JP6287804B2 - Direction estimation apparatus and direction estimation method - Google Patents

Description

  The present invention relates to an azimuth estimation technique, and more particularly to an azimuth estimation apparatus and an azimuth estimation method that estimate an azimuth after correcting an output signal from an angular velocity sensor.

In a vehicular navigation apparatus, generally, an optimum position is estimated by combining a position calculated from a self-contained navigation and a position calculated from a GPS (Global Positioning System). In the self-contained navigation, the current position is calculated by updating the previous positioning position based on the speed pulse indicating the speed of the vehicle and the turning angular speed of the vehicle measured by the angular speed sensor. According to such a type of navigation device, the position of the vehicle can be derived by self-contained navigation even in tunnels, underground parking lots and valleys of high-rise buildings where it is difficult to receive radio waves from GPS satellites. The angular velocity ω due to turning of the vehicle is derived by the following equation.
ω = (Vout−Voffset) / (S · cos α) (1)
Here, Vout is an output voltage of the angular velocity sensor, Voffset is an offset value of the angular velocity sensor, S (mV / deg / sec) is a sensitivity coefficient of the angular velocity sensor, and α (deg) is an angular velocity sensor with respect to the vertical axis. This is the inclination of the detection axis.

  In order to accurately determine the angular velocity, it is necessary to accurately determine the offset value of the angular velocity sensor and the sensitivity coefficient. The sensitivity coefficient of the angular velocity sensor generally differs depending on the individual difference of the angular velocity sensor and the influence of aging deterioration. In addition, the offset value may change due to a temperature change. In other words, the offset value is affected by temperature rise due to heat generated by the substrate used in the vehicle navigation device or heat generated by the vehicle engine when the vehicle navigation device is attached to the dashboard of the vehicle. Conventionally, the offset value of the angular velocity sensor has been corrected by using the output voltage from the angular velocity sensor when the vehicle is stopped or is traveling straight ahead when the angular velocity is “0”. However, it is difficult to periodically correct the offset value of the angular velocity sensor in cases where the frequency of stoppage of the vehicle is low, such as traveling on a highway or traveling for a long time in an area with low traffic volume. Is likely to get worse. The offset value correction during straight running so that the output voltage from the angular velocity sensor accurately becomes “0” is also easily influenced by the road shape and the driving situation of the driver, and thus it is difficult to perform it periodically. The sensitivity coefficient of the angular velocity sensor is derived from the azimuth change amount within the unit period and the output voltage of the angular velocity sensor. Therefore, as is apparent from the equation (1), the sensitivity coefficient of the angular velocity sensor is affected by the error of the offset value.

There has been proposed a technique for correcting an offset and a sensitivity coefficient of an angular velocity sensor even during traveling other than straight traveling. There, the offset value and sensitivity coefficient of the angular velocity sensor are corrected based on the average value of the output voltage of the angular velocity sensor during a predetermined period and the amount of change in the direction of the vehicle during the period during which the average value is calculated. More specifically, the offset value Voffset of the angular velocity sensor is derived as follows.
Voffset = 1 / n · ΣVout−1 / Δt · Δθ / n · S · cosα (2)
Here, n is the number of samples of the output voltage of the angular velocity sensor, Δt (sec) is the sampling interval, Δθ (deg) is the azimuth change amount, and α is the detection of the angular velocity sensor with respect to the vertical axis. It is the inclination of the axis. The azimuth change amount is obtained based on the GPS azimuth obtained from the GPS satellite or map data. The inclination of the detection axis of the angular velocity sensor with respect to the vertical axis is obtained by adding the mounting angle of the angular velocity sensor and the road inclination angle. The mounting angle of the angular velocity sensor is obtained from an acceleration sensor, and the road inclination angle is obtained from a GPS altitude change obtained from a GPS satellite. Further, the sensitivity coefficient of the angular velocity sensor is derived as follows when the amount of change in the corrected offset value is small, that is, in a stable state.
S = (1 / n · ΣVout−Voffset) · n / Δθ · Δt / cosα (3)
Here, Voffset is known and is a constant in a stable state where the amount of change in the corrected offset value is small (see, for example, Patent Document 1).

JP 2001-330454 A

  Under such circumstances, in order to improve the position estimation accuracy of the vehicle navigation device, it is desired to correct the output signal of the angular velocity sensor with high accuracy.

  The present invention has been made in view of such circumstances, and an object thereof is to provide a technique for correcting an output signal of an angular velocity sensor with high accuracy.

  In order to solve the above-described problem, an azimuth estimation apparatus according to an aspect of the present invention includes positioning data of an object measured based on a signal from a GPS satellite and an angular velocity of the object output from an angular velocity sensor. Based on the acquisition unit to be acquired, the sensitivity coefficient calculation unit for deriving the sensitivity coefficient of the angular velocity sensor based on the positioning data and angular velocity acquired in the acquisition unit, and the positioning data and angular velocity acquired in the acquisition unit, Acquired by the acquisition unit based on the offset value calculation unit for deriving the offset value of the angular velocity sensor, the offset value of the angular velocity sensor derived by the offset value calculation unit, and the sensitivity coefficient of the angular velocity sensor derived by the sensitivity coefficient calculation unit An angular velocity conversion unit that corrects the measured angular velocity, a correlation coefficient deriving unit that derives a correlation coefficient based on the positioning data and the angular velocity acquired by the acquisition unit, Based on the correlation coefficient derived in the coefficient deriving unit and the effectiveness of the positioning data acquired in the acquiring unit, the amount of change in the azimuth included in the positioning data acquired in the acquiring unit and the angular velocity conversion unit corrected A determining unit that determines a ratio when combining the angular velocity, and a change amount of the azimuth included in the positioning data acquired by the acquiring unit and an angular velocity corrected by the angular velocity converting unit according to the ratio determined by the determining unit. An update unit that combines and updates the azimuth according to the combined value.

  Another aspect of the present invention is a direction estimation method. In this method, the positioning data of an object measured based on a signal from a GPS satellite, the angular velocity of the object output from an angular velocity sensor, and the obtained positioning data and angular velocity are used. In addition, the step of deriving the sensitivity coefficient of the angular velocity sensor, the step of deriving the offset value of the angular velocity sensor based on the obtained positioning data and the angular velocity, the offset value of the derived angular velocity sensor, and the derived angular velocity sensor The step of correcting the acquired angular velocity based on the sensitivity coefficient, the step of deriving the correlation coefficient based on the acquired positioning data and the angular velocity, the derived correlation coefficient, and the obtained positioning data Based on the effectiveness, the step of determining the ratio when combining the amount of change in azimuth included in the acquired positioning data and the corrected angular velocity, and the determined ratio Comprising the acquired amount of change in azimuth angle that is included in the positioning data, and synthesizes the corrected angular velocity, and a step of updating the azimuth by synthesized value.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  According to the present invention, the output signal of the angular velocity sensor can be corrected with high accuracy.

It is a figure which shows the structure of the angular velocity calculation apparatus which concerns on Example 1 of this invention. It is a figure which shows the structure of the offset value calculating part of FIG. It is a figure which shows the structure of the offset value filter process part of FIG. It is a figure which shows the data structure of the table memorize | stored in the forgetting factor control part of FIG. It is a figure which shows the structure of the sensitivity coefficient calculating part of FIG. It is a figure which shows the data structure of the table memorize | stored in the forgetting factor control part of FIG. It is a figure which shows the outline | summary of the correlation coefficient calculated in the correlation coefficient calculating part of FIG. It is a flowchart which shows the derivation | leading-out procedure of the forgetting factor by the angular velocity calculation apparatus of FIG. It is a figure which shows the structure of the angular velocity calculation apparatus which concerns on Example 2 of this invention. It is a figure which shows the structure of the offset value calculating part of FIG. It is a figure which shows the structure of the offset value filter process part of FIG. It is a figure which shows the data structure of the table memorize | stored in the forgetting factor control part of FIG. It is a figure which shows the outline | summary of the correlation coefficient calculated in the correlation coefficient calculating part of FIG. It is a figure which shows the structure of the azimuth | direction estimation apparatus which concerns on Example 3 of this invention. It is a figure which shows the data structure of the table memorize | stored in the GPS effectiveness determination part of FIG. It is a figure which shows the data structure of the table memorize | stored in the synthetic | combination ratio determination part of FIG.

Example 1
Before describing the present invention specifically, an outline will be given first. An embodiment of the present invention relates to an angular velocity calculation device that is mounted on a vehicle or the like and derives an angular velocity due to turning of the vehicle. The angular velocity calculation device derives an angular velocity for the output voltage from the angular velocity sensor while using the offset value and the sensitivity coefficient. In Expression (3) for deriving the sensitivity coefficient of the angular velocity sensor, if Voffset is not stabilized due to the influence of a temperature change or the like, the error of the sensitivity coefficient may increase when the number of samples is increased. Further, in the equation (3) for deriving the sensitivity coefficient of the angular velocity sensor, when the azimuth change amount Δθ is obtained from the GPS azimuth obtained from the GPS satellite, the accuracy of the GPS azimuth depends on the reception status of the radio wave from the GPS satellite. It may deteriorate, and the error included in the sensitivity coefficient becomes large.

  Further, in the equation (3), when the direction change amount Δθ is obtained based on the map data, the road direction based on the map data and the traveling direction of the vehicle do not always coincide with each other, and thus are included in the sensitivity coefficient. The error can be large. Further, in the equation (3) for deriving the sensitivity coefficient of the angular velocity sensor, when the road inclination angle included in the inclination of the detection axis of the angular velocity sensor with respect to the vertical axis is obtained from the GPS altitude change amount acquired from the GPS satellite, The accuracy of the GPS azimuth may deteriorate due to the reception status of radio waves from the GPS satellite, and the error included in the sensitivity coefficient increases.

  In these cases, in order to improve the accuracy of deriving the angular velocity, it is required to reduce the error included in the sensitivity coefficient. Therefore, the angular velocity calculating apparatus according to the present embodiment uses the forgetting coefficient while changing the weighted average for the past value and the current value of the sensitivity of the angular velocity sensor that are sequentially derived. More specifically, the angular velocity calculation device calculates a correlation coefficient between the azimuth change calculated from the angular velocity sensor and the azimuth change calculated from GPS in a predetermined period, and changes the forgetting coefficient based on the correlation coefficient. To do. The larger the correlation coefficient, the closer the azimuth change calculated from different means such as the angular velocity sensor and GPS, and the higher the measurement accuracy of the azimuth change. Therefore, in such a case, the weighted average is calculated so as to increase the influence of the current value of the sensitivity of the angular velocity sensor. On the other hand, when the correlation coefficient is small, the measurement accuracy of the azimuth change is low, so the weighted average is calculated so that the influence of the current value of the sensitivity of the angular velocity sensor is small.

  FIG. 1 shows a configuration of an angular velocity calculation apparatus 100 according to the first embodiment of the present invention. The angular velocity calculation apparatus 100 includes a measurement unit 10, a parameter calculation unit 12, an angular velocity conversion unit 14, and a control unit 38. The measurement unit 10 includes a GPS positioning unit 20, an effectiveness determination unit 22, an attachment angle detection unit 24, an inclination angle detection unit 26, and an angular velocity sensor 28. The parameter calculation unit 12 includes an offset value calculation unit 30, a sensitivity coefficient. A calculation unit 32, a correlation coefficient calculation unit 34, and a positioning data storage unit 36 are included. Further, GPS positioning data 200, mounting angle 202, tilt angle 204, output signal 206, offset value 208, sensitivity coefficient 210, correlation coefficient 212, and storage data 214 are included as signals. The angular velocity calculation device 100 is installed in the vehicle with a predetermined mounting angle inclination.

  The GPS positioning unit 20 receives a signal from a GPS satellite (not shown) and calculates GPS positioning data 200. The GPS positioning data 200 includes the longitude and latitude, the GPS altitude that is the altitude of the vehicle, the GPS speed that is the moving speed, the GPS direction that is the direction of the vehicle, the PDOP (Position Division Precision), the number of captured satellites, and the like. Here, PDOP is an index of how the error of the GPS satellite position in the GPS positioning data 200 is reflected in the reception point position, and corresponds to the positioning error. Note that the GPS positioning data 200 may include values other than these. Further, since the GPS positioning data 200 may be calculated by a known technique, description thereof is omitted here. The GPS positioning unit 20 calculates the GPS positioning data 200 at every sampling interval, that is, periodically. The GPS positioning unit 20 sequentially outputs the GPS positioning data 200 to the validity determination unit 22.

  The validity determination unit 22 sequentially inputs the GPS positioning data 200 from the GPS positioning unit 20. The validity determination unit 22 determines the validity of each GPS positioning data 200 from the GPS positioning data 200. For example, when the value of PDOP is equal to or less than a first threshold value and the GPS speed is equal to or greater than a second threshold value, the validity determination unit 22 determines that the GPS direction corresponding to them is valid. judge. The validity determination unit 22 determines that the corresponding GPS azimuth is invalid when the above condition is not satisfied. This is because the accuracy of the GPS orientation tends to be low when the PDOP value is large or the GPS speed is low. More specifically, when the value of PDOP is 6 or less and the GPS speed is 20 km / h or more, the validity determination unit 22 represents the validity of the GPS direction with a flag.

  Moreover, the effectiveness determination part 22 determines with the said GPS speed being effective, when a GPS speed is more than a 3rd threshold value. Here, the third threshold value may be the same as the second threshold value. Further, the validity determination unit 22 determines that the GPS altitude is valid when the difference in GPS altitude in the predetermined period is equal to or smaller than the fourth threshold value. As a result of such processing, the validity determination unit 22 adds a flag indicating validity or invalidity to each value such as GPS azimuth included in the GPS positioning data 200 (hereinafter, a flag is added). GPS positioning data 200 is also referred to as “GPS positioning data 200”). The validity determination unit 22 sequentially outputs the GPS positioning data 200 to the attachment angle detection unit 24, the inclination angle detection unit 26, the offset value calculation unit 30, the sensitivity coefficient calculation unit 32, and the positioning data storage unit 36.

  The mounting angle detection unit 24 is, for example, an acceleration sensor (not shown) and the like, and sequentially receives the GPS positioning data 200 from the validity determination unit 22, and determines the inclination of the detection axis of the angular velocity sensor with respect to the vertical axis when the vehicle is in a horizontal state. Calculated as the mounting angle 202. The mounting angle detection unit 24 determines that the vehicle is in a horizontal state when a change in the GPS altitude of the input GPS positioning data 200 is minute, and at this time, the acceleration detected when the vehicle starts from a stop is detected. Based on this, the mounting angle 202 is calculated. In addition, since a well-known technique should just be used for an acceleration sensor apparatus, description is abbreviate | omitted here. The attachment angle detection unit 24 outputs the detected attachment angle 202 to the inclination angle detection unit 26.

  The inclination angle detection unit 26 sequentially inputs the GPS positioning data 200 from the validity determination unit 22, particularly the GPS altitude included in the GPS positioning data 200 and the attachment angle 202 from the attachment angle detection unit 24. The inclination angle detection unit 26 detects an average inclination angle of the vehicle (hereinafter referred to as “inclination angle 204”) at the sampling interval based on the sequentially input GPS altitude. More specifically, the tilt angle detection unit 26 sequentially calculates the difference between successive GPS altitudes, averages the calculation results, and then divides the average value by the sampling interval to derive the vehicle tilt angle. To do. Here, the interval between successive GPS altitudes corresponds to the sampling interval. The tilt angle detection unit 26 adds the mounting angle 202 to the calculated vehicle tilt angle, and sequentially outputs the tilt angle 204 to the offset value calculation unit 30, the sensitivity coefficient calculation unit 32, and the positioning data storage unit 36.

  The angular velocity sensor 28 corresponds to, for example, a gyro apparatus such as a vibration gyro, and detects a change in the traveling direction of the vehicle as a relative angular change of the vehicle. That is, the angular velocity sensor 28 detects the turning angular velocity of the vehicle. The detected angular velocity is output as an analog signal of 0V to 5V, for example. At that time, the positive angular velocity corresponding to the clockwise turn is output as a deviation voltage from 2.5V toward 5V, and the negative angular velocity corresponding to the counterclockwise turn is from 2.5V toward 0V. Output as deviation voltage. Further, 2.5V is an offset value of the angular velocity, that is, a zero point, and drifts under the influence of temperature and the like.

  The sensitivity coefficient (mV / deg / sec), which is about the deviation of angular velocity from 2.5V, is determined as a predetermined value that falls within the allowable error in a horizontal state. The cause of this allowable error is the individual difference of the gyro device, the secular change, the influence of temperature, and the like. The voltage value of the gyro device is AD converted by an AD (Analog to Digital) converter (not shown), for example, at a sampling interval of 100 msec, and the resulting digital signal is output. The digital signal corresponds to the aforementioned output voltage, and the term output signal 206 will be used below. In addition, since a well-known technique should just be used as a gyro apparatus, description is abbreviate | omitted here. The angular velocity sensor 28 outputs an output signal 206 to the offset value calculation unit 30, the sensitivity coefficient calculation unit 32, the positioning data storage unit 36, and the angular velocity conversion unit 14.

  The offset value calculation unit 30 receives the GPS positioning data 200 from the validity determination unit 22, the tilt angle 204 from the tilt angle detection unit 26, and the output signal 206 from the angular velocity sensor 28. The offset value calculation unit 30 also receives the sensitivity coefficient 210 from the sensitivity coefficient calculation unit 32. The offset value calculator 30 calculates an offset value (hereinafter referred to as “offset value 208”) of the angular velocity sensor 28 based on the GPS positioning data 200, the tilt angle 204, the output signal 206, and the sensitivity coefficient 210. Details of processing in the offset value calculation unit 30 will be described later. The offset value calculation unit 30 outputs the offset value 208 to the angular velocity conversion unit 14, the sensitivity coefficient calculation unit 32, and the correlation coefficient calculation unit 34.

  The sensitivity coefficient calculation unit 32 inputs the GPS positioning data 200 from the validity determination unit 22, the tilt angle 204 from the tilt angle detection unit 26, and the output signal 206 from the angular velocity sensor 28. The sensitivity coefficient calculation unit 32 also receives the offset value 208 from the offset value calculation unit 30 and the correlation coefficient 212 from the correlation coefficient calculation unit 34. The sensitivity coefficient calculation unit 32 is based on the GPS positioning data 200, the output signal 206, the tilt angle 204, and the offset value 208 that are input for a predetermined period, for example, 10 seconds. (Referred to as “sensitivity coefficient 210”). Details of processing in the sensitivity coefficient calculation unit 32, particularly processing using the correlation coefficient 212, will be described later. The sensitivity coefficient calculation unit 32 outputs the sensitivity coefficient 210 to the angular velocity conversion unit 14 and the offset value calculation unit 30.

  The positioning data storage unit 36 is configured by a ring buffer or the like on the memory, and sequentially receives the GPS positioning data 200 from the validity determination unit 22, the tilt angle 204 from the tilt angle detection unit 26, and the output signal 206 from the angular velocity sensor 28. input. The positioning data storage unit 36 acquires the GPS positioning data 200 determined to be valid by the validity determination unit 22, and is continuously valid for a predetermined period, for example, the period during which the sensitivity coefficient calculation unit 32 calculates the sensitivity coefficient 210. If it is, the GPS positioning data 200, the tilt angle 204, and the output signal 206 are stored as the stored data 214, and the validity flag of the stored data is validated. The stored data 214 is updated every time valid GPS positioning data is input. If the input GPS positioning data 200 is invalid, the positioning data storage unit 36 clears the stored data 214 and invalidates the validity flag. The positioning data storage unit 36 adds an validity flag to the stored storage data 214 and outputs it to the correlation coefficient calculation unit 34.

  The correlation coefficient calculation unit 34 sequentially inputs the offset value 208 from the offset value calculation unit 30 and the storage data 214 from the positioning data storage unit 36. When the validity flag added to the stored data 214 is valid, the correlation coefficient calculator 34 divides the difference between the integrated value of the output signal 206 and the offset value 208 by the cosine value of the inclination angle 204, and The correlation coefficient 212 is calculated for a GPS azimuth change amount included in the GPS positioning data 200 during a predetermined period, for example, a period during which the sensitivity coefficient calculation unit 32 calculates the sensitivity coefficient 210. Details of processing in the correlation coefficient calculation unit 34 will be described later. The correlation coefficient calculator 34 outputs the correlation coefficient 212 to the sensitivity coefficient calculator 32.

  The angular velocity converter 14 receives an output signal 206 from the angular velocity sensor 28, an inclination angle 204 from the inclination angle detector 26, an offset value 208 from the offset value calculator 30, and a sensitivity coefficient 210 from the sensitivity coefficient calculator 32. . The angular velocity conversion unit 14 calculates the angular velocity ω of the vehicle by calculating the above equation (1) based on the output signal 206, the tilt angle 204, the offset value 208, and the sensitivity coefficient 210. The control unit 38 controls the operation of the angular velocity calculation device 100 as a whole.

  This configuration can be realized in terms of hardware by a CPU, memory, or other LSI of any computer, and in terms of software, it can be realized by a program loaded in the memory, but here it is realized by their cooperation. Draw functional blocks. Accordingly, those skilled in the art will understand that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof.

  FIG. 2 shows a configuration of the offset value calculation unit 30. The offset value calculation unit 30 includes a state estimation unit 40, a state-specific offset value derivation unit 42, and an offset value filter processing unit 44. The state estimation unit 40 includes a stop estimation unit 50, a straight travel estimation unit 52, and a non-straight travel estimation unit 54. The state-specific offset value deriving unit 42 includes a stop offset value deriving unit 56, and a straight travel offset value. A deriving unit 58 and an offset value deriving unit 60 during non-straight running are included. Further, a temporary offset value 216 and traveling state information 218 are included as signals.

  The state estimation unit 40 receives the GPS positioning data 200, the tilt angle 204, and the output signal 206. The state estimation unit 40 estimates the traveling state of the vehicle in the stop estimation unit 50, the straight travel estimation unit 52, and the non-straight travel estimation unit 54. Here, as the running state of the vehicle, it is estimated whether the vehicle is in a stopped state or going straight, or the remaining state, that is, a non-straight running state. Further, the state estimation unit 40 outputs the determination result as the traveling state information 218 to the offset value filter processing unit 44.

  The stop estimation unit 50 acquires the GPS positioning data 200 determined to be effective by the effectiveness determination unit 22 (not shown). Further, the stop estimation unit 50 extracts the GPS speed from the GPS positioning data 200, and confirms whether the GPS speed is “0”. On the other hand, the stop estimation unit 50 calculates a variance value of the output signal 206 within a predetermined period, and compares the variance value with a fifth threshold value. The stop estimation unit 50 determines that the vehicle is in a stopped state when the GPS speed is 0 and the variance value is smaller than the fifth threshold value. As described above, when the GPS speed is low, the accuracy tends to be low. However, the stop estimation unit 50 determines that the stop has occurred by using the variance value of the output signal 206 together. Here, the predetermined period is, for example, 1 sec which is a sampling interval of the GPS speed. When the variance value of the output signal 206 is small during the predetermined period, it is estimated that the vehicle is in a stable state without shaking or the like. If the stop estimating unit 50 determines that the vehicle is not in the stopped state, the stop estimating unit 50 outputs the fact to the non-straight running estimation unit 54.

  The straight traveling estimation unit 52 acquires the GPS positioning data 200 determined to be effective by the effectiveness determination unit 22 (not shown). Further, the straight traveling estimation unit 52 extracts a GPS azimuth from the GPS positioning data 200 and derives a change amount of the GPS azimuth over a predetermined period (hereinafter referred to as “GPS azimuth change amount”). Further, the straight traveling estimation unit 52 confirms whether the GPS azimuth change amount is “0”. Further, the straight traveling estimation unit 52 calculates a variance value of the output signal 206 in a predetermined period, and compares the variance value with a sixth threshold value. Note that the sixth threshold value may be the same as the fifth threshold value. Here, the predetermined period is set to a period in which the GPS azimuth change amount is continuously zero, for example.

  The straight travel estimation unit 52 determines that the vehicle is traveling straight when the GPS heading change amount is 0 and the variance value is smaller than the sixth threshold value. If the variance value of the output signal 206 is small during the predetermined period, it is estimated that the vehicle is traveling straight without being affected by subtle meandering. Depending on the driver's driving situation and road shape, for example, in an urban area or the like, the detection frequency of the straight traveling state is generally lower than the determination of the stop state by the stop estimation unit 50, and the period is about several seconds. is there. When the straight traveling estimation unit 52 determines that the vehicle is not in the straight traveling state, the straight traveling estimation unit 52 outputs the fact to the non-straight traveling estimation unit 54. Here, when the stop estimation unit 50 determines that the vehicle is in the stopped state, and the straight traveling traveling estimation unit 52 determines that the vehicle is in the straight traveling state, the determination result of the stop estimating unit 50 is given priority. When the non-straight running estimation unit 54 inputs from the stop estimating unit 50 that the vehicle is not in a stopped state and from the straight running estimation unit 52, the vehicle is in a non-straight running state. Is determined.

  The state-specific offset value deriving unit 42 receives the GPS positioning data 200, the output signal 206, the tilt angle 204, and the sensitivity coefficient 210. The state-specific offset value deriving unit 42 sequentially derives the temporary offset value 216 of the angular velocity sensor 28 according to the traveling state of the vehicle estimated by the state estimating unit 40. When the stop estimation unit 50 determines that the vehicle is stopped, the stop offset value deriving unit 56 sequentially derives the temporary offset value 216 based on the output signal 206. When the straight travel estimation unit 52 determines that the vehicle is traveling straight, the straight travel offset value deriving unit 58 sequentially derives the temporary offset value 216 based on the output signal 206.

  Further, when the straight travel offset value deriving unit 58 determines that the vehicle is in the non-straight travel state, the non-straight travel offset value deriving unit 60 uses the GPS positioning data 200, the output signal 206, the tilt angle 204, and the sensitivity coefficient 210. The temporary offset value 216 is sequentially derived. That is, the non-straight running offset value deriving unit 60 derives the running state information 218 while changing the combination of the GPS positioning data 200, the output signal 206, and the like according to the traveling state of the vehicle. To do.

  The stop offset value deriving unit 56 sequentially derives the temporary offset value 216 of the angular velocity sensor 28 based on the output signal 206 when it is determined that the vehicle is stopped. Specifically, the offset value deriving unit 56 at the time of stop uses the fact that the turning angular velocity of the vehicle becomes “0” at the time of stop, and calculates the average value of the output signal 206 as the running state information 218. The straight traveling offset value deriving unit 58 sequentially derives the temporary offset value 216 of the angular velocity sensor 28 based on the output signal 206 when it is determined that the vehicle is traveling straight. Specifically, since the turning angular velocity of the vehicle is also 0 here, the straight traveling offset value deriving unit 58 calculates the average value of the output signal 206 as the traveling state information 218.

When it is determined that the vehicle is in the non-straight running state, the non-straight running offset value deriving unit 60 is based on the GPS azimuth amount, the inclination angle 204, the output signal 206, and the sensitivity coefficient 210 in the GPS positioning data 200. For example, the temporary offset value 216 in the sampling interval of the GPS direction is sequentially derived. Here, the temporary offset value 216 is derived as follows.
Goffset = 1 / n · ΣGout−Δθ · Gsensitivity · cos (α) (4)
Here, n is the number of samples of the output signal 206 at the GPS azimuth sampling interval, and ΣGout (mV) is the total value of the output signal 206 at the GPS azimuth sampling interval. Further, Δθ (deg) is a GPS azimuth change amount, Gsensitivity (mV / deg / sec) is a sensitivity coefficient 210, and α (deg) is an inclination angle 204 of the angular velocity sensor. Gout corresponds to Vout in Expression (1), and Gsensitivity corresponds to S in Expression (1).

  The sensitivity coefficient 210 is normally input from a sensitivity coefficient calculation unit 32 (not shown), but the sensitivity coefficient 210 may not be calculated yet in a state such as immediately after the angular velocity calculation apparatus 100 is activated. In such a case, the offset value deriving unit 60 during non-straight running uses the sensitivity coefficient 210 determined by the specifications of a gyro device (not shown) as an initial value. Further, the offset value deriving unit 60 during non-straight running may store the sensitivity coefficient 210 from the sensitivity coefficient calculating unit 32 at the end of the previous run and use it as an initial value.

  The offset value filter processing unit 44 inputs the temporary offset value 216 sequentially derived by the state-specific offset value deriving unit 42. The offset value filter processing unit 44 derives the offset value 208 of the angular velocity sensor 28 by performing statistical processing on the temporary offset value 216. The offset value 208 is indicated as Goffset, and this corresponds to Voffset in Expression (1). Hereinafter, the processing in the offset value filter processing unit 44 will be described with reference to FIG.

  FIG. 3 shows a configuration of the offset value filter processing unit 44. The offset value filter processing unit 44 includes an αi multiplier 70, an adder 72, a 1-αi multiplier 74, and a forgetting factor controller 76. As illustrated, the offset value filter processing unit 44 is configured to include an IIR (Infinite Impulse Response) filter, and the IIR filter configures a low-pass filter. The αi multiplier 70 multiplies the temporary offset value 216 by the forgetting factor “αi”. Here, “i” is 1 or 2. Therefore, the forgetting factor “αi” is a general term for α1 and α2. Α1 and α2 will be described later. The αi multiplier 70 outputs the multiplication result to the adder 72.

  The adder 72 sequentially adds the multiplication result from the αi multiplier 70 and the multiplication result from the 1-αi multiplier 74. The adder 72 sequentially outputs the addition result as an offset value 208. The 1-αi multiplier 74 multiplies the offset value 208 by a coefficient “1-αi”. Note that “αi” in the coefficient “1-αi” is the same as αi in the αi multiplier 70, and thus the description thereof is omitted here. The 1-αi multiplier 74 feeds back the multiplication result to the adder 72. The forgetting factor control unit 76 inputs the traveling state information 218. Further, the forgetting factor control unit 76 determines the value of the forgetting factor “αi” according to the state indicated by the traveling state information 218. Further, the forgetting factor control unit 76 sets the determined forgetting factor “αi” in the αi multiplication unit 70 and the 1-αi multiplication unit 74.

  FIG. 4 shows a data structure of a table stored in the forgetting factor control unit 76. As shown, a running state column 300 and a forgetting factor column 302 are included. The traveling state column 300 includes each state indicated by the traveling state information 218. The forgetting factor column 302 stores a forgetting factor “αi” corresponding to each state. That is, the forgetting factor “α1” is associated with the stop state and the straight traveling state, and the forgetting factor “α2” is associated with the non-straight traveling state. For example, α1> α2. Returning to FIG. The forgetting factor control unit 76 selects the forgetting factor “α1” or “α2” from the state indicated by the traveling state information 218 while referring to the table of FIG. In this way, the forgetting factor control unit 76 changes the forgetting factor at the time of the filtering process according to the traveling state of the vehicle.

  FIG. 5 shows the configuration of the sensitivity coefficient calculation unit 32. The sensitivity coefficient calculation unit 32 includes a sensitivity coefficient derivation unit 90, a sensitivity coefficient filter unit 92, and a forgetting factor control unit 94. The sensitivity coefficient deriving unit 90 receives the GPS positioning data 200, the tilt angle 204, the output signal 206, and the correlation coefficient 212. The sensitivity coefficient calculation unit 32 also inputs an offset value 208. The sensitivity coefficient deriving unit 90 sequentially derives a temporary sensitivity coefficient of the angular velocity sensor 28 based on the GPS positioning data 200, the tilt angle 204, the output signal 206, the correlation coefficient 212, and the offset value 208.

More specifically, if the GPS positioning data 200 indicates that the GPS orientation is valid, the sensitivity coefficient deriving unit 90 provides a temporary sensitivity coefficient of the angular velocity sensor 28 at the GPS orientation sampling interval as follows. Is calculated.
Gsensitivity = ((1 / n · ΣGout−Goffset) / cos (α)) / Δθ (5)
Here, Goffset is input from an offset value calculation unit 30 (not shown), but it is possible that the offset value 208 has not yet been calculated in a state such as immediately after the angular velocity calculation device 100 is activated. Since equation (5) includes division by Δθ, the sensitivity coefficient is calculated when Δθ is equal to or greater than a predetermined value. When the value of Δθ is equal to or smaller than the predetermined value, the sensitivity coefficient deriving unit 90 outputs the sensitivity coefficient corrected immediately before.

  The sensitivity coefficient filter unit 92 inputs the temporary sensitivity coefficient sequentially derived by the sensitivity coefficient deriving unit 90. The sensitivity coefficient filter unit 92 derives the sensitivity coefficient 210 of the angular velocity sensor 28 by performing statistical processing on the temporary sensitivity coefficient. The sensitivity coefficient 210 can be said to be a value for correcting the output signal 206 output from the angular velocity sensor 28. Similar to the offset value filter processing unit 44 shown in FIG. 3, the sensitivity coefficient filter unit 92 is configured by an IIR filter. The IIR filter forms a low-pass filter, and an instruction from the forgetting factor control unit 94 Based on the above, the forgetting factor in the IIR filter is set.

  The forgetting factor control unit 94 changes the forgetting factor in the sensitivity coefficient filter unit 92 according to the correlation coefficient 212. Further, the forgetting factor control unit 94 instructs the sensitivity factor filter unit 92 to use the changed forgetting factor. FIG. 6 shows a data structure of a table stored in the forgetting factor control unit 94. As shown, a condition column 400 and a forgetting factor column 402 are shown. In the condition column 400, “more than x” and “less than x” are shown as conditions of the correlation coefficient 212 for determining the forgetting coefficient. In the forgetting factor column 402, the value of the forgetting factor corresponding to each condition in the condition column 400 is shown. Specifically, the forgetting factor α10 is set for “x or more” and the forgetting factor α11 is defined for “less than x”. Further, a relationship of α11 <α10 is defined.

Here, the correlation coefficient which the correlation coefficient calculating part 34 calculates is demonstrated using FIG. FIG. 7 shows an overview of the correlation coefficient calculated by the correlation coefficient calculator 34, which is the difference between the output signal 206 input to the correlation coefficient calculator 34 and the offset value 208 at the inclination angle 204. A graph of the value divided by the cosine value and the GPS azimuth change amount included in the GPS positioning data 200 is shown. The correlation coefficient calculation unit 34 derives D _Gyro by dividing the difference between the integrated value of the output signal 206 and the offset value 208 by the cosine value of the tilt angle 204 as follows.

The right side of (Expression 5), which is a sensitivity coefficient calculation formula, represents the ratio between the angular velocity sensor output variation and the GPS azimuth variation, and D _Gyro corresponds to the angular velocity sensor output variation. The correlation coefficient calculation unit 34 derives a correlation coefficient C between the angular velocity sensor output change amount D _Gyro and the GPS azimuth change amount Δθ as follows.

The sensitivity coefficient 210 is generally affected by individual differences and aging deterioration, but its value changes little in the short term. Therefore, it can be said that the correlation coefficient C is large, in other words, the derivation accuracy of each variable on the right side of (Expression 5) is good and the error is small. That is, in FIG. 7, when the relationship between D _Gyro and Δθ is high, the derivation accuracy of each variable is good, and when the relationship between D _Gyro and Δθ is low, the derivation accuracy of each variable deteriorates. ing. The correlation coefficient calculator 34 outputs the correlation coefficient C as the correlation coefficient 212 to the sensitivity coefficient calculator 32. As described above, in the sensitivity coefficient filter unit 92, when the correlation coefficient 212 is greater than or equal to x, the derivation accuracy of the sensitivity coefficient 210 is high. An averaging process with a larger weight is performed.

  The operation of the angular velocity calculation apparatus 100 having the above configuration will be described. FIG. 8 is a flowchart showing the procedure for deriving the forgetting factor by the angular velocity calculation apparatus 100. When the storage data 214 is input from the positioning data storage unit 36 (Y in S10), the correlation coefficient calculation unit 34 calculates the GPS azimuth change amount and the angular velocity sensor output change amount for a predetermined period (S12). The correlation coefficient calculation unit 34 stands by when the storage data 214 is not input from the positioning data storage unit 36 (N in S10). The correlation coefficient calculation unit 34 calculates a correlation coefficient between the calculated GPS azimuth change amount and the angular velocity sensor output change amount (S14), and outputs the correlation coefficient to the forgetting factor control unit 94. The forgetting factor control unit 94 sets the forgetting factor α10 to the sensitivity factor filter unit 92 when the correlation factor is equal to or greater than the threshold value (Y in S16), and the correlation factor is less than the threshold value. (N in S16), the forgetting factor α11 is set in the sensitivity coefficient filter unit 92.

  According to the embodiment of the present invention, the forgetting factor is changed on the basis of the correlation factor, so that the sensitivity factor can be derived by filtering processing adapted to the GPS reception state. In addition, since the filtering process suitable for the GPS reception state is realized, the sensitivity coefficient derivation accuracy can be improved. In addition, since the correlation coefficient between the angular velocity sensor output change amount and the GPS azimuth change amount is calculated and the forgetting factor is controlled based on the correlation coefficient, Δθ obtained from the GPS azimuth change amount and the average inclination angle of the vehicle The influence of errors included in α can be reduced, and the sensitivity coefficient derivation accuracy can be improved. In addition, since the forgetting factor is set according to the correlation factor, it is possible to reduce the error included in Δθ obtained from the GPS azimuth change amount in Equation (5) and the average inclination angle α of the vehicle.

(Example 2)
Next, Example 2 will be described. As in the first embodiment, the second embodiment also relates to an angular velocity calculation device that derives an angular velocity using an offset value and a sensitivity coefficient with respect to an output voltage from the angular velocity sensor. In the equation (2) for deriving the offset value of the angular velocity sensor, when the azimuth change amount Δθ is obtained from the GPS azimuth obtained from the GPS satellite, the accuracy of the GPS azimuth deteriorates depending on the reception status of the radio wave from the GPS satellite. In some cases, the error included in the offset value increases.

  Further, in the equation (2), when the direction change amount Δθ is obtained based on the map data, the road direction based on the map data and the traveling direction of the vehicle do not always coincide with each other, and therefore are included in the offset value. The error can be large. Further, in the equation (2) for deriving the offset value of the angular velocity sensor, when the road inclination angle included in the inclination of the detection axis of the angular velocity sensor with respect to the vertical axis is obtained from the GPS altitude change amount acquired from the GPS satellite, The accuracy of the GPS azimuth may deteriorate due to the reception status of radio waves from the GPS satellite, and the error included in the offset value increases.

  In these cases, in order to improve the derivation accuracy of the angular velocity, it is required to reduce the error included in the offset value. For this reason, the angular velocity calculation apparatus according to the present embodiment uses the forgetting coefficient while changing the weighted average for the past value and the current value of the offset value of the angular velocity sensor that are sequentially derived. More specifically, the angular velocity calculation device calculates a correlation coefficient between the azimuth change calculated from the angular velocity sensor and the azimuth change calculated from GPS in a predetermined period, and changes the forgetting coefficient based on the correlation coefficient. To do.

  FIG. 9 shows a configuration of an angular velocity calculation apparatus 100 according to the second embodiment of the present invention. The angular velocity calculation device 100 is configured by the same members as in FIG. 1 except that the correlation coefficient calculation unit 34 outputs the correlation coefficient 212 to the offset value calculation unit 30. Below, it demonstrates centering on the difference from before.

  The offset value calculation unit 30 includes GPS positioning data 200 from the validity determination unit 22, a tilt angle 204 from the tilt angle detection unit 26, an output signal 206 from the angular velocity sensor 28, and a correlation coefficient from the correlation coefficient calculation unit 34. 212 is entered. The offset value calculation unit 30 also receives the sensitivity coefficient 210 from the sensitivity coefficient calculation unit 32. The offset value calculation unit 30 is based on the GPS positioning data 200, the tilt angle 204, the output signal 206, the sensitivity coefficient 210, and the correlation coefficient 212, and the offset value of the angular velocity sensor 28 (hereinafter referred to as “offset value 208”). Is calculated. Details of processing in the offset value calculation unit 30 will be described later. The offset value calculation unit 30 outputs the offset value 208 to the angular velocity conversion unit 14, the sensitivity coefficient calculation unit 32, and the correlation coefficient calculation unit 34.

  The correlation coefficient calculation unit 34 sequentially inputs the offset value 208 from the offset value calculation unit 30 and the storage data 214 from the positioning data storage unit 36. When the validity flag added to the stored data 214 is valid, the correlation coefficient calculation unit 34 divides the output signal 206 by the cosine value of the inclination angle 204 and the GPS azimuth included in the GPS positioning data 200. For the change amount, the correlation coefficient 212 is calculated in a predetermined period, for example, in a period in which the sensitivity coefficient calculation unit 32 calculates the sensitivity coefficient 210. Details of processing in the correlation coefficient calculation unit 34 will be described later. The correlation coefficient calculator 34 outputs the correlation coefficient 212 to the offset value calculator 30 and the sensitivity coefficient calculator 32.

  FIG. 10 shows a configuration of the offset value calculation unit 30. The offset value calculation unit 30 is configured by the same members as in FIG. 2 except that the correlation coefficient 212 is input to the offset value filter processing unit 44. Below, it demonstrates centering on the difference from before. The offset value filter processing unit 44 inputs the temporary offset value 216 sequentially derived by the state-specific offset value deriving unit 42. The offset value filter processing unit 44 derives the offset value 208 of the angular velocity sensor 28 by performing statistical processing on the temporary offset value 216. At that time, the offset value filter processing unit 44 changes the forgetting coefficient in the statistical processing according to the correlation coefficient 212. Hereinafter, the processing in the offset value filter processing unit 44 will be described with reference to FIG.

  FIG. 11 shows the configuration of the offset value filter processing unit 44. The offset value filter processing unit 44 is configured by the same members as in FIG. 3 except that the correlation coefficient 212 is input to the forgetting coefficient control unit 76. Below, it demonstrates centering on the difference from before. The forgetting factor control unit 76 inputs the correlation factor 212 and the running state information 218. Further, the forgetting factor control unit 76 determines the value of the forgetting factor “αi” according to the state indicated by the traveling state information 218 and the correlation coefficient 212. Further, the forgetting factor control unit 76 sets the determined forgetting factor “αi” in the αi multiplication unit 70 and the 1-αi multiplication unit 74.

  FIG. 12 shows a data structure of a table stored in the forgetting factor control unit 76. As shown, a running state column 500, a condition column 502, and a forgetting factor column 504 are included. The traveling state column 500 includes each state indicated by the traveling state information 218. In the condition column 502, “more than x” and “less than x” are shown as conditions of the correlation coefficient 212 for determining the forgetting coefficient. The forgetting factor column 504 stores a forgetting factor “αi” corresponding to each state. That is, the forgetting factor “α1” is associated with the stop state and the straight traveling state. Further, in the non-straight running state, “α21” is associated with “x or more” and “α22” is associated with “less than x”, corresponding to the condition column 502. In this manner, the forgetting factor control unit 76 selects the forgetting factor “α1” or “α2” from the state indicated by the traveling state information 218 while referring to the table of FIG. Therefore, the forgetting factor control unit 76 changes the forgetting factor at the time of the filtering process according to the traveling state of the vehicle.

Here, the correlation coefficient which the correlation coefficient calculating part 34 calculates is demonstrated using FIG. FIG. 13 shows an outline of the correlation coefficient calculated in the correlation coefficient calculation unit 34, which is a value obtained by dividing the output signal 206 input to the correlation coefficient calculation unit 34 by the cosine value of the inclination angle 204. The graph of the GPS azimuth | direction change amount contained in the GPS positioning data 200 is shown. The correlation coefficient calculator 34 derives D ′ _Gyro by dividing the integrated value of the output signal 206 by the cosine value of the tilt angle 204 as follows.

D ′ _Gyro also corresponds to the change amount of the angular velocity sensor output. The correlation coefficient calculation unit 34 derives a correlation coefficient C between the angular velocity sensor output change amount D ′ _Gyro and the GPS azimuth change amount Δθ as follows.

Here, the angular velocity sensor output change amount D ′ _Gyro is a value obtained by removing the Goffset term from the angular velocity sensor output change amount D _Gyro , and if the Goffset does not change during the calculation period of the correlation coefficient C, the azimuth change amount Δθ and Correlation coefficients with each of the same. Therefore, as described above, when the correlation coefficient C is large, it can be said that the accuracy of “1 / n · ΣGout”, “Δθ”, and “cos (α)” on the right side of Equation (4) is good. The correlation coefficient calculator 34 outputs the correlation coefficient C as the correlation coefficient 212 to the offset value calculator 30 and the sensitivity coefficient calculator 32. Note that the correlation coefficient 212 output to the sensitivity coefficient calculation unit 32 may be calculated by Expression (7).

  According to the embodiment of the present invention, since the forgetting factor is changed based on the correlation coefficient, the offset value can be derived by the filtering process adapted to the GPS reception state. In addition, since the filtering process suitable for the GPS reception state is realized, the accuracy of deriving the offset value can be improved. In addition, since the correlation coefficient between the angular velocity sensor output change amount and the GPS azimuth change amount is calculated and the forgetting factor is controlled based on the correlation coefficient, Δθ obtained from the GPS azimuth change amount and the average inclination angle of the vehicle The influence of the error included in α can be reduced, and the accuracy of deriving the offset value can be improved. In addition, since the forgetting factor is set according to the correlation factor, it is possible to reduce the error included in Δθ obtained from the GPS azimuth change amount in Equation (5) and the average inclination angle α of the vehicle.

(Example 3)
Next, Example 3 will be described. The third embodiment relates to an azimuth estimation apparatus that estimates an azimuth using the angular velocity derived in the conventional angular velocity calculation apparatuses. Even if the offset and sensitivity coefficient of the angular velocity sensor are corrected, a slight error may accumulate, and the error included in the position by the self-contained navigation may become significant. For this reason, when the vehicle travels in an area where GPS satellite acquisition is poor, the ratio of the self-contained navigation increases over a long period of time, and an error may be included in the estimated position.

  In this case, when the GPS satellite acquisition situation becomes good, it is necessary to increase the composition ratio of the positions calculated from the GPS and correct the estimated position error in a short period of time. Therefore, the direction estimation apparatus according to the present embodiment, when combining the direction calculated from the GPS and the angular velocity calculated from the angular velocity sensor, the direction change calculated from the angular velocity sensor and the direction change calculated from the GPS in a predetermined period, The correlation coefficient is calculated, and the composition ratio is changed based on the correlation coefficient.

  FIG. 14 shows a configuration of an azimuth estimation apparatus 150 according to Embodiment 3 of the present invention. The bearing estimation device 150 includes an angular velocity calculation device 100 and a dead reckoning position calculation unit 16. The dead reckoning position calculation unit 16 includes a GPS effectiveness determination unit 112, a composition ratio determination unit 114, and an estimated direction deriving unit 116. In addition, the GPS positioning data 200, the tilt angle 204, the output signal 206, the correlation coefficient 212, the angular velocity 220, the GPS effectiveness 222, and the composition ratio 224 are included as signals. Since the angular velocity calculation apparatus 100 is configured in the same manner as in FIG. 1 or FIG. 9, the description thereof is omitted here. The correlation coefficient calculation unit 34 included in the parameter calculation unit 12 also outputs the correlation coefficient 212 to the dead reckoning position calculation unit 16.

  The dead reckoning position calculation unit 16 inputs the GPS positioning data 200, the correlation coefficient 212, and the angular velocity 220 from the angular velocity calculation device 100, and derives the dead reckoning direction. The GPS effectiveness determination unit 112 receives the GPS positioning data 200 from the measurement unit 10, and provides a plurality of threshold values for the HDOP value, the number of received satellites, and the speed included in the GPS positioning data 200, and sets the respective values. In response, the GPS effectiveness 222 is determined step by step. The threshold value provided in the GPS validity determination unit 112 may be the same as or different from the threshold value provided in the validity determination unit 22.

  FIG. 15 shows a data structure of a table stored in the GPS effectiveness level determination unit 112. The condition column 600 shows the conditions of the HDOP value, the GPS speed, and the number of GPS reception satellites for determining the GPS effectiveness. Here, if “HDOP value is less than threshold C1,” “GPS speed is greater than or equal to threshold C2,” and “the number of GPS reception satellites is greater than or equal to threshold C3”, the accuracy of GPS positioning data 200 is high. Judgment is made and “high” is indicated in the GPS effectiveness column 602. Also, “HDOP value is greater than or equal to threshold C1” or “GPS speed is greater than 0 and less than threshold C2” or “the number of GPS reception satellites is greater than or equal to threshold C4 and less than threshold C3” If so, it is determined that the accuracy of the GPS positioning data 200 is low, and “low” is indicated in the GPS effectiveness column 602. Furthermore, if “the number of GPS reception satellites is less than the threshold value C4”, it is determined that the HDOP value and the GPS speed are invalid values, and “invalid” is indicated in the GPS validity column. Returning to FIG. The GPS effectiveness determination unit 112 outputs the result of the determination to the composite ratio determination unit 114 as the GPS effectiveness 222.

  The composite ratio determination unit 114 receives the correlation coefficient 212 from the parameter calculation unit 12 and the GPS effectiveness 222 from the GPS effectiveness determination unit 112, and based on the correlation coefficient 212 and the GPS effectiveness 222, the GPS A ratio (hereinafter also referred to as “compositing ratio”) for combining the GPS azimuth change amount and the angular velocity 220 included in the positioning data 200 is determined. FIG. 16 shows a data structure of a table stored in the synthesis ratio determination unit 114. When the GPS effectiveness 222 is “high” and the correlation coefficient 212 is equal to or greater than the threshold C5, the composition ratio is “GPS azimuth change 100%, angular velocity 0%”. If the threshold value is less than C5, the composition ratio is “GPS azimuth change amount 50%, angular velocity 50%”. When the GPS effectiveness 222 is “low” and the correlation coefficient 212 is equal to or greater than the threshold C5, the composition ratio is “GPS azimuth change 80%, angular velocity 20%”. If it is less than the threshold value C5, the composition ratio is “GPS azimuth change amount 30%, angular velocity 70%”.

  When the GPS effectiveness 222 is “not invalid” and the correlation coefficient 212 is equal to or greater than the threshold C5, the composition ratio is “GPS azimuth change 60%, angular velocity 40%”. Is less than the threshold value C5, the composition ratio is “GPS azimuth change 20%, angular velocity 80%”. When the GPS validity 222 is “invalid” and the correlation coefficient 212 is less than the threshold value C5, the composition ratio is “GPS azimuth change amount 0%, angular velocity 100%”. That is, the composition ratio determination unit 114 increases the ratio of the GPS azimuth change amount and decreases the angular velocity ratio as the GPS effectiveness 222 increases. Further, the composition ratio determination unit 114 increases the ratio of the GPS azimuth change amount and decreases the angular velocity ratio as the correlation coefficient 212 increases. Returning to FIG. The composition ratio determining unit 114 outputs the determined composition ratio 224 to the estimated orientation deriving unit 116.

  The estimated azimuth derivation unit 116 inputs the GPS positioning data 200 from the measurement unit 10, the synthesis ratio 224 from the synthesis ratio determination unit 114, and the angular velocity 220 from the angular velocity conversion unit 14. The estimated azimuth deriving unit 116 derives the azimuth change amount by synthesizing the GPS azimuth change amount included in the GPS positioning data 200 and the temporary offset value 216 based on the synthesis ratio 224. For example, if the north direction is expressed as 0 degrees / 360 degrees, the GPS direction change amount calculated from the input GPS positioning data 200 is 10 (degrees / second), the angular velocity 220 is 9 (degrees / second), and the composition ratio 224 Is “GPS azimuth variation 80%, angular velocity 20%”, the estimated azimuth derivation unit 116 sets the azimuth variation as 10 × 0.8 + 9 × 0.2 = 9.8 (degrees / second). calculate. Further, the estimated azimuth deriving unit 116 updates the estimated azimuth by adding the calculated azimuth change amount to the previously derived estimated azimuth.

  Note that the estimated azimuth deriving unit 116 determines that if the composition ratio 224 is “GPS azimuth change 100%, angular velocity 0%”, that is, if the composition ratio indicates use of only the GPS azimuth change, As the estimated orientation, the orientation included in the GPS positioning data 200 may be used. This corresponds to replacing the estimated direction before update with the GPS direction. By doing so, the GPS positioning state is improved even when the GPS positioning accuracy is poor and the composition ratio 224 is “GPS azimuth change 0%, angular velocity 100%” for a long period of time and the error of self-contained navigation is accumulated. Therefore, the accuracy of the estimated direction can be improved in a short period of time.

  According to the embodiment of the present invention, the ratio at the time of combining the azimuth change amount and the angular velocity is adjusted based on the correlation coefficient between the azimuth change amount calculated from the GPS azimuth and the angular velocity calculated from the angular velocity sensor. An azimuth change amount according to accuracy can be derived. Moreover, since the degree of correcting the dead reckoning direction is changed based on the composition ratio, the position estimation period can be shortened. Further, since the degree of correction of the dead reckoning direction is changed based on the composite ratio, the position estimation accuracy can be improved. In addition, when the composition ratio indicates the use of only the GPS azimuth change amount, the azimuth included in the positioning data is used as the presumed azimuth before updating, so that accumulated errors can be reduced in a short period of time. .

  In the above, this invention was demonstrated based on the Example. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to the combination of each component and each processing process, and such modifications are also within the scope of the present invention. .

  In the first to third embodiments of the present invention, the correlation coefficient calculation unit 34 fixes the calculation period when calculating the correlation coefficient. However, the present invention is not limited to this. For example, the correlation coefficient calculator 34 may adjust the calculation period for calculating the correlation coefficient in accordance with the amount of change in angular velocity. Here, the change amount of the angular velocity may be derived from the output signal 206, or may be derived from the GPS azimuth change amount Δθ included in the GPS positioning data 200. Thereby, when the change amount of the angular velocity is small, the calculation period is lengthened, and when the change amount of the angular velocity is large, the calculation period is shortened. According to this modification, a correlation coefficient corresponding to the amount of change in angular velocity can be derived.

  In the first to third embodiments of the present invention, the state estimating unit 40 determines the stop state using the GPS speed included in the GPS positioning data 200 as the vehicle traveling state determination. However, the present invention is not limited to this. For example, the state estimation unit 40 may input a vehicle speed pulse signal of a vehicle from a pulse detection unit (not shown) and determine the stop state based on the vehicle speed pulse. Here, the pulse detection unit is connected to a speed sensor (not shown), and the speed sensor is installed in the middle of the speedometer cable that rotates in response to the rotation of the drive shaft, and the vehicle speed pulse accompanying the rotation of the drive shaft. Output a signal. According to this modification, the speed of the vehicle can be measured by various means.

  In the first to third embodiments of the present invention, the validity determination unit 22 uses the PDOP to determine the validity of the GPS positioning data 200. However, the present invention is not limited to this. For example, the validity determination unit 22 may use GDOP (Geometric Dilution Of Precision), HDOP (Horizontal Division Of Precision), or a combination thereof. According to this modification, various parameters can be used for determination.

  In the first to third embodiments of the present invention, the offset value filter processing unit 44 and the sensitivity coefficient filter unit 92 are formed to include an IIR filter. However, the present invention is not limited to this. For example, the offset value filter processing unit 44 and the sensitivity coefficient filter unit 92 may be formed to include an FIR (Finite Impulse Response) filter. At that time, the forgetting factor is set as a tap factor. According to this modification, the degree of freedom of the filter configuration can be improved.

  In the first to third embodiments of the present invention, the forgetting factor control unit 76 defines the same value for the forgetting factor for the stop state and the forgetting factor for the straight traveling state. However, the present invention is not limited to this. For example, the forgetting factor control unit 76 may set different values for these. It can be said that different processing is performed for each of the three states. According to the present modification, three forgetting factors are prepared for the three states as the running state, so that filter processing suitable for each state can be realized.

  DESCRIPTION OF SYMBOLS 10 Measurement part, 12 Parameter calculation part, 14 Angular velocity conversion part, 16 Dead reckoning position calculation part, 20 GPS positioning part, 22 Effectiveness determination part, 24 Mounting angle detection part, 26 Inclination angle detection part, 28 Angular speed sensor, 30 Offset Value calculation unit, 32 Sensitivity coefficient calculation unit, 34 Correlation coefficient calculation unit, 36 Positioning data storage unit, 38 Control unit, 40 State estimation unit, 42 State-specific offset value derivation unit, 44 Offset value filter processing unit, 50 Stop estimation , 52 Straight running estimation unit, 54 Non-straight running estimation unit, 56 Offset value deriving unit at stop, 58 Offset value deriving unit during straight running, 60 Offset value deriving unit during non-straight running, 70 αi multiplier, 72 Adder 74 1-αi multiplication unit, 76 forgetting factor control unit, 90 sensitivity factor deriving unit, 92 Sensitivity coefficient filter unit, 94 forgetting factor control unit, 100 angular velocity calculation device.

Claims (5)

An acquisition unit that acquires positioning data of an object measured based on a signal from a GPS satellite and an angular velocity of the object output from an angular velocity sensor;
A sensitivity coefficient calculation unit for deriving a sensitivity coefficient of the angular velocity sensor based on the positioning data and the angular velocity acquired in the acquisition unit;
Based on the positioning data and angular velocity acquired in the acquisition unit, an offset value calculation unit that derives an offset value of the angular velocity sensor;
An angular velocity conversion unit that corrects the angular velocity acquired in the acquisition unit based on the offset value of the angular velocity sensor derived in the offset value calculation unit and the sensitivity coefficient of the angular velocity sensor derived in the sensitivity coefficient calculation unit; ,
A correlation coefficient deriving unit for deriving a correlation coefficient based on the positioning data and the angular velocity acquired in the acquisition unit;
Based on the correlation coefficient derived in the correlation coefficient deriving unit and the effectiveness of the positioning data acquired in the acquisition unit, the amount of change in the azimuth included in the positioning data acquired in the acquisition unit, A determination unit that determines a ratio when combining the angular velocity corrected in the angular velocity conversion unit;
Update by which the change amount of the azimuth included in the positioning data acquired in the acquisition unit and the angular velocity corrected in the angular velocity conversion unit are combined with the ratio determined in the determination unit, and the azimuth is updated with the combined value And
An azimuth estimation device comprising:   The determination unit increases the ratio of the change amount of the azimuth included in the positioning data acquired by the acquisition unit as the effectiveness of the positioning data acquired by the acquisition unit increases, and the correction is performed by the angular velocity conversion unit. The azimuth estimation apparatus according to claim 1, wherein a ratio of angular velocities is lowered.   The determination unit increases the ratio of change in azimuth included in the positioning data acquired in the acquisition unit as the correlation coefficient derived in the correlation coefficient deriving unit increases, and corrects in the angular velocity conversion unit The direction estimation apparatus according to claim 1, wherein the ratio of the angular velocity is reduced.   The update unit, when the ratio determined in the determination unit indicates the use of only the amount of change in the azimuth included in the positioning data acquired in the acquisition unit, as the direction before the update, in the acquisition unit The azimuth estimation apparatus according to any one of claims 1 to 3, wherein the azimuth included in the obtained positioning data is used. Obtaining positioning data of an object measured based on a signal from a GPS satellite and an angular velocity of the object output from an angular velocity sensor;
Deriving a sensitivity coefficient of the angular velocity sensor based on the obtained positioning data and angular velocity;
Deriving an offset value of the angular velocity sensor based on the obtained positioning data and angular velocity;
Correcting the obtained angular velocity based on the derived offset value of the angular velocity sensor and the derived sensitivity coefficient of the angular velocity sensor;
A step of deriving a correlation coefficient based on the obtained positioning data and angular velocity;
Based on the derived correlation coefficient and the effectiveness of the acquired positioning data, determining a ratio when combining the amount of change in azimuth included in the acquired positioning data and the corrected angular velocity;
A step of combining the amount of change in azimuth included in the acquired positioning data and the corrected angular velocity according to the determined ratio, and updating the azimuth with the combined value;
An azimuth estimation method comprising:

Images