hand. Is not always held in the hand, and the attitude of the mobile device is uncertain. In addition, when a pedestrian walks, the mobile device (including the built-in sensor) also vibrates as the pedestrian walks. Therefore, even when viewed from the pedestrian, the posture of the mobile device constantly vibrates and changes. . Therefore, in order to detect the position of a pedestrian using self-contained navigation, it is necessary to always specify the posture of the portable device carried by the pedestrian.

  The present invention has been made in view of such circumstances, and a posture specifying device that can specify its own posture even when a pedestrian carries it, a moving direction specifying device including the posture specifying device, and the movement It is an object of the present invention to provide a position specifying device provided with an azimuth specifying device, a computer program for realizing the posture specifying device, and a posture specifying method.

An attitude specifying device according to a first aspect of the invention is an attitude specifying device that specifies its own attitude with respect to a coordinate axis of a reference orthogonal three-dimensional coordinate, and has an axis that forms an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinate. Conversion means for performing coordinate conversion between orthogonal three-dimensional coordinates and the reference orthogonal three-dimensional coordinates, a three-axis acceleration sensor, coordinate conversion performed by the conversion means, and the apparatus orthogonal three-dimensional detected by the three-axis acceleration sensor An angle calculation unit that calculates the angle using an acceleration component at coordinates, a storage unit that stores information indicating the relationship between the acceleration components of each axis in the reference orthogonal three-dimensional coordinates, and the three-axis acceleration sensor and a variable calculating means for calculating the time or the acceleration variable in accordance with the sum of an appropriate number of times detected acceleration, the angle calculation means, said calculated in variable calculation means to the coordinate transformation performed by the conversion means Yes as using information stored in the acceleration variables and the storage means, characterized in that is arranged to identify said angle calculating means its orientation by the calculated angle.

Posture identifying apparatus according to the second invention, Oite the first shot bright, the angle calculating means is configured to use the gravitational acceleration at the reference orthogonal three-dimensional coordinate in the coordinate conversion performed by the conversion means It is characterized by that.

Posture identifying apparatus according to the third invention, Oite the first invention or the second shot bright, it is possible mobile by pedestrians, according to the change of the walking characteristics of the posture change or pedestrians when carrying its own It is characterized by comprising determination means for determining whether or not to identify its own posture.

A posture specifying apparatus according to a fourth aspect of the present invention includes, in the third aspect of the invention, an acceleration comparison unit that compares an absolute value of the acceleration detected by the three-axis acceleration sensor with a predetermined acceleration threshold value, and the determination unit includes the absolute value. Is greater than the acceleration threshold value, it is determined that it is determined not to specify its own posture.

A posture specifying apparatus according to a fifth aspect of the present invention includes, in the third aspect of the invention, a detection unit that detects a change in the walking cycle, the number of steps, or the walking speed based on a change in acceleration detected by the three-axis acceleration sensor. When the change is detected by the detection means, it is determined that the posture of the device is not specified.

A posture specifying apparatus according to a sixth aspect of the invention is characterized in that, in the fifth aspect of the invention, the determination unit determines that the posture of the determination is not specified when the stop is detected by the detection unit. To do.

The posture specifying device according to a seventh aspect of the present invention comprises, in the third aspect , an azimuth sensor, and an azimuth angle comparison unit that compares the azimuth angle detected by the azimuth sensor with a predetermined azimuth angle threshold value. When the azimuth angle is larger than the azimuth angle threshold, it is determined that the posture is not specified.

In any one of the first to seventh inventions, the posture specifying apparatus according to the eighth invention is an evaluation unit that evaluates the reliability of the angle calculated by the angle calculation unit using coordinate conversion performed by the conversion unit. And an invalidating unit that invalidates the calculated angle when the evaluation unit evaluates that the reliability is low.

In any one of the first to seventh inventions, the posture specifying device according to the ninth invention comprises a storage means for storing the angles calculated by the angle calculation means over a plurality of time points, and the angle calculation means. An angle difference calculating means for calculating an angle difference between the calculated angle and an angle stored in the storage means, and an angle calculated by the angle calculating means when the angle difference calculated by the angle difference calculating means is larger than a predetermined threshold value. And an invalidating means for invalidating.

A moving azimuth specifying device according to a tenth aspect of the invention is the three-axis geomagnetic sensor for coordinate conversion performed by the posture specifying device according to any one of the first to ninth inventions, a three-axis geomagnetic sensor, and the converting means. Using the detected geomagnetic component at the apparatus orthogonal three-dimensional coordinates and the angle calculated by the angle calculation means, the geomagnetism of each axis on the reference plane specified by the two axes of the reference orthogonal three-dimensional coordinates is calculated. It is characterized by comprising: a calculating means; and an azimuth specifying means for specifying its own moving azimuth by the geomagnetic component calculated by the geomagnetism calculating means.

A position specifying device according to an eleventh aspect of the invention is a moving direction specifying device according to the tenth aspect of the invention, a distance detecting means for detecting a moving distance, a moving distance detected by the distance detecting means, and a movement specified by the moving direction specifying device. And a position specifying means for specifying its own position based on the azimuth.

A computer program according to a twelfth aspect of the present invention is a computer program for causing a computer to function as means for specifying the posture of the apparatus with respect to the coordinate axes of the reference orthogonal three-dimensional coordinate, And an apparatus having an axis that forms an arbitrary angle and a conversion means for performing coordinate conversion between the orthogonal 3D coordinates of the apparatus and the reference orthogonal 3D coordinates, and the apparatus detected by the coordinate conversion performed by the conversion means and a 3-axis acceleration sensor An angle calculation means for calculating the angle using an acceleration component in orthogonal three-dimensional coordinates, and a variable calculation for calculating an acceleration variable according to an added value of acceleration detected for an appropriate time or an appropriate number of times by the 3-axis acceleration sensor. And the angle calculation means performs the acceleration variable calculated by the variable calculation means for the coordinate conversion performed by the conversion means, and the Yes as using information indicating a relationship between the acceleration components of each axis in the quasi-orthogonal three-dimensional coordinates, further characterized in that to function as a specific means for identifying its position by the calculated angle.

A posture specifying method according to a thirteenth aspect of the present invention is a posture specifying method for specifying a posture with a posture specifying device for specifying its own posture with respect to a coordinate axis of a reference orthogonal three-dimensional coordinate, wherein the posture specifying device includes the reference orthogonal The apparatus orthogonality detected by a three-axis acceleration sensor for coordinate conversion between the apparatus orthogonal three-dimensional coordinates having an axis that forms an arbitrary angle with each axis of the three-dimensional coordinates and the reference orthogonal three-dimensional coordinates. The angle is calculated using an acceleration component in three-dimensional coordinates, an acceleration variable corresponding to an added value of acceleration detected for an appropriate time or an appropriate number of times by the 3-axis acceleration sensor is calculated, and the calculated acceleration variable and the Information indicating the relationship between the acceleration components of the respective axes in the reference orthogonal three-dimensional coordinates is used for the coordinate conversion, and the posture of the apparatus is specified by the calculated angle.

In the first invention, the twelfth invention and the thirteenth invention, the reference orthogonal three-dimensional coordinate is (x, y, z). The reference orthogonal three-dimensional coordinates (x, y, z) have, for example, a reference surface specified by two axes (x axis, y axis) of three coordinate axes. The reference plane can be, for example, a plane parallel to the plane of the road or passage on which the pedestrian walks. Also, the y-axis direction can be the pedestrian walking direction, the x-axis can be the right lateral direction with respect to the pedestrian walking direction, and the z-axis can be the vertical direction. The apparatus orthogonal three-dimensional coordinates with respect to the reference orthogonal three-dimensional coordinates (x, y, z), where the apparatus orthogonal three-dimensional coordinates having an axis that makes an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinates are (X, Y, Z) The rotation angles of (X, Y, Z) are α (pitch angle), β (roll angle), and γ (yaw angle), respectively, and the direction in which the right screw advances is positive. In this case, between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z), a predetermined conversion formula (conversion) using the rotation angles α, β, γ as parameters. Means) is established.

  The angle (α, β) is calculated using this conversion formula and the acceleration component at the apparatus orthogonal three-dimensional coordinates (X, Y, Z) detected by the three-axis acceleration sensor. In this case, each axis of the triaxial acceleration sensor can be an X axis, a Y axis, and a Z axis. Further, if the yaw angle γ is 0 or known in advance, the posture specifying device can specify its own posture based on the calculated angle. That is, the orientation of the subject is specified by obtaining the rotation angle with respect to the reference orthogonal three-dimensional coordinates (x, y, z) serving as a reference. By using a three-axis acceleration sensor, even when a pedestrian carries it, his / her posture can be specified.

In addition , information indicating the relationship between the acceleration components of the respective axes at the reference orthogonal three-dimensional coordinates (x, y, z) is stored. For example, the acceleration components of the x-axis (right lateral direction of the pedestrian), the y-axis (walking direction), and the z-axis (vertical direction) are v1, v2, and v3, respectively. In addition, the magnitude of each acceleration component is the largest in the vertical direction, and the acceleration component in the walking direction and the lateral direction of the pedestrian is proportional to the acceleration component in the vertical direction. As information indicating the relationship between the acceleration components, for example, v1 = s1 · v3 and v2 = s2 · v3 can be expressed. Note that s1 and s2 are appropriate proportional coefficients. An acceleration variable corresponding to the added value of the acceleration detected for an appropriate time or an appropriate number of times by the three-axis acceleration sensor is calculated. The calculation of the acceleration variable in the apparatus orthogonal three-dimensional coordinates (X, Y, Z) is performed by, for example, detecting the maximum value or the minimum value (negative maximum value) of the acceleration detected by the three-axis acceleration sensor. It can be determined over time or by integrating or averaging over an appropriate number.

  Thereby, it is possible to perform conversion using data obtained by averaging accelerations in both the reference orthogonal three-dimensional coordinates (x, y, z) and the apparatus orthogonal three-dimensional coordinates (X, Y, Z). That is, the acceleration variable of each axis of the apparatus orthogonal three-dimensional coordinates (X, Y, Z) can be mapped to the acceleration component of each axis of the orthogonal three-dimensional coordinates (x, y, z). Calculated acceleration variables and stored accelerations when using a predetermined conversion formula (conversion means) between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z) By using the component, the yaw angle γ is calculated. Thereby, even when the yaw angle γ is indeterminate, even when a pedestrian carries it, his / her posture can be specified.

In the second invention, when using a predetermined conversion formula (conversion means) between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z), Gravitational acceleration g at reference orthogonal three-dimensional coordinates (x, y, z) is used. In the gravitational acceleration g, the x-axis component and the y-axis component are 0, and the z-axis component is g. Thereby, even when a pedestrian carries, own posture can be specified.

In the third invention, whether or not to identify one's posture in accordance with a change in own posture when carrying or a change in walking characteristics of a pedestrian (for example, whether or not walking is stopped, a change in walking direction, etc.) Determine. If there is no change in the walking characteristics of the pedestrian and the pedestrian walks at a constant rhythm, and the posture of the mobile device (posture specifying device) does not change, the data detected by the three-axis acceleration sensor also has periodicity associated with walking. There are no significant fluctuations in the maximum value or minimum value of. However, if there is a change in the posture of the person or the walking characteristics of the pedestrian, the data detected by the three-axis acceleration sensor also includes unexpected fluctuations. In such a case, the posture of the person is not specified. Like that. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

In the fourth invention, the absolute value of the acceleration detected by the three-axis acceleration sensor is compared with a predetermined acceleration threshold value, and if the detected absolute value of the acceleration is larger than the acceleration threshold value, it is determined that the posture is not specified. To do. For example, when a portable device (posture specifying device) held in a pocket or bag is taken out, when a portable device is stored in a pocket or bag, etc., when a hand holding the portable device is changed, or in a bag containing a portable device When the posture changes, the posture of the mobile device also changes, and at least temporarily, the absolute value of the acceleration detected by the three-axis acceleration sensor changes greatly. Even in such a state, if the posture is specified using acceleration data, the correct posture cannot be obtained. Therefore, if the detected absolute value of acceleration is larger than the acceleration threshold, it is determined that the posture is not specified. To do. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

In the fifth invention, when the posture of the portable device (posture specifying device) does not change, the data detected by the triaxial acceleration sensor also has periodicity accompanying walking, and the walking cycle, the number of steps, the walking speed, etc. There is no big fluctuation and it is almost constant. However, when the posture of the mobile device changes, the walking cycle, the number of steps, the walking speed, etc. also change. When a change in the walking cycle, the number of steps, or the walking speed is detected based on a change in the acceleration detected by the three-axis acceleration sensor, it is determined that the posture is not specified. As a result, when the posture of the mobile device changes, it is possible not to specify its own posture, but to prevent the specification of an incorrect posture and improve the accuracy.

In the sixth aspect of the invention, when the stop of walking is detected, it is determined that its own posture is not specified. When a pedestrian stops walking, for example, it may be possible to take out the mobile device (posture specifying device) from a pocket or a bag, and the posture of the mobile device is likely to change. By not specifying the posture of itself, it is possible to prevent the specification of an incorrect posture and improve the accuracy.

In the seventh invention, an azimuth angle detected by an azimuth sensor (for example, a geomagnetic sensor) is compared with a predetermined azimuth angle threshold value, and if the azimuth angle is larger than the azimuth angle threshold value, it is determined that its own posture is not specified. To do. For example, if the pedestrian has greatly changed the walking azimuth, when the posture is specified using acceleration data, the correct posture may not be obtained, so if the detected azimuth is greater than the azimuth threshold, It is determined that its own posture is not specified. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

In the eighth invention, the reliability of the calculated angle is expressed by a conversion formula (conversion means) between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z). ) And the calculated angle is invalidated if the reliability is low. For example, it can be determined whether or not the acceleration data detected by the three-axis acceleration sensor satisfies the above-described conversion formula, and if it is not satisfied, it can be determined that the reliability is low. Note that invalidation includes, for example, positively rejecting the calculated angle or not using the calculated angle without rejecting it. As a result, even if the angle for specifying the posture can be obtained, if the reliability of the calculation result is low, the calculation result is invalidated, and the accuracy is improved by preventing the wrong posture from being specified. Can do.

In the ninth invention, the angles calculated over a plurality of time points are stored. Thereby, data calculated in the past is collected. An angle difference between the newly calculated angle and the stored angle is calculated, and if the calculated angle difference is greater than a predetermined threshold, the calculated angle is invalidated. For example, when the calculated data is larger than the previously calculated data, it is determined that the acceleration data detected by the triaxial acceleration sensor includes some abnormality, and the calculated result is reliable. Otherwise, the calculation result is invalidated. Note that invalidation includes, for example, positively rejecting the calculated angle or not using the calculated angle without rejecting it. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

In the tenth invention, the calculated pitch angle α, roll angle β, yaw angle γ, geomagnetism of each axis of the apparatus orthogonal three-dimensional coordinates (X, Y, Z) detected by the triaxial geomagnetic sensor, and reference orthogonality Using geomagnetism (eg, geomagnetism of the z-axis component) at three-dimensional coordinates (x, y, z), geomagnetism of two axes (eg, x-axis and y-axis) orthogonal to each other on the reference plane (eg, horizontal plane) The component is calculated, and its own moving direction is specified by the calculated geomagnetic component. For example, the x-axis component and the y-axis component of geomagnetism are mx and my, respectively. If the azimuth angle between the walking azimuth of the pedestrian (the direction of the y-axis) and the true east is θ (counterclockwise is positive), the azimuth angle θ can be obtained by tan θ = my / mx. When θ is 0, the walking direction is true east, and when θ is π / 2, the walking direction is true north. By using the three-axis acceleration sensor and the three-axis geomagnetic sensor, the walking direction of the pedestrian can be obtained.

In the eleventh aspect of the invention, the own position is specified by detecting the moving distance and specifying the moving direction. For example, a distance sensor can be used to detect the movement distance. Moreover, a triaxial acceleration sensor and a triaxial geomagnetic sensor can be used for specifying the moving direction. Thereby, even if the reception level of the radio wave from the GPS satellite is insufficient or cannot be received, the position of the pedestrian can be specified by the self-contained navigation or the map matching method.

  In the present invention, even when a pedestrian carries it, his / her posture can be specified.

  Hereinafter, the present invention will be described with reference to the drawings illustrating embodiments. FIG. 1 is a block diagram showing an example of the configuration of a position detecting device 100 as a position specifying device according to the present invention. The position detection device 100 is portable for pedestrians (including pedestrians traveling on bicycles) and has a function of specifying its own posture and a function of detecting a pedestrian's moving direction (walking direction) as will be described later. And a function of detecting the position of a pedestrian.

  The position detection device 100 includes a control unit 10 that controls the entire device, a communication unit 11, a positioning unit 20, a GPS 12, a map database 13, a posture specifying unit 30 as a posture specifying device, an orientation specifying unit 14, a storage unit 15, and an operation unit. 16, an output unit 17, a position detection processing unit 40, and the like. In addition, the moving direction specifying device can be configured by the posture specifying unit 30, the direction specifying unit 14, and the like. In addition, it can also be set as the structure which is not equipped with GPS12. Even in this case, the position of the pedestrian can be detected by the self-contained navigation or the map matching method.

  The positioning unit 20 includes a distance sensor 21, a triaxial acceleration sensor 22, a triaxial geomagnetic sensor 23, an orientation correction unit 24, a sensor calibration unit 25, and the like. The posture specifying unit 30 includes a coordinate conversion unit 31, an angle calculation unit 32, a determination unit 33, and the like. The position detection processing unit 40 includes a position estimation unit 41, an error calculation unit 42, an attribute determination unit 43, a walking behavior determination unit 44, a position update unit 45, a reliability calculation unit 46, an evaluation unit 47, and the like.

  The communication unit 11 is an optical beacon, an electric wave beacon, an RFID or a DSRC, a narrow-area communication function for communicating with a road device, a mid-range communication function such as a wireless LAN such as a UHF band or a VHF band, or a mobile phone Wide-area communication functions such as PHS, multiplex FM broadcasting or Internet communication. The communication unit 11 uses, for example, mid-range communication such as wireless LAN in the vicinity of an intersection, and is communicated by road-to-road communication between road devices, road-to-vehicle communication between road devices and vehicles, or vehicle-to-vehicle communication. Map information or intersection signal information. Here, examples of the road device include a traffic information collection device such as an ultrasonic sensor, an optical beacon, or an image sensor, an information board device that provides traffic information in characters or figures, a signal control device, and the like. Moreover, the communication part 11 can also acquire the signal information or map information of the intersection around a pedestrian from center apparatuses, such as an information processing center or a traffic control center, using wide communication, such as a mobile telephone.

  The communication unit 11 has a communication function for communicating with the base station, receives radio waves from a plurality of base stations, and outputs reception results to the positioning unit 20. In addition, the communication unit 11 outputs the position information of the communication point obtained by the narrow area communication with the road device to the positioning unit 20.

  The positioning unit 20 measures the position of the pedestrian from time to time (for example, every time 0.5 seconds, 1 second, etc., every 1 m, 2 m, etc.) (determines the positioning position) The moving distance and moving direction (positioning direction) are stored in the storage unit 15 together with the time as a positioning locus.

  The distance sensor 21 includes an acceleration sensor that can detect a speed in a very short time, a moving distance, a step sensor that can detect a relatively long moving distance, and the like. Here, for example, if an acceleration sensor is used as the step sensor, steep data generated in accordance with the walking pitch can be obtained, and the number of steps and the walking speed can be obtained by counting this number. Also, in this case, when riding a bicycle and stroking the pedal, or when going up and down the stairs of a pedestrian bridge or an underground crossing passage, the characteristics of steep data, for example, peak values (walking strength) are different. Therefore, it is possible to specify a walking place to some extent. Thereby, the position of a pedestrian can be detected for each short-distance and short-distance walk in self-contained navigation. If the GPS satellite positioning accuracy is very good and there are no buildings or the like outside the metropolitan area, the walking distance is calculated based on the difference in GPS positioning without using the step sensor. Also good. In the present application, the walking speed is used as a concept including a walking pitch (the number of steps per unit time).

  The triaxial acceleration sensor 22 detects acceleration in directions of three axes orthogonal to each other (for example, the X axis, the Y axis, and the Z axis), and outputs the detected acceleration to the posture specifying unit 30.

  The triaxial geomagnetic sensor 23 detects the geomagnetism in the directions of three axes orthogonal to each other (for example, the X axis, the Y axis, and the Z axis), and outputs the detected geomagnetism to the posture specifying unit 30.

  The azimuth correction unit 24 identifies the road on which the pedestrian walks by map matching, determines the estimated position of the pedestrian, and changes the positioning azimuth (movement azimuth) over a predetermined movement distance (for example, 20 m). If not, the positioning direction is corrected to the direction of the road on the map. Details of the azimuth correction will be described later.

  The sensor calibration unit 25 calibrates the distance sensor 21 (acceleration sensor, step count sensor) and the like. For example, when it is determined that the passage between the two points is certain by the map matching method, the distance between the road map and the distance between them The number of steps sensor is calibrated from the number of steps measured in the step. Note that the walking distance may be corrected by a positioning locus between two points.

  The GPS 12 receives radio waves from a plurality of GPS satellites and measures the position of the pedestrian. In addition to GPS 12, DGPS (differential GPS) can also be mounted. DGPS can receive FM broadcast or medium wave transmitted from a reference station whose position is known in advance, and can correct the displacement of the positioning position obtained by GPS 12 and can improve the accuracy of the position of the pedestrian. . Note that it is also possible to perform positioning in a form that combines a method of roughly positioning a position with radio waves from a plurality of base stations of a mobile phone and GPS. As a result, even if it is difficult to receive radio waves from GPS satellites indoors, the probability that a position with poor position accuracy can be obtained is increased. Normally, the position of the pedestrian can be detected by satellite navigation or hybrid navigation using the positioning unit 20 and the GPS 12, but if positioning by the GPS 12 cannot be performed, the positioning unit 20 can be used independently. The position of a pedestrian can be detected by navigation or map matching.

  The posture identifying unit 30 identifies the posture of the device using data (acceleration) detected by the triaxial acceleration sensor 22. More specifically, the posture specifying unit 30, as will be described later, has a rotation angle α (pitch angle), β (roll angle), γ (yaw) with respect to a reference orthogonal three-dimensional coordinate (x, y, z) as a reference. The angle is determined to identify its own posture. Hereinafter, a method for specifying the posture will be described.

  FIG. 2 is an explanatory diagram showing a coordinate system for specifying the posture. As shown in FIG. 2, a reference orthogonal three-dimensional coordinate having two axes (x axis, y axis) orthogonal to the reference plane and an axis orthogonal to the two axes (z axis) is (x, y, z). And The reference plane can be, for example, a plane parallel to the plane of the road or passage on which the pedestrian walks. Even when the road or passage is inclined, the reference plane remains horizontal because the pedestrian does not incline. Also, the y-axis direction can be the pedestrian walking direction, the x-axis can be the right lateral direction with respect to the pedestrian walking direction, and the z-axis can be the vertical direction.

  Next, the apparatus orthogonal three-dimensional coordinate having an axis that makes an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinate (x, y, z) is defined as (X, Y, Z), and the reference orthogonal three-dimensional coordinate (x, The rotation angles of the device orthogonal three-dimensional coordinates (X, Y, Z) with respect to y, z) are α (pitch angle), β (roll angle), and γ (yaw angle), respectively. And The apparatus orthogonal three-dimensional coordinates (X, Y, Z) indicate the coordinate system of the position detection apparatus 100, for example, the three axes (X axis, Y axis, Z axis) of the three axis acceleration sensor 22 and the three axis geomagnetic sensor 23. 3 axes (X axis, Y axis, Z axis). In the following, in order to simplify the explanation, the three axes (X axis, Y axis, Z axis) of the three axis acceleration sensor 22 and the three axes (X axis, Y axis, Z axis) of the three axis geomagnetic sensor 23 are used. However, the three axes of the three-axis acceleration sensor 22 and the three axes of the three-axis geomagnetic sensor 23 do not have to be the same as long as the angles of the corresponding axes are known in advance.

  The coordinate conversion unit 31 performs coordinate conversion between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z). That is, coordinate conversion between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z) is performed using the pitch angle α, roll angle β, and yaw angle γ as parameters. As conversion equations, they can be expressed by equations (1) and (2).

  The angle calculation unit 32 calculates the pitch angle α, the roll angle β, and the yaw angle γ by applying the data (acceleration) detected by the triaxial acceleration sensor 22 to the conversion formula of the coordinate conversion unit 31. Hereinafter, a method for calculating the angle will be described.

The accelerations of the X-axis, Y-axis, and Z-axis components detected by the triaxial acceleration sensor 22 are AC _X , AC _Y , and AC _Z. The x-axis, y-axis, and z-axis components of the gravitational acceleration g at the reference orthogonal three-dimensional coordinates (x, y, z) are 0, 0, and g, respectively. By substituting this into equation (1), equation (3) can be obtained. In addition, the relationship of the equation (4) is established between the accelerations of the X-axis, Y-axis, and Z-axis components AC _X , AC _Y , and AC _Z.

  Since the gravitational acceleration g is known, the pitch angle α and the roll angle β can be calculated by the equations (3) and (4). If the yaw angle γ is 0 or is known in advance, the posture of itself can be specified by calculating the pitch angle α and the roll angle β.

  When a pedestrian walks with the position detection device 100 being carried, the position detection device 100 vibrates in accordance with walking and receives a dynamic acceleration fluctuation. FIG. 3 is an explanatory diagram showing an example of acceleration data detected by the triaxial acceleration sensor 22 during walking. In the example of FIG. 3, the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z) coincide with each other, and a pedestrian is walking almost straight.

  As shown in FIG. 3, when a pedestrian is walking, the X-axis and Y-axis accelerations on the horizontal plane (reference plane) are very small although dynamic acceleration affects them. On the other hand, the acceleration of the Z-axis increases due to the influence of a walking motion that raises the foot against gravity and lowers the foot in the direction of gravity, that is, the influence of dynamic acceleration in the vertical direction.

  Therefore, in this case, with respect to the data of each axis of the triaxial acceleration sensor 22, an average value (for example, an average value, a median value, an intermediate value between the maximum value and the minimum value) based on a plurality of times of the walking cycle is used. These are hereinafter referred to as “standard values.”) By calculating ACX, ACY, and ACZ, average acceleration can be obtained.

  Next, a case where the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z) do not match will be described. For example, there is a case where the X axis and the Y axis are not on the horizontal plane (reference plane, xy plane). This is because when the position detection device 100 is stored in a pocket or a bag of clothes of a pedestrian, the device orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z) are obtained. This is because the positions of the position detection device 100 are not always consistent and are not determined.

  Also in this case, the position detecting device 100 vibrates in accordance with the walking of the pedestrian and receives a dynamic acceleration fluctuation, but can be considered not to receive a large vibration other than the vibration accompanying the walking. Therefore, the standard values ACX, ACY, and ACZ of the acceleration data can be applied to the equations (3) and (4) as they are.

The gravitational acceleration g may be a constant, but may change depending on the position of the pedestrian, and the value of 1 g shown in the example of FIG. 3 is an intermediate value between the maximum value and the minimum value of the acceleration AC _Z. However, it is preferable to use the equation (4) instead of the gravitational acceleration g because a deviation may occur when the average value of the acceleration is obtained by averaging the vibrations.

  Next, a method for calculating the yaw angle γ using the acceleration detected by the triaxial acceleration sensor 22 when the yaw angle γ is unknown will be described. FIG. 3 described above shows that the apparatus orthogonal three-dimensional coordinates (X, Y, Z) coincide with the reference orthogonal three-dimensional coordinates (x, y, z), and the pedestrian is walking almost straight. The acceleration data detected by the axial acceleration sensor 22 is shown. Therefore, when the pedestrian walks almost straight, the acceleration of each axis of the reference orthogonal three-dimensional coordinates (x, y, z) shows the same characteristics. That is, the acceleration in the z-axis direction (vertical direction) is the largest, and the acceleration in the x-axis direction (straight side of the pedestrian) and the y-axis direction (walking and walking) is smaller than the acceleration in the z-axis direction. The acceleration in the z-axis direction is mainly caused by raising and lowering the foot, and the acceleration in the y-axis direction is caused to keep the walking speed constant when alternately stepping on the foot. The acceleration is generated because the pedestrian's body swings from side to side with walking.

  As shown in FIG. 3, although the instantaneous acceleration fluctuates, the characteristics of the acceleration of each axis of the reference orthogonal three-dimensional coordinates (x, y, z) have periodicity associated with walking, and at an appropriate time. When the instantaneous acceleration values are averaged or when the instantaneous acceleration values are added and averaged an appropriate number of times, the acceleration of each axis is considered to approach a constant value. Accordingly, it is possible to define an acceleration component obtained by averaging the accelerations of the respective axes of the reference orthogonal three-dimensional coordinates (x, y, z).

  FIG. 4 is an explanatory diagram showing an example of the acceleration component of the reference orthogonal three-dimensional coordinates (x, y, z). The example of FIG. 4 shows an acceleration component obtained by averaging instantaneous values of fluctuating acceleration over an appropriate time. When the instantaneous values of the fluctuating acceleration are averaged, it is considered that the acceleration component of each axis (x axis, y axis, z axis) of the reference orthogonal three-dimensional coordinates (x, y, z) is constant. As shown in FIG. 4, each acceleration component is set to v1, v2, and v3.

  Each acceleration component has the largest acceleration component in the vertical direction (z-axis direction), and the acceleration component in the walking direction (y-axis direction) and in the lateral direction of the pedestrian (x-axis direction) is the acceleration in the vertical direction. Assuming that there is a proportional relationship with the component, the relationship of Equation (5) is established. Note that s1 and s2 are appropriate proportionality coefficients, and the relationship of Expression (5) is stored in advance in the storage unit 15 as information indicating the relationship between the acceleration components of the respective axes in the reference orthogonal three-dimensional coordinates. Can do. In FIG. 4, the acceleration component v3 is the largest and the acceleration component v2 is greater than v1, but the magnitude of the acceleration component varies depending on the walking state of the pedestrian. The magnitude relationship between the acceleration components can be changed by appropriately determining the proportionality coefficient.

  On the other hand, a value obtained by averaging the accelerations detected by the triaxial acceleration sensor 22 for an appropriate time or an appropriate number of times is defined as an acceleration variable. Any acceleration averaging method can be used, for example, as long as it can average the instantaneous values of fluctuating acceleration using a value corresponding to the added value of acceleration detected for a suitable time or a suitable number of times. Such a method may be used. For example, the maximum value or minimum value (negative maximum value) of the acceleration of each axis (x axis, y axis, z axis) of the reference orthogonal three-dimensional coordinates (x, y, z) is set for an appropriate time or appropriately. Can be obtained by integrating or averaging. It can also be obtained by integrating the instantaneous acceleration value over an appropriate time. Alternatively, it can be obtained by integrating only positive acceleration or only negative acceleration. That is, if the acceleration variable of each axis of the apparatus orthogonal three-dimensional coordinates (X, Y, Z) is defined as an average of instantaneous values of acceleration detected by the three-axis acceleration sensor 22, it is detected by the three-axis acceleration sensor 22. The acceleration variable can be calculated by applying a method similar to the above-described method to the acceleration.

  Thereby, it is possible to perform conversion using data obtained by averaging accelerations in both the reference orthogonal three-dimensional coordinates (x, y, z) and the apparatus orthogonal three-dimensional coordinates (X, Y, Z). That is, the acceleration variable of each axis of the apparatus orthogonal three-dimensional coordinates (X, Y, Z) can be mapped to the acceleration component of each axis of the reference orthogonal three-dimensional coordinates (x, y, z).

  The acceleration variables for the X, Y, and Z axes of the apparatus orthogonal three-dimensional coordinates (X, Y, Z) are ABX, ABY, ABZ, and the x-axis, y of the reference orthogonal three-dimensional coordinates (x, y, z) Assuming that the acceleration components of the axes and z-axis are v1, v2, and v3, Expression (1) to Expression (6) are obtained using Expression (5).

By defining the variables p1 and p2 with Expression (7) and substituting them into Expression (6), Expression (8) is obtained. Expression (8) can be expressed by Expression (9) to Expression (11), and variables p1, p2, and v3 can be obtained from Expression (12). Here, U ′ represents a transposed matrix of the matrix U, and (U′U) ⁻¹ represents an inverse matrix of the matrix (U′U). The yaw angle γ can be calculated from the calculated p1 and p2 by the equation (13).

  As the proportional coefficients s1 and s2 used in the above description, predetermined values can be used, or values obtained by learning based on data collected after the fact can be used. For example, s1 and s2 may be determined so as to satisfy the expressions (7) and (12).

  The determination unit 33 determines whether or not the posture specifying unit 30 specifies the posture of itself (position detection device 100). More specifically, the determination unit 33 specifies its own posture in accordance with a change in its own posture when it is carried or a change in walking characteristics of the pedestrian (for example, whether or not walking is stopped, a change in the walking direction). It is determined whether or not.

  When there is no change in the walking characteristics of the pedestrian and the posture of the position detection device 100 does not change, the data detected by the three-axis acceleration sensor 22 also has periodicity associated with walking, and the maximum data There is no significant fluctuation in the value or minimum value. However, if there is a change in the posture of the person or a change in the walking characteristics of the pedestrian, the data detected by the three-axis acceleration sensor 22 also includes unexpected fluctuations. Do not. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

  First, the posture change of the position detection device 100 will be described. The absolute value of the acceleration detected by the three-axis acceleration sensor 22 is compared with a predetermined acceleration threshold value. When the detected absolute value of the acceleration is larger than the acceleration threshold value (when it is an abnormal value), the position detection device 100 has a large posture. It is determined that the posture of the position detection device 100 is not specified as having changed. For example, when the position detection device 100 held in a pocket or a bag is taken out, when the position detection device 100 is stored in a pocket or a bag, the hand holding the position detection device 100 is changed, or the position detection device 100 is inserted. When the posture of the bag changes, the posture of the position detection device 100 also changes, and the absolute value of the acceleration detected by the triaxial acceleration sensor 22 changes greatly at least temporarily. Even in such a state, if the posture is specified using acceleration data, the correct posture cannot be obtained. Therefore, if the detected absolute value of acceleration is larger than the acceleration threshold, it is determined that the posture is not specified. To do. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

  Further, the absolute value of the geomagnetism detected by the triaxial geomagnetic sensor 23 (or the azimuth sensor) is compared with a predetermined geomagnetic threshold, and when the detected absolute value of the geomagnetism is larger than the geomagnetic threshold, or the detected azimuth angle is predetermined. If it is larger than the azimuth angle threshold, it can be determined that the posture of the position detection device 100 is not specified, assuming that the walking direction of the pedestrian has greatly changed. Even in such a state, if the user's posture is specified using the acceleration data, the correct posture cannot be obtained, and therefore it is determined that the user's posture is not specified. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

The determination of the abnormal value accompanying the posture change of the position detection device 100 described above can be performed as follows, for example. Assuming that the acceleration of the X-axis, Y-axis, and Z-axis components is AC _X , AC _Y , and AC _Z , in the case of abnormality, AC _X −ACxmax> T1x, AC _Y −ACymax> T1y, AC _Z −ACzmax> T1z, AC _{X -ACxmin <-T2x, AC Y -ACymin} <-T2y, or AC _Z -ACzmin <either -T2z is established. Here, T1x, T1y, T1z, T2x, T2y, and T2z are positive constants. ACxmax, ACymax, and ACzmax are data based on a constant or the maximum value of each axis of acceleration data when the posture is stable in the past, such as an average value or an exponential smoothing value. ACxmin, ACymin, and ACzmin are data based on a constant or the minimum value of each axis of acceleration data when the posture is stabilized in the past, and are an average value, an exponential smoothed value, or the like.

  In addition, when the posture of the position detection device 100 does not change, the data detected by the three-axis acceleration sensor 22 also has periodicity associated with walking, and the walking cycle, the number of steps, the walking speed, and the like are not substantially changed and are substantially constant. . However, when the posture of the position detection device 100 changes, the walking cycle, the number of steps, the walking speed, and the like also change. When a change in the walking cycle, the number of steps, or the walking speed is detected based on a change in acceleration detected by the triaxial acceleration sensor 22, it is determined that the posture of the position detection device 100 is not specified. As a result, when the posture of the position detection device 100 is changed, the posture of the position detection device 100 is not specified, and it is possible to prevent the specification of an incorrect posture and improve the accuracy.

  For example, when the number of steps is HW, when the absolute value of (HW−HWave) is larger than the threshold value Hk, it can be determined that there is an abnormality in the data periodicity. Here, HWave is, for example, a constant or an average value or exponential smoothing value of step count data when the posture is stable in the past.

  Next, changes in walking characteristics will be described. As an example of the change in walking characteristics, for example, when walking stop is detected, it is determined that the position detection device 100 is not specified in posture. When the pedestrian stops walking, for example, it is considered that the position detection device 100 is taken out from a pocket or a bag, and the posture of the position detection device 100 is likely to change. By not specifying the position of the position detection device 100, it is possible to prevent specifying the wrong position and improve the accuracy. Specifically, it can be determined that walking is stopped when the walking cycle, the number of steps, and the like cannot be detected.

  The determination by the determination unit 33 may be performed using a result obtained by continuously performing the above-described determination for a certain period, or only a part of the above-described determination may be performed.

  The posture specifying unit 30 converts the reliability of the angle calculated by the angle calculation unit 32 between the apparatus orthogonal three-dimensional coordinates (X, Y, Z) and the reference orthogonal three-dimensional coordinates (x, y, z). When the evaluation is performed using (conversion means) and it is evaluated that the reliability is low, the calculated angle is invalidated. For example, the reliability of the solution (calculated angle) depends on how much the acceleration data detected by the triaxial acceleration sensor 22 satisfies the equation (7) with p1, p2, and v3 obtained by the above equation (12). Can be determined. If the equation (7) is not sufficiently satisfied using the redundancy of the equation for the variable, it can be determined that the reliability of the calculation result is low. Note that invalidation includes, for example, positively rejecting the calculated angle or not using the calculated angle without rejecting it. As a result, even if the angle for specifying the posture can be obtained, if the reliability of the calculation result is low, the calculation result is invalidated, and the accuracy is improved by preventing the wrong posture from being specified. Can do.

  Further, the posture specifying unit 30 stores the angles calculated over a plurality of time points in the storage unit 15. Thereby, data calculated in the past is collected. The posture identifying unit 30 calculates an angle difference between the newly calculated angle and the stored angle, and invalidates the calculated angle when the calculated angle difference is greater than a predetermined threshold. For example, when the calculated data is larger than the data calculated in the past, it is determined that the abnormality is included in the acceleration data detected by the triaxial acceleration sensor 22, and the calculated result is reliable. If there is no, invalidate the calculation result. As a result, it is possible to prevent the specification of an incorrect posture and increase the accuracy.

  The posture specification (calculation of rotation angles α, β, γ) by the posture specification unit 30 can be performed as appropriate. For example, if there is no change in the posture of the position detection device 100, it is considered that the posture does not change except for vibration caused by walking, so the frequency of specifying the posture can be reduced. For example, after the determination unit 33 determines that the posture of the position detection device 100 is changing, the posture is specified using data detected by the triaxial acceleration sensor 22 at a timing when the posture is determined to be stable. Thereafter, until it is determined that the posture of the position detection device 100 has changed, it is not necessary to determine the posture as it is, or to specify the posture appropriately in order to increase the accuracy of the posture. You may make it obtain | require the average. In addition, in the case where the posture is appropriately specified and averaged, various methods such as an average obtained by removing the median value, the maximum value, and the minimum value can be employed in addition to obtaining a simple average value. Further, these values can be used as past collected data.

  The direction specifying unit 14 specifies the walking direction (moving direction) of the pedestrian using the geomagnetism of the X-axis, Y-axis, and Z-axis components detected by the triaxial geomagnetic sensor 23.

Hereinafter, a method for specifying the walking direction will be described. Assume that the x-axis, y-axis, and z-axis components of the geomagnetism m at the reference orthogonal three-dimensional coordinates (x, y, z) are mx, my, and mz, respectively. The vertical component mz is known if the approximate position of the pedestrian, that is, the longitude and latitude are determined, and the geomagnetic vertical component mz can be stored in association with the latitude and longitude. Further, when the geomagnetism of the X-axis, Y-axis, and Z-axis components detected by the triaxial geomagnetic sensor 23 is MG _X , MG _Y , MG _Z , Expression (14) is established. In addition, the relationship of Expression (15) is established for the geomagnetism mx, my, mz of the x-axis, y-axis, and z-axis components.

  By substituting the pitch angle α, roll angle β, and yaw angle γ calculated using the triaxial acceleration sensor 22 into the equation (14), mx and my can be obtained.

  The azimuth specifying unit 14 calculates the geomagnetic component of two axes (for example, the x axis and the y axis) orthogonal to each other on the reference plane (for example, the horizontal plane) using the angle calculated by the angle calculation unit 32, and calculates the calculated geomagnetic component To identify its own moving direction. For example, the x-axis component and the y-axis component of geomagnetism are mx and my, respectively. Further, if the azimuth angle between the walking azimuth of the pedestrian (the direction of the y-axis) and the true east is θ (counterclockwise is positive), the azimuth angle θ can be obtained by Expression (16).

  When the azimuth angle θ is 0, the walking azimuth is true east, and when θ is π / 2, the walking azimuth is true north. By using the three-axis acceleration sensor and the three-axis geomagnetic sensor, the walking direction of the pedestrian can be obtained. In this case, a deviation angle (for example, 5 to 7 degrees) which is a difference between the azimuth direction and the true azimuth, that is, a deviation between true north and magnetic north can be taken into consideration.

  After the attitude of the position detection device 100 is specified by the attitude specifying unit 30, the acceleration component of the reference orthogonal three-dimensional coordinate is obtained using the expressions (1) and (2) unless the specified attitude changes. Can do. The z-axis acceleration component (vertical component) of the reference orthogonal three-dimensional coordinate can be used to calculate the pedestrian's moving distance (walking distance) through the step count measurement. Thus, only when the posture of the position detection device 100 can be specified by the posture specifying unit 30, the movement distance and acceleration can be accurately obtained using the data detected by the distance sensor 21 and the triaxial acceleration sensor 22. it can.

  The positioning unit 20 obtains the data detected by the distance sensor 21, the pedestrian's moving direction (walking direction) specified by the direction specifying unit 14, positioning data such as the positioning result obtained by the GPS 12, and the communication unit 11. The positioning position and the positioning position error are calculated based on the received radio wave reception result from the base station or the position information of the communication point obtained by the narrow area communication with the road device. Hereinafter, a positioning position and a calculation method of the error will be described.

FIG. 5 is an explanatory diagram showing an example of the error range of the positioning position. In the orthogonal coordinate system (x direction and y direction) on a plane parallel to the plane of the road or passage on which the pedestrian walks, the error range of the position detected by the narrow area communication with the GPS, base station or road device, As an example, the rectangular area (x-direction length is 4a and y-direction length is 4b) is set. That is, the positioning position is the center position of the rectangular area, and the error range extends from the center position by a range of ± 2a in the x direction and by a range of ± 2b in the y direction. For example, when 2a is set to 2 sigma, the variance in the x direction is a ² and the standard deviation can be set to a. When 2b is set to 2 sigma, the variance in the y direction is b ² and the standard deviation can be set to b.

  Since the error due to the narrow area communication with the road device is smaller than that of the GPS or the base station communication (for example, the error range is several m), the error range is determined by using the narrow area communication with the road device or the GPS. Or, it depends on whether base station communication is used. Further, for example, when using GPS, the error range is temporally determined by environmental conditions, more specifically, GPS reception level, number of captured satellites, 2D or 3D positioning, CEP (Circular Error Probability). To change. In the case of base station communication, the error range varies with time depending on the communication level with the base station, the communication range of the base station, and the like. A predetermined constant in which the error range is set to be large in advance, a constant determined in advance according to the place or time, or the like may be used. The shape of the error range is not limited to a rectangular shape, and may be an arbitrary shape such as a circle or an ellipse. For example, when positioning is performed using only GPS, when the environmental conditions are favorable, an error range of about 10 to 20 m can be set.

  Hereinafter, a method for calculating the positioning position of the pedestrian will be described. The positioning position is expressed by a two-dimensional vector in an orthogonal coordinate system, but in three dimensions, only altitude information is added and can be easily expanded. Further, in the following description, it is formulated by time, but in actual processing, processing may be performed for each unit travel distance instead of processing for each unit time. In the following, capital letters are assumed to be vectors or matrices.

  If the position P (t) of the pedestrian at time t is expressed by equation (17), walking at time t + 1 (the interval between times t and t + 1 is a predetermined time, for example, 1 second, 0.5 seconds, etc.). The person's position P (t + 1) can be expressed by Expression (18). Alternatively, the time at which the pedestrian has traveled a predetermined travel distance (eg, 1 m, 2 m, etc.) from time t can be set as time t + 1. Note that “T” added to the vector means transposition. Moreover, Formula (18) shows a pedestrian's dynamic characteristic. Note that the position P (t) of the pedestrian at time t is the true position (actual position) of the pedestrian, and is an unobservable position due to the presence of an unknown error. That is, the positioning position of the pedestrian is to obtain an optimum estimated position with respect to the true position P (t).

  Here, D (t) is expressed by Equation (19), d (t) is the distance that the pedestrian has moved (walked) from time t to time t + 1, and θ (t) is the orthogonal coordinate system. This is the azimuth angle of pedestrian movement (walking). E (t) is expressed by equation (20), and e (t) is an error of the moving distance d (t). Further, the variance Q (t) of the error E (t) is expressed by the equation (21), and q is the error variance in the unit distance movement and can be a constant value.

  Further, at time t, the position S (t) detected by GPS, base station communication, or communication with a road device can be expressed by Expression (22). Here, G (t) is an error of the position S (t), and the covariance matrix R (t) of the error G (t) can be expressed by Expression (23). In Expression (23), a and b are each one-fourth of the length in the x direction and the y direction of the rectangular area that is the error range shown in FIG. That is, the covariance matrix R (t) is composed of variances in the x and y directions when 2a and 2b are 2 sigma. Even if the average value of E (t) and G (t) is 0, generality is not lost.

  The optimum estimated position H (t) of the pedestrian position P (t) at time t is expressed by a recurrence formula as shown in Expression (24) by the Kalman filter.

Here, Γ (t) is the variance of the estimation error of the estimated position H (t), and can be expressed by a recurrence formula like Formula (25). Further, “−1” attached to a matrix means an inverse matrix of the matrix. Further, the estimated position H (0) at the initial time 0 and the variance Γ (0) of the estimation error can be expressed by Expression (26) and Expression (27), respectively. Here, M is a priori information on the initial position of the pedestrian, and Σ is its error variance. If there is no a priori information, M = 0 and Σ ⁻¹ = 0, and the estimated position H (0) at the initial time 0 and the variance Γ (0) of the estimated error are expressed by the equations (28) and (29), respectively. ).

Note that since the equation (22) is obtained when the position is detected by GPS, base station communication, or communication with a road device, the equation (22) is obtained while communication with the GPS, base station communication, or road device is not performed. Assuming that the errors a and b in (23) are sufficiently large, in Equation (24), if R ⁻¹ (t) = 0, the estimated position is repeatedly calculated using Equation (24) as it is. Can do. That is, in this case, it is equivalent to measuring the position only by the self-contained navigation.

  The error e (t) of the movement distance d (t) varies depending on the type of distance sensor. For example, when a plurality of types of distance sensors are used at the same time, a combined error obtained by combining the errors of the sensors may be set. For example, if the distances obtained by the two types of sensors are d1 and d2, and the variances of the errors are q1 and q2, respectively, the coupling distance is set by the equation (30), and the variance as the coupling error is expressed by the equation ( 31).

  In the above formulation, in order to simplify the description, it is assumed that there is no error in the azimuth (azimuth angle) θ, but linear approximation is performed in consideration of the error f (t) in the azimuth (azimuth angle) θ. Thus, the above formulation can be easily extended. In this case, the error variance of the error f (t) is defined similarly to the errors e (t) and G (t). Alternatively, the position error variance can be increased in consideration of the azimuth error.

  For example, when the error F (t) is added to the measured value of the azimuth θ at time t, if the higher order error is ignored, the equation (18) can be expanded as the equations (32) and (33). Can do.

  In this case, equation (21) may be replaced with equation (34). Here, u is the variance of the error of the azimuth θ. It should be noted that instead of the equations (32) to (34), Q (t) in the equation (21) may be set to a larger value. In the above mathematical formula, it is formulated as two-dimensional position detection, but it may be formulated in three dimensions by adding a height dimension.

  The map database 13 stores a wide range of map information. In addition, according to the position of a pedestrian, the map information of the vicinity can also be acquired by communication from the outside, such as a center apparatus or a road device, and memorize | stored.

  FIG. 6 is a schematic diagram showing an example of map information, and FIG. 7 is an explanatory diagram showing an example of attributes of areas on the map. When detecting the position of a pedestrian, it becomes more complicated and difficult than when detecting the position of a vehicle. In other words, in the case of a vehicle, the position of the vehicle can be detected by map matching between the estimated position and the roadway on the map, whereas in the case of a pedestrian, walking other than the pedestrian sidewalk is possible. There are various areas where a person can walk. Moreover, the necessity of detecting the position of a pedestrian is high not only outdoors but indoors.

  In addition, when detecting the position of a pedestrian, detailed map matching such as separation of a sidewalk and a roadway is required, so detailed data is also required as map information. However, it is not necessary to store a wide range of map information in the storage unit 15 of the position detection device 100, and may be acquired by communication from the outside such as an information center device or a road device as appropriate according to the position of the pedestrian. .

  As shown in FIG. 6, there are various areas such as a pedestrian road (sidewalk), a roadway, a pedestrian crossing, a building, a retail store, a park, and a pond on the map. Therefore, as shown in FIG. 7, the attributes of the areas on the map are defined to classify the areas on the map. The attribute is first divided into a walkable area and a prohibited area. Forbidden areas are areas where entry of pedestrians is generally prohibited, such as restricted areas, rivers, ponds, seas, lakes, swamps, ponds, cliffs, railway sites, imperial palaces, etc. It is.

  The walking area is divided into an indoor area, a road area, and other areas. The indoor area is, for example, a building, an underpass, a station building, a store, a retail store, a house, a factory, an underground mall, a building interior, and the like. Road areas include, for example, pedestrian roads, sidewalks in the case of main roads with separate sidewalks and roadways, pedestrian crossings, pedestrian overpasses (pedestrian bridges), underground crosswalks, railroad crossings, and other privately owned roads. Called a road). The other area is, for example, a roadway, a park, a playground, or any other outdoor area that can be freely walked. In addition, when a pedestrian's walk is limited in an indoor area | region or another area | region, a walk path (road) can be set in it and it can also be set as a road area | region.

  As shown in FIG. 6, the road area is a pedestrian road, a main road sidewalk, a pedestrian crossing, a pedestrian overpass, an underground crossing passage, and an indoor area is a building, a retail store, a house, a retail store There are roads. The prohibited areas are ponds, and the other areas are highway roads and parks. Based on the information of FIGS. 6 and 7, roads on the map (road areas, line segments for map matching) can be set. In FIG. 7, the attributes are divided into walkable areas and prohibited areas, and then the walkable areas are divided into indoor areas, road areas, and other areas. As examples of each attribute, pedestrian roads, crosswalks, Although pedestrian overpasses are listed, the attributes are not limited to the hierarchical structure as described above, and pedestrian roads, pedestrian crossings, pedestrian overpasses, etc. can be defined as one attribute. . Moreover, the example of FIG. 7 is an example, Comprising: It is not limited to this.

  FIG. 8 is an explanatory diagram showing an example of setting a road on a map. As shown in FIG. 8, one or more line segments (standard gait line, road on the map) for map matching are defined for roads in the road area, indoor area, and other areas, and each line segment is connected. Set the relationship information. The road (road area) on the map can be constituted by links and nodes, or a width corresponding to the road width for the line segment can be set. A node is a part of the road on the map, and the node defines the connection point of the link, ie the connectivity of the road on the map. The road to be set is synonymous with the road area.

  For example, in the example of FIG. 6, the road on the map is set as shown in FIG. Roads on the map can include road width information and connection information. In addition, the indoor area and the prohibited area may include information on the range of each area. The other area may include area range information, or if it is determined that the other area is not a road area, indoor area, or prohibited area, do not include the area range information in the other area. May be. Although not shown in FIGS. 6 and 8, a point or road (passage) connecting the regions may be set on the map. For example, if there is a communication port (stairs, elevator, escalator, etc.) on the road leading to an underground area in an indoor area or a subway station building, the position of the pedestrian can be changed when moving from the road to the indoor area. It can be corrected.

  The storage unit 15 stores various information received via the communication unit 11, positioning data measured by the positioning unit 20, processing results processed by the posture specifying unit 30, the direction specifying unit 14, the position detection processing unit 40, and the like. . In addition, when any one or some of the control unit 10, the posture specifying unit 30, the direction specifying unit 14, the position detection processing unit 40, and the like are configured by a CPU, a RAM, etc., a computer program that defines the processing procedure of each unit is stored. You can also

  The operation unit 16 includes various operation buttons and functions as a user interface between the pedestrian and the position detection device 100. For example, the operation unit 16 receives an operation for starting or stopping the operation of the position detection device 100 by an operation of a pedestrian.

  The position detection processing unit 40 performs a process for detecting the position of the pedestrian.

  The position estimation unit 41 updates the estimated position or the estimated position calculated last time (for example, may be the latest or may be two or three times before), and is detected as the position of the pedestrian. Based on the positioning position trajectory calculated by the positioning unit 20 (positioning trajectory), an estimated position on the map (estimated position trajectory) is calculated. More specifically, the trajectory and the estimated position of the estimated position are calculated by obtaining a trajectory along the positioning trajectory from the estimated position or the detection position calculated last time to the positioning position.

  The error calculation unit 42 calculates an error range of the estimated position estimated by the position estimation unit 41. More specifically, the error calculating unit 42 sets the error range of the estimated position to a predetermined value when initially registering the estimated position or updating the estimated position as will be described later. For example, when the estimated position is initially registered, the error range of the estimated position can be set to an error range of the positioning position (for example, 20 to 200 m). Further, when the estimated position is updated at a characteristic point such as a curve or an intersection on the road, the minimum error (for example, a range of about the road width) can be obtained.

  After the error calculation unit 42 sets the error range of the initially registered estimated position or the updated estimated position to a predetermined value, the pedestrian's moving distance or movement from the initially registered or updated estimated position to the set predetermined value. An error range is calculated by adding a value corresponding to the direction (for example, an increase in positioning error). As a result, once the position of the pedestrian is determined (confirmed), and after setting the error range at that position to a predetermined value, the positioning error increases with an increase in the positioning trajectory (movement distance or direction) However, an appropriate error range of the estimated position can be obtained according to the positioning locus. Note that the error range can always be an appropriate predetermined constant value (for example, 100 m).

  The attribute determination unit 43 determines the attribute of the area on the map that exists within the error range of the estimated position calculated by the error calculation unit 42. If a plurality of attributes exist within the error range of the estimated position, it can be determined that the attribute is the area closest to the estimated position (the distance is short). Alternatively, a priority order can be set for each attribute, and it can be determined that the attribute has the highest priority order. In this case, the priority order of the attributes may be, for example, the road area, the indoor area, and the other area from the higher rank, or the road area, the other area, and the indoor area from the highest rank. .

  The walking behavior determination unit 44 determines the walking behavior of the pedestrian based on the positioning data obtained by the positioning unit 20. The walking behavior indicates a walking characteristic of a pedestrian and includes a traveling characteristic when riding a bicycle. The walking behavior is, for example, the start of walking, walking speed, fluctuation in walking speed, walking strength (for example, the level intensity indicated by acceleration by the step sensor), fluctuation in walking strength, per unit time The number of steps (in the case of a bicycle, the number of pedaling), the variation in the number of steps, the walking direction, the walking stop, and the like.

  The walking behavior determination unit 44 determines the presence or absence of walking disturbance based on the walking behavior. For example, when the walking speed is low or the number of steps is small with reference to the predetermined walking speed or the number of steps per unit time, the walking speed is not substantially constant and the walking speed varies frequently, This is the case when there is meandering or circulation, or when the strength of walking is unstable.

  The position update unit 45 performs initial registration of the estimated position based on the positioning position. Further, the position update unit 45 uses the map matching method to walk within the error range based on the attribute within the error range determined by the attribute determination unit 43 and the estimated position calculated by the position estimation unit 41. The estimated position is updated to a position considered to be present, and the updated position is detected as the position of the pedestrian. In this case, the most probable estimated position is detected as the position of the pedestrian based on the evaluation coefficient calculated by the evaluation unit 47. Details of the evaluation coefficient will be described later. Note that the reciprocal of the evaluation coefficient can be defined as the degree of correlation, and the degree of correlation can be used.

  The reliability calculation unit 46 calculates the reliability of the positioning data measured by the positioning unit 20. More specifically, the reliability calculation unit 46 performs self-contradiction of data in a single sensor such as the distance sensor 21, the three-axis acceleration sensor 22, the three-axis geomagnetic sensor 23, data contradiction between sensors, or map information. The determination of the presence / absence of abnormality such as inconsistency with GPS, the calculation result of the GPS 12 positioning, and the use environment index indicating the reliability of base station communication are performed. In addition, when it is determined that there is an abnormality in the sensor or the like, a usable sensor is selected, and for an unusable sensor, a process for making the error variance of the sensor infinite (or increased) is performed. Do. Thereby, the availability of the sensor is always monitored. Hereinafter, details of the processing in the reliability calculation unit 46 will be described.

  The determination of the presence or absence of an abnormality in the GPS 12 is made by, for example, determining whether there is a self-contradiction in temporal changes such as a positioning position determined based on the positioning result obtained by the GPS 12, a walking speed of the pedestrian, and a moving direction. be able to. When there is an abnormality, the error variance of the GPS 12 data is set to infinity so that it is not used. In addition, when such an abnormal situation continues for a predetermined time and / or a predetermined distance, the reliability of the GPS 12 is assumed to be low. In addition, when such an abnormal situation disappears, the GPS 12 data is used, and if the normal situation continues for a predetermined time or a predetermined distance, the GPS reliability is restored to a normal value. .

  For example, if the GPS 12 reception level, the number of captured satellites, 2D or 3D positioning, or CEP (Circular Error Probability) is below a predetermined standard value, the GPS 12 is considered abnormal. Make the error variance of the data infinite and avoid using it. In addition, when such an abnormal situation continues for a predetermined time and / or a predetermined distance, the reliability of the GPS 12 is assumed to be low. Further, when such an abnormal situation disappears, the GPS 12 data is used, and if the normal situation continues for a predetermined time or a predetermined distance, the reliability of the GPS 12 is restored to a normal value. .

  Base station communication usage environment indicators include, for example, when the communication level with the base station, the communication range of the base station, etc. falls below a predetermined standard value, the error distribution of data due to base station communication is infinite as abnormal And avoid using it. In addition, when such an abnormal situation continues for a predetermined time and / or a predetermined distance, it is assumed that the reliability of base station communication is low. In addition, when such an abnormal situation disappears, the base station communication data is used, and if the normal situation continues for a predetermined time and / or a predetermined distance, the reliability of the base station communication is increased. Return to normal value. In the above description, GPS and base station communication are distinguished from each other, but a positioning method combining GPS and base station communication may be used.

  The evaluation unit 47 calculates an evaluation coefficient for evaluating the estimated position calculated by the position estimation unit 41 and the estimated position updated by the position update unit 45. The evaluation coefficient is a coefficient for evaluating the certainty of the estimated position. For example, the smaller the evaluation coefficient, the larger the certainty (probability) of the estimated position. The evaluation coefficient is, for example, the positional deviation between the estimated position and the positioning position, the average of the positional deviation between the estimated position at the characteristic point such as a curve or an intersection, and the estimated position and the road at the characteristic point such as a curve or an intersection. And the average of the positional deviation (distance deviation). Details of the evaluation coefficient will be described later.

  The output unit 17 includes a liquid crystal display panel, a speaker, etc., and displays the position of the pedestrian on the map and notifies the pedestrian of necessary information when displaying the position of the pedestrian, or Output voice or sound to call attention.

  Next, the position detection process by the map matching method of the position detection apparatus 100 will be described. In the following description, it is assumed that no passage (road) is set in the indoor area and other areas. When a passage is set, the passage can be handled as a road.

  FIG. 9 is an explanatory view showing an example of new registration of an estimated position. The new registration (initial registration) of the estimated position is a process performed when the map matching process is started or when there is no estimated position candidate.

  Whether or not to newly register the estimated position is determined by, for example, the following condition (1) and condition (2). That is, when both the condition (1) and the condition (2) are satisfied, the new registration of the estimated position is not performed, and when either the condition (1) or the condition (2) is not satisfied, the new registration of the estimated position is performed. I do. Condition (1) is when the reliability of the sensor or the like is equal to or lower than a predetermined threshold (low reliability), for example, when the reliability of the GPS 12 is bad. Further, the condition (2) is a case where there is a disturbance of walking over a predetermined range (time and / or distance). That is, when the condition (1) and the condition (2) are satisfied, there is a high possibility that the position of the pedestrian is in the indoor area. Therefore, the map matching process is not performed until the pedestrian enters the outdoor area from the indoor area.

  As shown in FIG. 9, it is determined whether or not there is a road area, that is, a road on the map, within the error range of the positioning position A. If there is a road (road area), the closest to the positioning position A The point on the road is registered as a new estimated position corresponding to the positioning position A. In FIG. 9, since there are two roads within the error range, the points M and N closest to the positioning position A on each road are registered as estimated positions. In this case, it is also possible to register in advance a road whose orientation substantially coincides with the orientation of the positioning locus up to the positioning location A or the positioning orientation at the positioning location A. If there is no road where the orientation of the positioning trajectory or the positioning orientation substantially matches the road orientation, the system waits without newly registering the estimated position until there is a road that can be registered within the error range.

  When the estimated position is newly registered, the error range of the positioning position corresponding to the estimated position is set (registered) as the error range of the newly registered estimated position. In the example of FIG. 9, the error range of the estimated positions M and N inherits the error range of the estimated position A. Further, based on the positional deviation between the newly registered estimated positions M and N and the corresponding positioning position A, evaluation coefficients for the estimated positions M and N are calculated.

  FIG. 10 is an explanatory diagram showing an example of calculation of an evaluation coefficient for a newly registered estimated position. The example of FIG. 10 shows a case where the estimated position M is newly registered in association with the positioning position A. If the coordinates of the positioning position A are (xa, ya) and the coordinates of the newly registered estimated position M are (x, y), the evaluation coefficient C of the estimated position M can be C = C1 + C2. Here, C1 = | x−xa | and C2 = | y−ya |. That is, the evaluation coefficient C of the estimated position M can be the difference between the x coordinate of the estimated position M and the positioning position A, and the difference of the y coordinate between the estimated position M and the positioning position A. In this case, the smaller the evaluation coefficient, the higher the probability (probability) of the estimated position. The evaluation coefficient C is not only calculated for each x coordinate and each y coordinate, but can also be used as one index by the sum of absolute values of the x coordinate and the y coordinate or the square root of the square sum. Thereby, it is possible to grasp how much the estimated position M is relative to the positioning position A.

  Next, an example of calculating the locus of the estimated position after new registration will be described. FIG. 11 is an explanatory diagram illustrating a calculation example of the locus of the estimated position. In the example of FIG. 11, the estimated position M is newly registered in association with the positioning position A. The calculation example of the locus of the estimated position shown in FIG. 11A is to obtain the locus of the estimated position by translating (shifting) the positioning locus from the positioning position A with the estimated position M as a starting point. .

  In the case of FIG. 11B, in the case of FIG. 11A, when the azimuth difference between the direction of the road on the map and the direction of the locus of the estimated position or the direction at the estimated position is smaller than a predetermined threshold value. The estimated position is updated (corrected) on the road every time the predetermined time has elapsed and / or the predetermined distance has been moved. Note that when the estimated position is updated to a position on the road, a value corresponding to the positional deviation may be added to the evaluation coefficient.

  In the case of FIG. 11C, in the case of FIG. 11A, when the azimuth difference between the azimuth of the road on the map and the azimuth of the locus of the estimated position or the azimuth at the estimated position is smaller than a predetermined threshold value. Always update (correct) the estimated position on the road. Note that when the estimated position is updated to a position on the road, a value corresponding to the positional deviation may be added to the evaluation coefficient.

  In the example of FIG. 11D, in the case of FIG. 11C, when the azimuth difference between the azimuth of the road on the map and the azimuth of the locus of the estimated position or the azimuth at the estimated position exceeds a predetermined threshold value. Indicates that updating (correcting) the position on the road is not performed. In addition, in the case of FIG. 11 (a) and FIG.11 (b), it can also correct on a map, when displaying on a liquid crystal display panel.

  The example of FIG. 11E is an example of correcting the estimated position when the road is represented by a line segment, while the estimated position is obtained when the road is represented by a two-dimensional figure as it is. This is an example of correction.

  The error range of the estimated position can be a positioning error (for example, a range of 20 to 200 m) in the new registration (initial registration) of the estimated position. Further, the error range of the estimated position can be set to the minimum error range (for example, about the road width) when the estimated position is updated (corrected) to a position on the road at a characteristic point such as a curve or an intersection. . Thereafter, as the pedestrian walks, positioning errors are accumulated, so that the error range of the estimated position can be increased. As a result, once the position of the pedestrian is determined (confirmed), and after setting the error range at that position to a predetermined value, the positioning error increases with an increase in the positioning trajectory (movement distance or direction) However, an appropriate error range of the estimated position can be obtained according to the positioning locus. The error range of the estimated position can always be an appropriate predetermined constant value (for example, 100 m). Thereby, the processing effort of position detection can be reduced.

  Next, a method for updating the estimated position will be described. The update of the estimated position includes, for example, the attribute of the area on the map within the error range of the estimated position, the connection characteristic of the road, the connection characteristic between the road area and other areas, the walking behavior of the pedestrian, the reliability of the positioning data ( Reliability), evaluation coefficient of estimated position, etc. If the estimated position is not valid, the estimated position is rejected.

  FIG. 12 is an explanatory diagram showing an example of updating the estimated position. In the example of FIG. 12, it is assumed that the pedestrian is walking outdoors. Corresponding to the positioning trajectory up to the positioning position A, the positioning position A is changed to the positioning position A by the trajectory of the estimated position from the estimated position updated last time (for example, may be the most recent time or may be two or three times before). Correspondingly, there are two estimated positions M and N. Since a road (road area) exists within the error range of the estimated position M, the estimated position M is updated to a position on the road. Further, a value corresponding to the positional deviation between the estimated position M and the updated estimated position may be added to the evaluation coefficient of the estimated position M and inherited as the updated estimated coefficient of the estimated position. In addition, at a point having a characteristic such as a curve or an intersection, the estimated position can be updated by correcting the displacement of the estimated position M, and the evaluation coefficient and error range can be updated. In this case, as the error range to be updated, for example, a minimum value (a range about the road width) can be set. It is possible to determine whether or not the azimuth difference between the azimuth of the estimated position and the azimuth of the road is smaller than a predetermined threshold, and when the azimuth difference is larger than the threshold, the estimated position may not be updated on the road.

  Within the error range of the estimated position N, the road (road area) in which the estimated position updated last time is outside the error range, so it is determined whether there is another road within the error range. If another road exists, the estimated position is updated to the position of the road, and the error range of the estimated position N, the error range of the estimated position with the updated evaluation coefficient, and the evaluation coefficient are taken over. If it is determined that there is no road to be updated, the attribute of the estimated position is determined, and if the determined attribute is another region, the position obtained by accumulating the amount of change in the walking trajectory accompanying the walking of the pedestrian from the estimated position N Is assumed to be the locus of the estimated position, and the evaluation coefficient and error range are taken over. If the determined attribute is a prohibited region, the estimated position N is rejected (see FIG. 12).

  FIG. 13 is an explanatory diagram showing another example of updating the estimated position. The example of FIG. 13 is a case where the road where the estimated position updated last time (for example, may be the latest or may be two or three times before) branches to two roads. Assume that there is one estimated position M corresponding to the positioning position A by the locus of the estimated position from the previously updated estimated position corresponding to the positioning locus up to the positioning position A. Since one branched road (road region) exists within the error range of the estimated position M, the estimated position M is updated to a position on the road. Further, a value corresponding to the positional deviation between the estimated position M and the updated estimated position may be added to the evaluation coefficient of the estimated position M and inherited as the updated estimated coefficient of the estimated position. The estimated position on the other road that is outside the error range of the estimated position M is rejected. Even in this case, the estimated position can be updated by correcting the displacement of the estimated position M at a characteristic point such as a curve or an intersection, and the evaluation coefficient and error range can be updated.

  FIG. 14 is an explanatory diagram showing another example of updating the estimated position. The example of FIG. 14 is also a case where the road where the estimated position updated last time (for example, the most recent time or two or three times before) exists may branch into two roads. Assume that there is one estimated position M corresponding to the positioning position A by the locus of the estimated position from the previously updated estimated position corresponding to the positioning locus up to the positioning position A. Since both branched roads (road areas) exist within the error range of the estimated position M, the estimated position M is updated to a position on each road. The updated estimated position is set as the estimated position of the first candidate and the estimated position of the second candidate. Further, a value corresponding to the positional deviation between the estimated position M and the estimated position of the first candidate and the estimated position of the second candidate is added to the evaluation coefficient of the estimated position M, and the estimated position of the first candidate and the second candidate You may make it inherit as an evaluation coefficient of an estimated position. Even in this case, the estimated position can be updated by correcting the displacement of the estimated position M at a characteristic point such as a curve or an intersection, and the evaluation coefficient and error range can be updated.

  Next, a method for calculating the evaluation coefficient of the estimated position at a characteristic point such as a curve or an intersection will be described. FIG. 15 is an explanatory diagram illustrating an example of calculation of an evaluation coefficient of an estimated position at a feature point. In the example of FIG. 15, it is assumed that there are estimated positions M1, M2, and M3 that are updated to positions on the road corresponding to the positioning positions A1, A2, and A3 at the characteristic points B1, B2, and B3 such as a curve. The positional deviation between the positioning position A1 and the estimated position M1 is (x1, y1), the positional deviation between the positioning position A2 and the estimated position M2 is (x2, y2), and the positional deviation between the positioning position A3 and the estimated position M3 is Let (x3, y3). In addition, the positional deviation between the estimated position updated last time (for example, the latest position or a plurality of previous times such as two times or three times) and the corresponding positioning position is (x0, y0). The evaluation coefficient of the estimated position (for example, estimated position M3) can be obtained as an average value of the difference in positional deviation between the estimated position and the positioning position.

  That is, the average value of the difference in positional deviation in the x direction is {| x1-x0 | + | x2-x1 | + | x3-x2 |} / 3, and the average value of the positional deviation in the y direction is { | Y1-y0 | + | y2-y1 | + | y3-y2 |} / 3. It can be said that as the positional deviation at each of the feature points B1, B2, and B3 is equal, the evaluation coefficient becomes smaller and the probability (probability) of the estimated position is larger. Thereby, it is possible to grasp how much the estimated position is relative to the positioning position.

  In addition, as an evaluation coefficient, the square of the difference of the position shift in each feature point is totaled and averaged, and the square root of the average value can be obtained. Further, instead of the average value, other statistical values such as a median value and a numerical value that is a half of the sum of the maximum value and the minimum value can be used.

  Next, the correction of the positioning direction will be described. When the road is identified and its estimated position is determined by the map matching method, and the positioning direction does not change more than a predetermined walking distance (for example, 20 m), the positioning direction can be corrected to the road direction on the map. it can. The timing for correcting the positioning direction is, for example, when a situation where the specified estimated position is unique continues for a predetermined time (for example, 1 minute) or a predetermined distance (for example, 50 m), or for all specified positions The situation where the road direction on the map corresponding to the estimated position is within a predetermined threshold (for example, 2 degrees) continues for a predetermined time (for example, 1 minute) and / or for a predetermined distance (for example, 50 m). In this case, the positioning azimuth does not change more than a predetermined walking distance (for example, 20 m).

  Further, the following correction timing conditions can be added. As the condition (1), if the difference between the estimated position and the positioning position is within a predetermined threshold (for example, 50 m) for a predetermined time (for example, 1 minute) or a predetermined distance (for example, 50 m), (2) When the probability that the geomagnetic sensor is normal for a predetermined time (for example, 1 minute) or a predetermined distance (for example, 50 m) is low, for example, continuous for a predetermined walking distance (for example, 20 m) or more When the geomagnetic sensor is not normal, as the condition (3), when the probability that the GPS 12 is normal for a predetermined time (for example, 1 minute) and / or a predetermined distance (for example, 50 m) is low, for example, This is a case where the GPS 12 is not normal continuously for a predetermined walking distance (for example, 50 m) or more. The correction of the positioning direction may be limited to the case where the road direction on the map is straight.

  Next, a display example for displaying the position of a pedestrian on a map will be described. When the position of the pedestrian is displayed on the output unit 17, when there is no estimated position, a positioning position or a temporary position is displayed. When there is an estimated position, the estimated position (including the updated estimated position) is displayed. .

  For example, when there is one estimated position, the estimated position (or including the updated estimated position) is displayed on the map, and the error range of the displayed estimated position is also displayed. When there are a plurality of estimated positions, one estimated position with the smallest evaluation coefficient can be displayed.

  Further, when there are a plurality of estimated positions, an area including the plurality of estimated positions is displayed as the estimated position. In this case, a range including the error range of each estimated position may be displayed. In addition, the estimated position and its error range can be displayed at the same time by expanding or reducing the size of the estimated position error range according to the size of the estimated position evaluation coefficient. In the case where there are a plurality of values, a range including an error range of each estimated position may be displayed. Moreover, you may display the gravity center position of several estimated position.

  When the position specified (detected) by updating the estimated position is displayed as the position of the pedestrian, if there are a plurality of specified positions, the most likely specific position is determined according to the magnitude of the calculated or corrected evaluation coefficient. It may be displayed, or a plurality of specific positions may be displayed, or some of the specified positions may be selected and displayed. Also, at a certain point in the process of detecting the position of the pedestrian, the position cannot be detected temporarily with high accuracy, and if the evaluation coefficient increases and the detected position is displayed, the pedestrian Even if there is a possibility that the wrong position is displayed, if the accuracy of the specific position is sufficiently secured as a result of the subsequent positioning, the position will be identified after the time when the certainty of the position can be secured. The position can also be displayed.

  Next, a procedure of position detection processing including posture specifying processing will be described. 16, 17 and 18 are flowcharts showing the procedure of the position detection process. The control unit 10 performs its own posture specifying process, moving direction specifying process, position detecting process, and the like in cooperation with each unit in the position detection apparatus 100. The control unit 10 determines whether or not a search for an initial position is necessary (S11).

  Unlike in the case of pedestrians, in the case of a vehicle, position detection is almost impossible, and even when the power is turned off, the detected position is saved, so there is almost no need to search for the initial position. There is nothing. However, in the case of a pedestrian, the position may not be detected when coming out of the basement and the initial position needs to be searched. The search for the initial position is usually performed by communication with the base station because positioning by the GPS 12 takes time.

  When the search for the initial position is necessary (YES in S11), the control unit 10 searches for the initial position (S12), and sets the search position as a temporary positioning position (S13). When the search for the initial position is not necessary (NO in S11), that is, when one or a plurality of estimated positions are already held, the control unit 10 does not perform the processes of Step S12 and Step S13, but will be described later in Step S14. Perform the process.

  The control unit 10 determines whether or not there is communication with the road device (S14), and when there is communication with the road device (YES in S14), the positioning position is corrected to the communication position (S15) and held estimation All positions are rejected (S16). The error due to narrow area communication (local communication) with the on-road device is considerably smaller than GPS etc., and the accuracy is high. Therefore, when the communication position is obtained by local communication, this communication position is the most reliable. It can be an estimated position.

  The control unit 10 registers the corrected positioning position as an estimated position (S17), and registers the attribute, error range, and evaluation coefficient of the map area of the estimated position (S18). Thereby, the new registration (initial registration) of the estimated position is completed, and the error range of the estimated position and the evaluation coefficient are set. When there is no communication with the road device (NO in S14), the control unit 10 performs the process of step S17 and registers the provisional positioning position as the estimated position.

  The control unit 10 acquires positioning data (S19) and specifies its own posture (S20). Details of the posture specifying process will be described later. The control unit 10 confirms whether there is an abnormality in the positioning data (S21), and calculates the reliability of the positioning data according to the confirmation result (S22). The control unit 10 obtains the movement distance and movement direction based on the positioning data, calculates the positioning position (S23), and calculates the positioning error of the positioning position (S24). The moving direction can be obtained by the direction specifying unit 14. As described above, the positioning position can be calculated by a combination of the self-contained navigation or map matching method using the distance sensor 21 and the direction specifying unit 14 and the satellite navigation using the GPS 12 or the like. However, if there are high obstacles such as buildings around and the positioning performance of the GPS satellite is not good, positioning can be performed only by the self-contained navigation or the map matching method.

  The control unit 10 acquires the walking behavior of the pedestrian (S25), and determines whether or not the map information and the signal information are updated from the external device (S26). When the map information and signal information are updated (YES in S26), the control unit 10 acquires map information and signal information (S27). When the map information and the signal information are not updated (NO in S26), the control unit 10 performs the process of step S28 described later without performing the process of step S27.

  The control unit 10 determines whether or not the map matching function is stopped (S28). The condition for determining whether or not the map matching function is stopped is, for example, when the walking speed of the pedestrian is larger than a predetermined threshold or when the position of the pedestrian is detected over a predetermined distance in the indoor area. . The predetermined threshold can be set to 60 km / h, for example, and when the walking speed becomes larger than the predetermined threshold, the pedestrian uses a vehicle such as a train or a car, or other transportation, and maps Stops pedestrian position detection by matching. Further, the predetermined distance can be set to 5 km, for example, and when the position of the pedestrian is detected over a predetermined distance in the indoor area, the pedestrian can use other transportation such as a train or a subway. Is used, the position detection of the pedestrian by map matching is stopped. Thereby, erroneous position detection can be prevented.

  When the map matching function is not stopped (NO in S28), the control unit 10 performs an update process of the estimated position (S29). After updating the estimated position, the control unit 10 determines the necessity of stopping the map matching function (S30), and if there is a necessity to stop (YES in S30), the control unit 10 stops the map matching function and estimates. All positions are rejected (S31).

  The control unit 10 determines whether or not the condition for initial registration (new registration) of the estimated position is satisfied (S32). If the condition is satisfied (YES in S32), the initial registration of the estimated position is performed based on the positioning position. (New registration) is performed (S33). When the map matching function is stopped (YES in S28), the control unit 10 performs the process of step S32. If there is no need to stop the map matching function (NO in S30), the control unit 10 performs a process in step S34 described later, and does not satisfy the conditions for initial registration (new registration) of the estimated position (NO in S32). ), Similarly, the process of step S34 described later is performed.

  The control unit 10 determines the presence / absence of the estimated position (including the updated estimated position) (S34). If there is an estimated position (YES in S34), the estimated position is displayed on the output unit 17 (S35). If there is no (NO in S34), the positioning position or the temporary position is displayed on the output unit 17 (S36).

  The control unit 10 determines whether or not the positioning azimuth correction condition is satisfied (S37). When the condition is satisfied (YES in S37), the positioning azimuth is corrected to the road direction on the map (S38). When the positioning azimuth correction condition is not satisfied (NO in S37), the control unit 10 performs the process of step S39 described later without performing the process of step S38.

  The controller 10 determines whether or not the sensor needs to be calibrated (S39). If calibration is necessary (YES in S39), the sensor is calibrated (S40). When sensor calibration is not necessary (NO in S39), the control unit 10 performs the process of step S41 described later without performing the process of step S40. The control unit 10 determines whether or not there is an instruction to end the process (S41). If there is no instruction (NO in S41), the process from step S11 is continued, and if there is an instruction (YES in S41), the process ends. .

  Next, the posture specifying process will be described. FIG. 19 is a flowchart showing the procedure of the posture specifying process. The control unit 10 determines whether or not it is necessary to specify the posture (S201), and when it is not necessary to specify the posture (NO in S201), the process ends. When it is necessary to specify the posture (YES in S201), the control unit 10 acquires data of the triaxial acceleration sensor 22 (S202), and determines whether the posture can be specified (S203).

  If the posture can be specified (YES in S203), that is, if there is no change in the posture of the position detection device 100, or if there is no change in the walking characteristics of the pedestrian, the control unit 10 has collected sufficient data. It is determined whether or not (S204).

  If sufficient data has not been collected (NO in S204), the control unit 10 continues the process from step S202 and collects data. If sufficient data has been collected (YES in S204), the control unit 10 averages the data and calculates the standard value of the acceleration data of each axis (X axis, Y axis, Z axis) (S205). .

  The control unit 10 calculates the acceleration variable of each axis (X axis, Y axis, Z axis) from the data of the triaxial acceleration sensor 22 (S206). The control unit 10 calculates the pitch angle α and the roll angle β from the standard value of the acceleration data (S207), and calculates the yaw angle γ from the acceleration variable (S208).

  The control unit 10 compares the pitch angle α, roll angle β, and yaw angle γ calculated this time with the previously calculated pitch angle α, roll angle β, and yaw angle γ, and averages the current data according to the comparison result. Alternatively, it is rejected (invalidated) (S209), and the process is terminated. That is, when it is determined that the reliability of the data calculated this time is low, or when it is determined that the difference from the past data is large, the control unit 10 rejects (invalidates) the current data.

  When the posture cannot be specified (NO in S203), the control unit 10 discards the acquired data (S210) and determines whether or not a predetermined time has elapsed or a predetermined distance has been moved (S211).

  If the predetermined time has not elapsed or the predetermined distance has not been moved (NO in S211), the control unit 10 continues the processing from step S202. When the predetermined time has elapsed or the predetermined distance has been moved (YES in S211), the control unit 10 determines that the posture cannot be specified continues, stops the posture specification (S212), and performs the process. finish.

  As described above, according to the present invention, by using the three-axis acceleration sensor built in the portable device, the posture of the portable device can be specified even when the pedestrian carries it. Further, by using a triaxial geomagnetic sensor, the walking direction of the pedestrian can be specified. As a result, the walking direction of the pedestrian can be accurately identified regardless of the posture of the mobile device, so that even in an environment where GPS cannot be used, a pedestrian based on autonomous navigation can be used. Can be detected. Moreover, since the posture of the portable device can be specified, when the distance sensor or the acceleration sensor is used when obtaining the movement distance or position of the pedestrian carrying the portable device, the movement distance or acceleration is accurately obtained. Can do.

  In the above-described embodiment, some examples have been shown as methods for specifying the position of a pedestrian by associating the walking behavior and reliability of positioning data with the map information together with the physical position or orientation. In addition, for example, when you get on an elevator or escalator, specify the location of the escalator (walking stop, change in altitude or pressure, taking into account the reliability of positioning data), if you get off at the underground city or subway connection To the contact point (considering changes in walking speed and walking strength, changes in altitude or atmospheric pressure, reliability of positioning data), identification at the ticket vending machine position when stopping at the ticket vending machine in the subway station building (Walking stop, taking into account the reliability of positioning data), specifying a stop (walking stop, change in position), etc. when you get on a bus or streetcar at a stop, register a specific position such as a public object in the map information Leave It can become.

  In the above example, the map area attribute, position detection, position detection error range, etc. are all expressed in two dimensions to simplify the expression. However, even if altitude information is included, it can be expressed in three dimensions. Good. This can improve the accuracy of pedestrian overpasses and underground crossing passages that are related to altitude, and it is also possible to estimate whether it is on the floor of a building in an indoor area, whether it is on the rooftop, etc. It becomes. In addition, the above description describes a case where a pedestrian wears a portable device during normal walking or bicycle riding, but the present invention is not limited to this, and the pedestrian directly Put your bag or mobile device in a travel case with wheels, carts, prams, wheelchairs, etc., temporarily install or temporarily place it, carry it by a pedestrian, push or pull the car, or pull the wheel by hand. It may be a case where it is rotating and passing through an area where a pedestrian can pass.

  In addition, in the above, position detection is stopped when a pedestrian gets on a train or a car, but a map matching type position detection system is prepared separately when a person gets on a train or gets on a car. Thus, the position detection may be continuously performed. Thereby, it is not necessary to perform position detection from the beginning, and a more effective system can be constructed. Furthermore, in the above example, the form in which all the data necessary for position detection is collected and detected in the mobile device is shown. However, the present invention is not limited to this. For example, the server may execute the position detection process on the server, and the result may be transmitted to the mobile device. Or you may share execution of a process. This can reduce the burden on the mobile device.

  The position detection device described above can be applied to an information terminal device such as a mobile phone, a PDA (Personal Digital Assistant), a PHS, a notebook personal computer, a music player, and a mobile game device, or a mobile terminal device.

  The mathematical formulas for estimating the position of the pedestrian shown in the above-described embodiment are merely examples, and the mathematical formulas are not limited thereto, and mathematical formulas appropriately modified can be used.

  The disclosed embodiments are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a block diagram which shows an example of a structure of the position detection apparatus as a position specific device based on this invention. It is explanatory drawing which shows the coordinate system for specifying an attitude | position. It is explanatory drawing which shows an example of the acceleration data detected with the 3-axis acceleration sensor at the time of a walk. It is explanatory drawing which shows an example of the acceleration component of a reference | standard orthogonal three-dimensional coordinate. It is explanatory drawing which shows the example of the error range of a positioning position. It is a schematic diagram which shows an example of map information. It is explanatory drawing which shows an example of the attribute of the area | region on a map. It is explanatory drawing which shows the example of a setting of the road on a map. It is explanatory drawing which shows an example of the new registration of an estimated position. It is explanatory drawing which shows an example of calculation of the evaluation coefficient of the newly registered estimated position. It is explanatory drawing which shows the example of calculation of the locus | trajectory of an estimated position. It is explanatory drawing which shows an example of the update of an estimated position. It is explanatory drawing which shows the other example of the update of an estimated position. It is explanatory drawing which shows the other example of the update of an estimated position. It is explanatory drawing which shows an example of calculation of the evaluation coefficient of the estimated position in a feature point. It is a flowchart which shows the procedure of a position detection process. It is a flowchart which shows the procedure of a position detection process. It is a flowchart which shows the procedure of a position detection process. It is a flowchart which shows the procedure of an attitude | position specific process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Position detection apparatus 10 Control part 11 Communication part 12 GPS
DESCRIPTION OF SYMBOLS 13 Map database 14 Orientation specific part 15 Memory | storage part 16 Operation part 17 Output part 20 Positioning part 21 Distance sensor 22 3-axis acceleration sensor 23 3-axis geomagnetic sensor 24 Direction correction part 25 Sensor calibration part 30 Attitude specification part 31 Coordinate conversion part 32 Angle Calculation unit 33 Determination unit 40 Position detection processing unit 41 Position estimation unit 42 Error calculation unit 43 Attribute determination unit 44 Walking behavior determination unit 45 Position update unit 46 Reliability calculation unit 47 Evaluation unit

Claims (13)

A posture identifying device that identifies its posture with respect to a coordinate axis of a reference orthogonal three-dimensional coordinate,
Conversion means for performing coordinate conversion between an apparatus orthogonal three-dimensional coordinate having an axis that forms an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinate and the reference orthogonal three-dimensional coordinate;
A 3-axis acceleration sensor;
Angle conversion means for calculating the angle using coordinate conversion performed by the conversion means and acceleration components in the apparatus orthogonal three-dimensional coordinates detected by the three-axis acceleration sensor ;
Storage means for storing information indicating a relationship between acceleration components of the respective axes in the reference orthogonal three-dimensional coordinates;
Variable amount calculating means for calculating an acceleration variable according to an added value of acceleration detected at an appropriate time or an appropriate number of times by the three-axis acceleration sensor ;
The angle calculation means includes
The acceleration variable calculated by the variable calculation means and the information stored in the storage means are used for coordinate conversion performed by the conversion means,
Posture identifying apparatus characterized by is arranged to identify its orientation by the angle calculated by the angle calculating means. The angle calculation means includes
The posture identifying device according to claim 1, characterized in that are configured to use the gravitational acceleration at the reference orthogonal three-dimensional coordinate in the coordinate conversion performed by the conversion means. It can be carried by pedestrians,
In response to changes in gait characteristic of the mobile during its posture change or pedestrian, according to claim 1 or claim 2, characterized in that it comprises determining means for determining whether to identify its orientation Posture identification device. An acceleration comparison means for comparing the absolute value of the acceleration detected by the three-axis acceleration sensor with a predetermined acceleration threshold;
The determination means includes
4. The posture identifying apparatus according to claim 3 , wherein when the absolute value is larger than the acceleration threshold, it is determined that the posture is not specified. Detection means for detecting a change in the walking cycle, the number of steps, or the walking speed based on a change in acceleration detected by the three-axis acceleration sensor,
The determination means includes
4. The posture identifying apparatus according to claim 3 , wherein when a change is detected by the detecting means, it is determined that the posture of the device is not specified. The determination means includes
6. The posture identifying apparatus according to claim 5 , wherein when the walking stop is detected by the detecting means, it is determined that the posture is not specified. An orientation sensor;
An azimuth angle comparing means for comparing the azimuth angle detected by the azimuth sensor with a predetermined azimuth angle threshold;
The determination means includes
4. The posture identifying device according to claim 3 , wherein when the azimuth angle is larger than the azimuth angle threshold value, it is determined that the posture of itself is not identified. An evaluation unit that evaluates the reliability of the angle calculated by the angle calculation unit using coordinate conversion performed by the conversion unit;
The posture specifying device according to any one of claims 1 to 7 , further comprising: an invalidating unit that invalidates the calculated angle when the evaluation unit evaluates that the reliability is low. Storage means for storing the angles calculated by the angle calculation means over a plurality of time points;
Angle difference calculating means for calculating an angle difference between the angle calculated by the angle calculating means and the angle stored by the storage means;
If the angle difference calculated by the angle difference calculating means is larger than the predetermined threshold value, any one of claims 1 to 7, characterized in that it comprises a disabling means for disabling the angle calculated by said angle calculating means The attitude | position identification apparatus as described in one. The posture identifying device according to any one of claims 1 to 9 ,
A triaxial geomagnetic sensor;
The coordinate transformation performed by the transformation means is specified by the two axes of the reference orthogonal three-dimensional coordinates using the geomagnetic component in the apparatus orthogonal three-dimensional coordinates detected by the three-axis geomagnetic sensor and the angle calculated by the angle calculation means. A geomagnetism calculating means for calculating a geomagnetic component of each axis on the reference plane;
A moving direction specifying device comprising: direction specifying means for specifying its moving direction based on a geomagnetic component calculated by the geomagnetic calculating means. The moving direction specifying device according to claim 10 ,
A distance detecting means for detecting a moving distance;
A position specifying device comprising: position specifying means for specifying its own position based on the moving distance detected by the distance detecting means and the moving direction specified by the moving direction specifying device. A computer program for causing a computer to function as means for specifying an attitude of a device with respect to a coordinate axis of a reference orthogonal three-dimensional coordinate,
Computer
Conversion means for performing coordinate conversion between an apparatus orthogonal three-dimensional coordinate having an axis that forms an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinate and the reference orthogonal three-dimensional coordinate;
Angle conversion means for calculating the angle using coordinate conversion performed by the conversion means and acceleration components in the apparatus orthogonal three-dimensional coordinates detected by a three-axis acceleration sensor;
A variable calculation means for calculating an acceleration variable according to an added value of acceleration detected for an appropriate time or an appropriate number of times by the three-axis acceleration sensor;
To function,
The angle calculation means includes
Information indicating the relationship between the acceleration variable calculated by the variable calculation means and the acceleration component of each axis in the reference orthogonal three-dimensional coordinates is used for coordinate conversion performed by the conversion means;
Furthermore, the computer program is made to function as a specifying means for specifying its own posture based on the calculated angle. A posture specifying method for specifying a posture with a posture specifying device for specifying its own posture with respect to a coordinate axis of a reference orthogonal three-dimensional coordinate,
The posture specifying device includes:
Performing coordinate conversion between the apparatus orthogonal three-dimensional coordinates having an axis that forms an arbitrary angle with each axis of the reference orthogonal three-dimensional coordinates and the reference orthogonal three-dimensional coordinates;
The angle is calculated using the acceleration component in the apparatus orthogonal three-dimensional coordinates detected by the three-axis acceleration sensor for the coordinate conversion,
Calculating an acceleration variable according to an added value of acceleration detected at an appropriate time or an appropriate number of times by the three-axis acceleration sensor;
Information indicating the relationship between the calculated acceleration variable and the acceleration component of each axis in the reference orthogonal three-dimensional coordinates is used for the coordinate conversion,
A posture identifying method characterized by identifying one's posture based on a calculated angle.

Images