JP3875714B2 - Moving body acceleration / distance estimation circuit, moving body positioning device, and moving body positioning method - Google Patents

Description

  The present invention relates to a moving body positioning apparatus and a moving body positioning method for obtaining an accurate moving position of a moving body based on the acceleration of the moving body and positioning information of the moving body using a satellite.

In conventional mobile body position estimation, an INS that calculates the position of a mobile body by inertial navigation calculation by integrating the movement of the mobile body based on an acceleration detector or the like mounted on the mobile body itself called an IMU (Inertial Measurement Unit) (Inertial Navigation System) There was a navigation called navigation.
In addition, there is a navigation called GPS (Global Positioning System) navigation in which the position of a moving body is calculated by positioning navigation based on the propagation time difference of the positioning signal based on the positioning signal from the positioning satellite.
Also, as the composite navigation, Kalman filter that inputs position information obtained from INS navigation and GPS navigation and calculates and outputs an error estimation value and a state estimation value for correction based on the input time difference. A configuration is also known in which this output is input to the INS navigation using. Based on the feedback amount, the INS navigation calculation unit is influenced by other position information because position information and speed information of another navigation are input. In that case, generally, the output of this composite navigation is used as the position where the moving body has moved.

  As the operation of the Kalman filter, calculation processing such as differentiation is performed, and the difference between the navigation state quantities consisting of the position and speed information output by the above two navigations and the errors of various sensors are estimated and output in real time. Is returned to the INS navigation and used for correction input of the navigation. In the above-described compound navigation, the input / output of the Kalman filter has a configuration called loose coupling or tight coupling. Loose coupling refers to a configuration in which both the speed and position, which are outputs of INS navigation, and the speed and position from GPS navigation are input, and the output of the Kalman filter is fed back to INS navigation. Tight coupling refers to a configuration in which a GPS navigation input signal is input to a Kalman filter instead of a position output from GPS navigation.

Note that the GPS position information is not yet highly accurate, and the DGPS (correcting the error amount of the measurement point based on the error amount of the GPS signal at a plurality of reference points whose coordinate values are accurately known). Differential Global Positioning System) is known. Non-Patent Document 1 discloses a method of knowing an error amount at a measurement point a certain distance away from a reference point in a certain range centered on a specific reference point from an error amount at a plurality of reference points.
Japanese translation of PCT publication No. 2003-509671 Special Table 2000-502802 JP 2001-264409 A JP 2002-196060 A US Pat. No. 6,311,129 RTK Networks based on Geo ++ GNSMART®-Concepts, Implementation, Results (Gerhard Wuebbena, Andreas Bagge, Martin Schmitz) International Technical Meeting, ION GPS-01, Sept. 11-14, 2001)

The conventional mobile position estimation apparatus is configured as described above, and the input of INS navigation that obtains position information by inputting based on IMU acceleration detection indicates that the acceleration value obtained by the IMU is accurate. Therefore, it is difficult to say that the accuracy of the position information that is the output of the INS navigation is accurate.
In the case of GPS navigation, since the position information obtained by GPS navigation includes various error factors, the accuracy of the output position by GPS navigation is not accurate.
In addition, there is a composite navigation that uses GPS navigation in combination to correct the output position of INS navigation via a Kalman filter. However, since the position information obtained by GPS navigation also includes various error factors, GPS navigation is used. Even if it correct | amends with the positional information obtained by (3), the precision is not necessarily exact.

  The present invention has been made to solve the above-described problem, and improves the accuracy of the input side of the distance information including the acceleration signal in the inertial navigation so that the calculated navigation position of the moving body is an accurate value. Another object of the present invention is to improve the accuracy of the pseudo distance on the input side in positioning navigation, further improve the quality of feedback information, and obtain more accurate position information as a result.

A mobile positioning device according to the present invention receives information from a positioning satellite, obtains a plurality of pseudo-ranges from the positioning satellite to the mobile body and a navigation calendar of the positioning satellite, and also uses a positioning navigation to determine a positioning navigation speed and a positioning navigation position. In a positioning device that obtains the navigation position of a moving body by a positioning navigation unit that outputs, an inertial navigation unit that outputs inertial navigation speed and inertial navigation position by inertial navigation, and a Kalman filter that outputs state estimation values,
The Doppler observation value between the positioning satellite and the moving body, which has been error-corrected with the wide area correction information, is input, and the relative speed between the positioning satellite and the moving body is calculated based on the navigation calendar and the state estimation value output by the Kalman filter. The Doppler estimated value between the positioning satellite and the moving object is obtained, the Doppler estimated value is compared with the error-corrected Doppler observed value, and a predetermined pseudo distance is calculated from the plurality of pseudo distances based on the comparison result. A satellite information selection unit that selects and outputs as a selected pseudorange;
The carrier phase information obtained from the signal received from the positioning satellite is input, and the input carrier phase information, the selected pseudo-range output from the satellite information selection unit, and the state estimation value output from the Kalman filter are calculated and simulated. A pseudo-distance estimation unit that obtains a distance estimate and outputs the pseudo-range estimate obtained;
The positioning navigation unit receives the pseudo-range estimation value output from the pseudo-range estimation unit as an input, obtains the positioning navigation speed and the positioning navigation position, and outputs the positioning navigation speed and the positioning navigation position.
The Kalman filter includes the pseudo-range estimation value output from the pseudo-range estimation unit, the positioning navigation speed output from the positioning navigation unit, the positioning navigation position, the inertial navigation speed output from the inertial navigation unit, and the inertial navigation position. Was used as an input to calculate the estimated state value.

  According to this configuration or method, the position based on the accurate inertial navigation calculation or the positioning navigation calculation can be determined based on the information obtained by increasing the signal accuracy of the acceleration signal including speed detection on the input side or by increasing the accuracy by the correct positioning satellite. There is an effect to be obtained.

Embodiment 1 FIG.
FIG. 1 shows a schematic configuration diagram of the entire mobile body positioning device in this embodiment. The mobile body positioning device is mounted on the mobile body. In FIG. 1 and subsequent figures, black circles indicate connection points. White circles indicate adders. Lines indicate signals (signal lines). The arrow at the line end indicates the output destination of the signal indicated by the line. If the signal input to the adder has a minus sign (-), it indicates that the signal value is subtracted. FIG. 2 is a diagram showing a detailed configuration of the SF (scale factor) / speed estimation unit 24 and the acceleration estimation unit 25.

  In FIG. 1, the mobile body positioning device includes a satellite positioning navigation system (GPS navigation system), an inertial navigation system (INS navigation system), and a Kalman filter 16. Here, the satellite positioning navigation is a positioning calculation by measuring the distance between the positioning satellite and the moving body, more precisely, the distance between the positioning satellite and the receiver (GPS unit 11) mounted on the moving body by the propagation time of the radio wave. Is to do. As a result, the position of the moving body is output as the calculation result in the GPS navigation unit based on the pseudorange information obtained by receiving the signal from the positioning satellite. In addition to the GPS navigation position, the Kalman filter, in addition to the speed obtained by the GPS navigation unit, is output as a GPS navigation state quantity.

  In addition, the inertial navigation includes an acceleration detector mounted on a moving body, a vehicle speed detecting section (moving body speed detecting section) (VMS section) for detecting vehicle speed (rotation) pulses of wheels, and Forward / Reverse indicating the traveling direction of the moving body. (F / R) Using an input from a detection unit or the like, an acceleration estimation unit that outputs a distance estimation value or an acceleration estimation value indicating a moving distance of the moving body is provided, and a navigation calculation is performed based on this output to output an INS navigation position. . Thus, using the stable and correct SF, correct information can be input to the INS navigation unit based on the acceleration and distance information corrected by the alignment. Similarly to the GPS navigation unit, an INS navigation state quantity composed of speed and acceleration calculated in association with the INS navigation unit is output to the Kalman filter.

  The satellite positioning navigation system (GPS navigation system) has the following elements. First, there is a GPS unit 11 that receives signals from positioning satellites such as GPS satellites and quasi-zenith satellites and obtains observation amounts such as pseudoranges and Doppler excursions. Among the observed quantities from the GPS unit 11, a pseudo distance is input, and a position error signal based on a plurality of electronic reference points is input as wide area correction information, and predetermined correction information is selected to correct the pseudo distance. Next, there is a DGPS (Differential Global Positioning System) unit 12 for outputting the corrected pseudoranges in time series. Furthermore, when the observation amount 12a including the corrected pseudo distance obtained in time series from the DGPS unit 12 is input, it is determined whether or not the information is based on multipath. A satellite that discards the corrected pseudorange and determines that the signal is based on the correct path instead of the multipath, and outputs the selected pseudorange 13a by selecting the correct path signal by outputting the corrected pseudorange. There is an information selector 13.

Further, there is a pseudo distance estimation unit 14 that receives the selected pseudo distance 13a determined to be a correct path, outputs a pseudo distance estimation value 14a that is determined to be more correct after being corrected with various other signals.
Thus, by combining the DGPS unit, the satellite information selection unit, and the pseudo distance estimation unit, only the correct pseudo distance is selected and input to the GPS navigation unit. A GPS navigation unit (satellite positioning navigation unit) that receives the pseudo-range estimation value 14a as input and outputs an INS navigation state quantity 15a composed of a position (GPS navigation position) by satellite positioning navigation and a related speed by using this input. 15).

In addition, as an INS inertial navigation system (INS navigation system), an IMU unit 21 that detects the angular acceleration of the moving body, the acceleration in the vertical and horizontal directions, and the acceleration in the traveling direction (distance direction), the forward traveling or the backward traveling of the moving body. Output from the F / R unit 22 for detecting the vehicle, a vehicle speed detection unit 23 (moving body speed detection unit) for detecting a pulse signal (vehicle speed pulse) generated according to the amount of movement of the moving body, and the vehicle speed detection unit 23. There is an SF (scale factor) / speed estimation unit 24 that multiplies the number of vehicle speed pulses per unit time as a vehicle speed detection signal by a scale factor (SF) and outputs a moving body speed estimation value 24a.
The SF / speed estimator 24 calculates the coordinate conversion value of the GPS navigation state quantity (position, speed) 15a, which is an output from the GPS navigation unit 15, and the INS navigation unit, based on whether or not the speed of the moving object can be estimated. From the quantity obtained from the coordinate conversion value of the INS navigation state quantity (position, speed) 26a, which is the output from 26, an output with higher quality is selected and the scale factor is corrected.

  Further, in the INS navigation system, the traveling direction acceleration obtained from the IMU unit 21 is corrected with the above-mentioned moving body speed estimated value 24a, and further, alignment correction described in detail later is performed, and the corrected estimated acceleration is output. The acceleration estimation unit 25 that outputs the distance estimation values 45a and 47a by calculating in detail, and the estimated acceleration 25a is further inputted as the difference between the error estimation value 16b of the Kalman filter output, and from the acceleration estimation unit 25 The INS navigation unit 26 calculates and outputs the INS navigation position using the distance estimation value and the state estimation value 16c from the Kalman filter 16 as inputs.

In addition, the INS navigation unit 26 receives the GPS navigation state quantity 15a and the INS navigation state quantity 26a, the pseudo distance estimation value 14a output from the pseudo distance estimation unit 14 and the navigation calendar 11a as outputs of both navigation units. Kalman filter that outputs an error estimation value 16b and a state estimation value 16c that are feedback outputs to the side, and outputs a state estimation value 16a that is a feedback output to the pseudorange estimation unit 14 and the satellite information selection unit 13 on the GPS navigation side There are 16.
The Kalman filter 16 performs tight coupling using the pseudo-range estimation value 14a and loose coupling using the GPS navigation state quantity and the INS navigation state quantity as inputs from the GPS navigation unit 15 and the INS navigation unit 26, respectively. They are also used together, and when estimating and outputting differences in navigation state quantities and errors of various sensors in real time, more detailed information is input, and therefore more accurate errors are estimated and output.
In FIG. 1, the GPS navigation position that is a part of the navigation state quantity 15a output from the GPS navigation unit 15 and the INS navigation position that is a part of the navigation state quantity 26a output from the INS navigation part 26 are shown in the latter part. Although used as appropriate, the description is omitted here because it is not related to the configuration and operation of the present embodiment.

The detailed configuration of the moving body acceleration / distance estimation circuit, that is, the SF / speed estimation unit 24 and the acceleration estimation unit 25 will be described with reference to FIG.
The vehicle speed detector 23 is an element that detects the speed (vehicle speed) of the moving body. This may be a digital signal or an analog signal. For example, a pulse is generated each time the tire rotates, and this generated pulse is transmitted. Therefore, if the vehicle speed pulses per unit time are counted, the vehicle's current speed equivalent can be determined. More specifically, if the vehicle speed pulse is detected every time the tire rotates, the speed can be determined by multiplying it by a conversion factor corresponding to the tire circumference, because it is the number of pulses per hour. At this time, the traveling direction is discriminated based on a forward or backward signal from the F / R unit 22. The SF / speed estimation unit 24 inputs the vehicle speed pulse to the SF calculation circuit 51. By the way, the scale factor, which is a conversion factor from the number of vehicle speed pulses per unit time to the speed, varies as described later. Therefore, it is necessary to update following the fluctuation, and to estimate the speed with the updated scale factor. Corresponding to this, there is an estimation availability determination circuit 52 that determines whether or not the scale factor can be updated. This condition will be described in the explanation of operation.

  If it is determined by the estimation availability determination circuit 52 that the update is possible, the input for scale factor update is output to the Kalman filter 53 as an input switching (or selection) instruction as to which of the GPS navigation system and the INS navigation system is appropriate. The state quantity related to the position in this case is expressed in the earth coordinate system for both GPS navigation and INS navigation, but the coordinate system of the moving body is different from this. Therefore, it is necessary to convert both coordinate systems, and the coordinate conversion circuits 54 and 55 do this. The SF calculation circuit 51 receives the vehicle speed pulse from the vehicle speed detector 23 and the output from the Kalman filter 53 constituting the calculation loop, updates and holds the scale factor, and then multiplies the vehicle speed pulse to multiply the mobile object. The speed estimated value 24a is output.

On the other hand, the acceleration estimation unit 25 first inputs various detected acceleration signals including the traveling direction acceleration 21 a from the IMU unit 21. In FIG. 2, although not shown in detail in FIG. 1, inclination correction by an inclination detector 41 such as an inclination sensor of the traveling direction acceleration obtained by the IMU unit 21 is also shown. The tilt detector 41 detects that the moving body is tilted. In FIG. 2, the output correction amount 41a is converted from the detected tilt amount into the X-axis direction of gravity acceleration, that is, the traveling direction acceleration. To output the acceleration correction amount to be added or subtracted.
The IMU traveling direction acceleration 21a obtained from the IMU unit 21 is corrected with the inclination correction amount 41a to obtain the traveling direction acceleration 21b, and the traveling direction acceleration 21b is input to the acceleration estimation unit 25.
FIG. 4 also shows the relationship between the mobile body positioning device and the input in this embodiment. In this case, the GPS receiver mounted on the mobile body, which is an automobile, causes the pseudorange, Doppler deviation, etc. from the GPS satellite. The information is received, and the GPS navigation distance is obtained by the GPS navigation unit. At this time, there are an ionosphere and a troposphere (not shown) in the upper layer between the GPS satellite and the moving body, and these fluctuations become an error factor of the pseudorange. Then, the acceleration is detected by the IMU unit mounted on the vehicle, the vehicle speed pulse is detected by the vehicle speed detector, and the position moved in the traveling direction by the INS navigation unit is obtained therefrom.

Returning to FIG. 2, there is still a variation factor in the traveling direction acceleration 21b. Of these, a configuration for stationary alignment correction will be described.
First, in order to perform stationary alignment, the value in a stationary state must be measured. However, even if the alignment is first obtained at rest and the bias value is obtained, the value changes depending on the acceleration change received by the car and the sensor characteristic change thereafter. Therefore, unless the stationary alignment considered to be correct is performed every moment and the traveling direction acceleration signal is corrected with the bias value, the traveling direction acceleration is not correct. In the present embodiment, a slight stable period during traveling is detected, stationary alignment is performed during this small period, a bias value included in the acceleration signal is detected or measured, and the traveling method is performed using the detected bias value. Correct the acceleration.
First, a stationary determination circuit 42 for determining a stationary condition is provided, and the stationary alignment detection circuit 43 includes a bias included in the acceleration signal only when it is determined that the stationary circuit is stationary due to coincidence of information that can be determined to be a plurality of stationary. Detect value. The bias value is held until the next update. In this way, the acceleration estimation unit 25 receives the traveling direction acceleration 21b obtained by correcting the traveling direction acceleration obtained from the IMU unit 21 with the correction amount output by the inclination detector 41 as the correction value 41a, and is held by the stationary alignment detection circuit 43. It corrects with a bias value, and also correct | amends also with the correction amount by the movement alignment mentioned later, and outputs it as the estimated acceleration 25a.

  Further, the acceleration estimation unit 25 outputs (IMU) integrators 44 and 45 on the side for outputting a distance estimation value, and (VMS) (VMS means a moving body speed estimation value side) distance estimation value. And a moving alignment detection circuit 46 that outputs an acceleration error estimated value 73a and a speed error estimated value 76a by using these two distance estimated values and two speeds on the IMU side and the VMS side as time series inputs. .

FIG. 3 is a flowchart showing the functions and operations of the SF / speed estimation unit 24 and the acceleration estimation unit 25. The operation until the corrected acceleration in the traveling direction is obtained from the acceleration and speed obtained by the IMU mounted on the moving body and the vehicle speed detection will be described with reference to FIGS.
2 and 3, the estimation possibility determination circuit 52 in the SF / speed estimation unit 24 is coordinated by the coordinate conversion circuits 54 and 55 of the GPS navigation state quantity 15a and the INS navigation state quantity 26a in S201 of FIG. The value corresponding to the distance obtained by converting the system into the coordinate system of the moving body and the detected vehicle speed estimated value 24a are monitored, the moving body is traveling at a certain speed or more, and the acceleration input is small or the speed fluctuation is small. If the condition is that the vehicle is traveling stably, the information quality is good, and the distance can be calculated, it is determined that the scale factor (SF) can be estimated. As described above, the scale factor is a conversion coefficient used when converting the number of vehicle speed pulses into the distance, but this value varies depending on the behavior of the vehicle that is the moving body, such as the acceleration fluctuation that the moving body receives. ing. Therefore, the coefficient is updated on the assumption that the fluctuation is small when the above stable running condition is satisfied.

If the determination of the stable condition is determined with only one input, there is a possibility of an erroneous determination, but even if not all the conditions are satisfied, if a plurality of representative monitoring target inputs indicate the stable driving condition, The estimability determination circuit 52 outputs a scale factor (SF) update control signal to the internal Kalman filter 53 of the SF / speed estimator 24 in S202. Further, an input switching signal for selection indicating whether to update the scale factor (SF) by using either the converted value of the GPS navigation state quantity 15a or the converted value of the INS navigation state quantity 26a is output.
In S203, the SF calculation circuit 51 forms a closed loop with a Kalman filter 53 in which the conversion value of the INS navigation state quantity or the conversion value of the GPS navigation state quantity is set, and calculates the update value of the scale factor (SF). To do. Thereafter, the updated value of this SF is held, and based on the number of vehicle speed pulses per unit time obtained by the vehicle speed detection unit 23, the coefficient is multiplied by a coefficient, and in S204, the moving body speed estimated value is obtained. 24 a is output to the acceleration estimation unit 25, the estimation availability determination circuit 52, and the internal Kalman filter 53.
Updating the stable SF value following this SF variation greatly contributes to improving the position calculation accuracy of the moving object.

On the other hand, on the acceleration estimation unit 25 side, the stationary determination circuit 42 is S211, which means that the vehicle speed pulse from the vehicle speed detection unit 23 is hardly output, but the moving body speed estimation value 24a is below the threshold value. In addition, the stationary state of the moving object is determined based on a plurality of stationary information indicating that the vehicle has not moved even from the GPS navigation state quantity converted by the coordinate conversion circuit 54.
FIG. 5 is a block diagram showing details of the stationary alignment detection circuit 43. When the stationary determination circuit 42 checks the input signal and determines that it is in the stationary state described above, the update is activated. Further, in S212a, the standard deviation of the IMU traveling direction acceleration 21b is obtained by the standard deviation calculation circuit 62, and a state in which the calculation can be performed by the calculation possibility determination circuit 65 is detected by those time series signals by the delay circuits 63 and 64.
When it is determined that the static alignment can be calculated in this manner, the static alignment calculation circuit 66 measures the bias value 43a using the low-pass filter from the corrected IMU traveling direction acceleration 21b. That is, in S212b, a bias value as a correction amount is obtained from the output of the stationary alignment detection circuit 43.
In this example, as a condition for performing the stationary alignment, the combination of the above three conditions in which the vehicle speed pulse is hardly output, the navigation state quantity is not changed, and the alignment can be calculated has been described. However, as is the case when the SF is updated, there are various variations as inputs for monitoring whether or not the update condition is satisfied, and other appropriate monitoring inputs may be adopted.

For example, FIG. 6 is an example of an actual measurement value in which the horizontal axis indicates time and the vertical axis indicates acceleration. The solid line with many fluctuations in the figure indicates the measured value of the acceleration in the traveling direction, and the measured acceleration constantly fluctuates with time. A horizontal line horizontally in the center indicates the stationary acceleration, that is, the bias value, held by the stationary alignment detection circuit 43. The stillness determination circuit 42 performs a stillness determination in a slight time axis direction at a point A in the figure, and then, as shown in a circled portion, the stationary alignment detection circuit 43 determines the acceleration value at rest. Has been corrected and updated in steps.
As described above, when the static alignment can be calculated, the detection of the bias and the correction of the traveling direction acceleration by the static alignment detection circuit 43 are continuously updated, and the updated value is held. And the correction | amendment by stationary alignment is performed appropriately following the state change of various vehicles accompanying the movement of a moving body.

Further, the speed estimation value is output in S214 by detecting the movement alignment in S213a by the integrator 44 and the movement alignment detection circuit 46 whose detailed configuration is shown in FIG. 7 and the processing in S213b.
The movement alignment detection circuit 46 first samples and holds the detected vehicle speed estimated value 24 a by the sample / hold circuit 71. Then, the delay circuit 72 is used to differentiate the difference between the sample hold value and the integrator 44 obtained by integrating the acceleration corrected by the bias value obtained by further performing stationary alignment on the traveling direction acceleration 21b from the IMU side. The acceleration error estimation circuit 73 is used to output an acceleration error estimation value 73a. Further, the difference between the moving distance obtained by integrating the detected vehicle speed estimated value 24a by the integrator 47 and the distance by the integrator 45 is estimated by the speed error estimating circuit 76 using the sample / hold circuit 74 and the delay circuit 75, A speed error estimated value 76a is obtained.
That is, the estimation calculation performed by the acceleration error estimation circuit 73 and the speed error estimation circuit 76 in FIG. 7 obtains, for example, the mean square value of those samples from the past four points to the latest point. For example, the calculation performed by the speed error estimating circuit 76 is not a polygonal line obtained by obtaining sample values of 5 points indicated by 1 to 5 on the horizontal axis in FIG. It is approximating with a straight thick line which is a linear regression line. The difference between the detected vehicle speed estimated value obtained by the acceleration estimating unit 25 and the distance integrated value by the acceleration from the IMU side is sampled at five points in time series. And what approximated these 5 points | pieces by the primary regression line is shown to Fig.8 (a) thru | or (d) in order by changing time. 8A to 8D is a speed estimated value that is an output of the acceleration estimating unit 25. In FIG. 2, the difference between the output 76a from the speed error estimating circuit 76 and the integrator 44 is the same. It is. The calculation by the acceleration error estimation circuit 73 is the same as that by the speed error estimation circuit 76 described above.

Thus, the acceleration estimation unit 25 outputs two distance estimated values, that is, (IMU) distance estimated value 45a and (VMS) distance estimated value 47a in S215 using the corrected acceleration input from the IMU and the moving body speed estimated value. To do. The SF / speed estimation unit 24 and the acceleration estimation unit 25 constitute an acceleration / distance estimation circuit.
The INS navigation unit 26 may receive the speed estimation value 25b as an input, but the two distance estimation values 45a and 47a from the acceleration estimation unit 25 and the acceleration output of the acceleration estimation unit 25 are also error estimation values of the Kalman filter 16. The acceleration corrected in step 1 is used as an input. Of course, the angular acceleration other than the traveling direction acceleration is also input to the INS navigation unit 26 after being feedback-corrected from the IMU unit 21 with the error estimated value 16b, which is the output of the Kalman filter, and used for the navigation calculation.
Since the INS navigation unit has a general configuration, detailed description of its operation is omitted. However, INS navigation position information is obtained using an estimated acceleration value and a distance estimate value that are considered to be more correct as inputs.

The operation of the GPS system will be described.
As is well known, GPS is a positioning system using satellites that cover the entire earth, and a group of satellites for this purpose is called a GPS satellite. However, since the position is not always near the zenith, a group of three satellites called quasi-zenith satellites can be used.
The quasi-zenith satellite will be explained. The quasi-zenith satellite orbits once a day according to the rotation of the earth so as to have an inclination angle of about 45 degrees from the equator plane. Note that the inclination angle from the equator plane may be arbitrarily set depending on the design. In addition, as an example, three aircraft are arranged so as to be 120 degrees apart at the ascending intersection eclipse (intersection with the equator plane). The locus of the quasi-zenith satellite orbit projected onto the ground surface is orbiting so that the quasi-zenith satellite draws an “eight-letter shape” or “tears drop shape” when the ground is fixed. The three quasi-zenith satellites have different orbital planes, but they change over every 8 hours (meetings), so they are located above Japan without a break. Also, when considering the region in Japan, there will always be a quasi-zenith satellite with an elevation angle of 70 degrees or more. Because it is located above Japan without a break, there is always a quasi-zenith satellite with an elevation angle of 70 degrees or more, and when a receiver receives a radio wave from the quasi-zenith satellite on the ground, it is less likely to be blocked in the valleys of the building. .

In the present embodiment, the DGPS method is used to eliminate the influence of satellite orbit error and ionospheric delay error because of the necessity of processing in the subsequent stage, and the corrected pseudo distance 12a, that is, a more precise pseudo distance, Doppler Observed values can be obtained. FIG. 4 also describes typical types of information obtained from signals emitted from positioning satellites, that is, navigation values such as pseudoranges to measurement points, Doppler excursions, almanac / ephemeris.
The GPS receiving unit 11 outputs the pseudo distance, the Doppler shift, and the carrier phase signal described above as detailed data as the observation amount 12a. It also provides quality information.
Further, the conventional DGPS obtains an error amount at a reference station whose position is determined in advance, and the error amount is the sum of ionospheric delay, troposphere delay, and satellite orbit error. However, ionospheric delay, tropospheric delay, satellite orbit error, etc., increase in value depending on the baseline length, and as the distance from the reference point increases, that is, the farther the moving body is from the reference point, the worse the positioning accuracy. There was a problem such as.

FIG. 9A is an error explanatory diagram obtained by approximating the error distribution in the DGPS method shown in Non-Patent Document 1 by multi-plane approximation, and FIG. 9B is a diagram illustrating an error at a measurement point on a certain error plane. It is. An error plane that estimates a position-dependent error component from observation information of a plurality of reference points and linearly approximates the error component around each reference point is shown. In the figure, if there is information from a plurality of electronic reference points, at the measurement point S at the coordinate values X1, Y1, with respect to the error e3 at the reference point in R3 approximated by the R3 plane, from the reference point to the measurement point. E3 + Δe can be obtained by adding the error vector amount Δe on the surface proportional to the distance. This is because the above-mentioned conventional DGPS simply calculates only e, but the error can be further corrected by Δe.
In the present embodiment, the DGPS unit 12 obtains the wide area correction information as the plurality of reference point information, and obtains e3 + Δe by the above calculation. That is, the fine pseudo-range is corrected by the error obtained by this calculation, and the post-correction pseudo-distance 12a is obtained as more accurate distance information. Since the moving body moves toward the destination in order, the approximate plane is selected in order of R1, R3, and R4 in FIG. A corrected pseudorange 12a having a high value is obtained.
In the DGPS unit, the Doppler deviation, which will be described later, is also corrected more finely than the value obtained by the normal GPS unit output, and therefore a more correct value is output.

Next, multipath will be described. FIG. 10 shows a path of a radio wave coming from a satellite, and shows a directly received path described as a direct signal Vse, and a path reflected and reached by a surrounding building described as a multipath signal. . A signal that arrives after reflection in this way is called a multipath. In multipath, the propagation path is longer than the normal path, which causes an offset in the positioning calculation result. In this figure, only the multipath is received. Thus, the positioning accuracy is improved by identifying the case where multipath is mainly received and removing the signal component.
The so-called Doppler deviation, in which the carrier frequency of the radio wave coming from the satellite changes depending on the movement of the satellite as the transmission source and the movement on the receiving side, can be calculated from the propagation direction components of the movement speeds of both. The Doppler amount due to the movement of the satellite is common to both the direct signal and the multipath signal, and this is Vse. Assuming that the moving speed of the moving body on the receiving side is Vr in the figure, the Doppler amounts generated in the direct signal and the multipath signal are Vre and Vre ′, respectively, as shown. In the case of the figure, it can be seen that the amount of Doppler depending on the moving speed on the receiving side differs not only in absolute value but also in sign.
It is possible to calculate the positioning satellite's coordinates and moving speed by the navigation calendar 11a. Further, since a general positioning satellite is far away, if the approximate position and approximate speed of the positioning device are known, the Doppler amount can be predicted. Therefore, the effect of multipath can be estimated by comparing the Doppler prediction amount with the Doppler amount of the actually received signal.

FIG. 11 is a diagram showing a detailed configuration of the satellite information selection unit 13 that removes a signal considered to be a multipath signal from a signal from the satellite via the DGPS unit 12.
FIG. 12 is a flowchart showing the selection function and operation. In FIG. 12 and FIG. 1, the observation amount 12a that is the output of the DGPS unit 12 includes the corrected pseudorange and the Doppler observation value corrected with the wide area correction information.
Here, the satellite information selection is to delete the pseudorange determined to be received by multipath from the calculation target data as data including an error, and to select the pseudorange information based on the direct signal. For this purpose, the Doppler amount of the GPS signal is estimated and compared with the actual observation value to determine whether it is multipath. For example, when the observed value is significantly different from the estimated value, it can be determined that the observed value is based on multipath. Then, using this determination result, signals from positioning satellites determined to be multipath are discarded, and only correct signals that are direct signals are selected and output to the subsequent stage.

First, the GPS signal is received, and the satellite position calculation circuit 83 calculates the position of the satellite in step S301 based on the navigation calendar 11a included therein. Further, among the position state quantities output from the Kalman filter 16, the position of the moving body is obtained from the position information, and the difference between the two is obtained. Based on this, the line-of-sight vector calculation circuit 84 calculates the line-of-sight vector in S302.
From the navigation calendar 11a, the velocity of the satellite obtained by the satellite velocity calculation circuit 82 is compared with the value indicating the velocity in the state estimated value 16a from the Kalman filter, and a temporary Doppler is calculated from the relative velocity difference between the positioning satellite and the moving body. A quantity is obtained. The dot product estimated value 81a of the line-of-sight vector component is obtained by the inner product circuit 85. That is, the Doppler value converted in the traveling direction is obtained.

The multipath determination circuit 86 calculates the difference between the Doppler estimated value 81a for the traveling direction obtained above and the accurate error-corrected Doppler observation value (12a) obtained from the DGPS signal by the absolute value circuit 87. Obtained in S304. Of the position state quantities output by the previous Kalman filter 16, the gaze vector component of the moving body speed is obtained by the inner product circuit 88. The absolute value of the line-of-sight vector component is obtained by the absolute value circuit 89, and this value is compared with the absolute value of the previous Doppler difference under a predetermined condition. In S305a, the determination circuit 90 makes a comparison determination.
The criterion for determining multipath is that the absolute value difference between the line-of-sight vector components is large. If this difference is greater than or equal to a predetermined value, the multipath determination circuit 86 issues a discard selection control signal 86a, and based on this, the selection circuit 91 discards the pseudo distance as being multipath in S311.
Thus, in S306, only the pseudo distance and related information that are considered to be correct are output from the satellite information selection unit 13 to the next stage pseudo distance estimation unit 14 as the selected pseudo distance 13a. That is, the multipath determination can be correctly performed based on the Doppler observation value corrected by the wide area correction information by the DGPS unit 12.
In order for the above selection to be performed correctly, it is essential to use the corrected Doppler value from the DGPS unit 12. In other words, the GPS unit output Doppler value can only obtain an inaccurate value, but the correct selection can be made by using the Doppler observation value included in the corrected pseudorange 12a corrected by the wide area correction information via the DGPS unit 12. Can do.

FIG. 13A is a diagram showing a detailed configuration of the pseudo distance estimation unit 14, and FIG. 14 is a flowchart showing functions and operations of the pseudo distance estimation unit 14.
The observation amount of the pseudo distance includes the influence of the clock error of the GPS receiver. In FIG. 13A, the pseudo distance conversion circuit 101 converts the clock error amount into a distance value by the clock error estimated value 15b obtained from the calculation result in the GPS navigation unit 15, and the selected pseudo distance 13a from the previous stage is converted by this value. In step S401, the receiver clock error is removed.
The operation of the pseudo distance smoothing calculation will be described based on the detailed configuration diagram shown in FIG. The difference in the carrier phase 12b output from the GPS unit 11 and error-corrected by the DGPS unit 12 is output by the difference between the delay circuit 121 and the subtractor to obtain the fluctuation amount of the carrier phase, and the adaptive calculation circuit 122 As input. In S402, the adaptive calculation circuit 122 weights the time series using the carrier phase fluctuation, the selected pseudo-range corrected by the above-described clock error conversion, and the Kalman filter output state estimation value 16a to perform adaptive calculation. The multiplier K of the multiplication circuit 123 is controlled.
On the other hand, in S403, the pseudo-range prediction value is obtained by adding the fluctuation amount of the carrier wave phase and the delay circuit 124. In S404, the difference between the pseudo distance predicted value and the pseudo distance previously corrected is multiplied by K, which is an output instruction of the adaptive calculation circuit 122, and the multiplier circuit 123 outputs the difference. To obtain a pseudo distance smoothing value 102a which is an output of the pseudo distance smoothing calculation circuit 102.
That is, the noise contained in the pseudorange observation is removed using the carrier phase observation with excellent quality. The two are weighted and added, and the smoothing constant K is adaptively changed in time series in accordance with changes in the behavior of the moving object and the pseudorange, thereby improving the noise removal characteristics.

Returning to FIG. 13A, the difference between the satellite position calculated by the satellite position calculation circuit 103 using the input navigational calendar and the moving body position in the state estimated value 16a is calculated, and the absolute value circuit. To 105. In addition, the elevation angle to the positioning satellite from which the pseudo-range information is obtained by the elevation angle calculation circuit 104 is obtained from the output of the satellite position calculation circuit 103, and the tropospheric error that has not yet been removed by the weighting circuit 106 and the estimation calculation circuit 108 in S405. Is estimated. At the same time, a position error due to an unremoved multipath error or the like is estimated by the weighting circuit 107 and the estimation calculation circuit 109 from the pseudo distance quality information from the GPS unit 11. The total estimation error and the previously compared difference are multiplied by the multiplication circuit 110, and the estimation calculation circuit 111 calculates the mean value of the past to obtain a correction value 111a.
In step S406, the pseudo distance estimation unit 14 finally outputs the pseudo distance estimation value 14a corrected with the correction value 111a, which is used as an input to the GPS navigation unit 15.
In this case as well, the carrier phase and pseudorange may be false pseudorange estimates if values from the GPS unit 11 are used. The correct pseudo-range estimated value 14a is obtained for the first time using the error-corrected value via the DGPS unit 12.

Since the configuration of the GPS navigation unit 15 is well known, its operation will not be described in detail. Since the corrected pseudo-range estimated value 14a is used as an input, a more correct value can be obtained as the obtained GPS navigation position.
As an input to the Kalman filter 16, in the loose coupling, the INS navigation position from the INS navigation unit 26 and the GPS navigation position from the GPS navigation unit 15 are used. In this embodiment, the Doppler value from the pseudo-range estimation unit 14 is also used. The pseudo-range estimated value 14a including the close coupling is also used.
The feedback output of the Kalman filter 16 is used as an estimated state value 16c for input to the INS navigation unit 26 and acceleration correction using the estimated error value 16b. For the GPS system, the estimated value 16a is used as the estimated state value 16a. It is output to the satellite information selection unit 13.

Thus, for a INS system, when a stable input is obtained, the scale factor (SF) is obtained and updated with a reduced error, and the value is held until the next update, and still alignment can be performed. The state is detected, the bias value is updated, the movement alignment is estimated from the feedback amount, and a value that seems to be correct is estimated, and the acceleration error is corrected using these values to obtain a correct acceleration input. That is, when the vehicle speed pulse input is in a predetermined stable input state, the bias factor included in the acceleration signal is measured only when the scale factor is corrected and it is clear that the vehicle is still still. The accuracy of navigation calculation is improved by using an acceleration / distance estimation circuit that corrects the input acceleration signal with a bias value.
For the GPS system, the pseudorange error is corrected by the DGPS unit including wide area correction information from a plurality of electronic reference points, the influence of multipath is excluded, the carrier phase and clock error estimation, and the root mean square calculation of past samples Improve the quality of the input information by value estimation and residual error removal by itself, and also perform pseudo-range estimation integrated with the Kalman filter, which itself uses the pseudo-range estimation value from the pseudo-range estimation unit to improve estimation accuracy, The accuracy of navigation calculation is increased.

In the above description, each component has been described as hardware. However, each component incorporates a software program having the same function in a ROM or a nonvolatile RAM, and reads them out by a microprocessor, as shown in FIGS. The configuration shown in FIG. 14 may be executed. In that case, the function of step S in each flowchart is the same as the function of the element in the corresponding configuration diagram. In addition, the GPS navigation unit 15 and the INS navigation unit may have a function by software by incorporating a program in the same manner as other elements.
Each of the above components may be realized by a combination of hardware and software.

It is a block diagram of the mobile body positioning apparatus in Embodiment 1 of this invention. 3 is a detailed configuration diagram of a scale factor / speed estimation unit and an acceleration error estimation unit in the first embodiment. FIG. 6 is a flowchart showing functions and operations of a scale factor / speed estimation unit and an acceleration error estimation unit according to Embodiment 1. FIG. It is a figure explaining the acceleration which a mobile body receives. 3 is a detailed configuration diagram of a stationary alignment detection circuit according to Embodiment 1. FIG. It is a figure which shows the example of a static acceleration value (bias value) obtained by real acceleration data and a static alignment detection circuit. 2 is a detailed configuration diagram of a movement alignment detection circuit according to Embodiment 1. FIG. It is a figure which shows the example of the inclination coefficient obtained by a movement alignment detection circuit. It is error explanatory drawing which approximated the error distribution in a DGPS system in many planes. It is a figure explaining the influence by multipath and multipath. 3 is a detailed configuration diagram of a satellite information selection unit in Embodiment 1. FIG. 3 is a flowchart showing functions and operations of a satellite information selection unit in Embodiment 1. FIG. 3 is a detailed configuration diagram of a pseudo distance estimation unit according to Embodiment 1. FIG. FIG. 6 is a flowchart showing functions and operations of a pseudo distance estimation unit in the first embodiment.

Explanation of symbols

  12 DGPS section, 13 satellite information selection section, 14 pseudorange estimation section, 15 GPS navigation section, 16 Kalman filter, 24 scale factor (SF) / speed estimation section, 25 acceleration estimation section, 26 INS navigation section, 42 stationary determination circuit, 43 stationary alignment detection circuit, 46 moving alignment detection circuit, 51 SF calculation circuit, 52 estimation possibility determination circuit, 53 Kalman filter, 81 Doppler estimation circuit, 86 multipath determination circuit, 91 selection circuit, 102 pseudo distance smoothing calculation circuit, S201 SF Estimable determination step, S202 calculation input switching setting step, S203 SF update step, S204 moving body speed estimation value calculation step, S211 stationary determination step, S212a stationary alignment detection step, S212b by stationary alignment Correction step, S213a Movement alignment detection step, S213b Movement alignment processing step, S214 Speed estimation value calculation step, S215 Distance estimation value calculation step, S301 Satellite position calculation step, S302 Line-of-sight vector calculation step, S303 Doppler estimation value calculation step, S304 Step of detecting difference between Doppler estimated value and Doppler observed value, S305a Step of comparing difference between estimated value and observed value and line-of-sight vector component of moving body speed, 305b Step of determining whether comparison result is within predetermined value S306 Step of outputting pseudo distance as normal path, S311 Data discarding step as multipath, S401 Carrier phase fluctuation calculation step, S402 Pseudo distance and Kalman filter after phase fluctuation and clock error correction Step S403, an adaptive calculation step based on the state estimation value, S403, a pseudo-range prediction value calculation step based on the delay output and the phase fluctuation, S404, and a prediction after the multiplier K times that is an adaptive calculation result of the difference between the pseudo-range prediction value and the pseudo-range input. A pseudo-range smooth value acquisition step based on a difference from the value, and a pseudo-range estimated value acquisition step based on a difference between the pseudo-range smooth value and the estimation calculation result.

Claims (3)

Used in a positioning device that determines the navigation position of a moving object with a positioning navigation unit that outputs positioning navigation speed and positioning navigation position by positioning navigation, and an inertial navigation unit that outputs inertial navigation speed and inertial navigation position by inertial navigation In the mobile object acceleration / distance estimation circuit,
1)
1-1) A scale factor calculation circuit that inputs a moving body speed detection signal obtained along with the movement of the moving body and obtains a coefficient for converting the moving body speed detection signal into the moving body speed as a scale factor;
1-2) When a plurality of predetermined conditions are satisfied, information including the positioning navigation speed and the positioning navigation position, which are outputs from the positioning navigation section, and the inertial navigation speed and the inertial navigation, which are outputs from the inertial navigation section. A Kalman filter that selects at least one of the information including the position and outputs the selected navigation information and the scale factor calculation circuit output to the scale factor calculation circuit,
The Kalman filter and the scale factor calculation circuit constitute a closed loop. The scale factor calculation circuit receives the output from the Kalman filter as an input, updates the scale factor, and detects the moving body speed based on the updated scale factor. A scale factor / speed estimation unit that adjusts the signal and outputs a moving body speed estimation value;
2)
2-1) A moving body acceleration signal is input, and the moving body is determined to be stationary based on a plurality of signals for detecting the stationary state of the moving body, and the bias value included in the traveling direction acceleration signal when the stationary state is determined. A static alignment detection circuit for detecting
2-2) Using the differential of the difference between the moving body speed estimated value and the value obtained by integrating the traveling direction acceleration signal with the bias value corrected by the output of the stationary alignment detection circuit as an input, the average value of the differential input is In order to further correct the traveling direction acceleration signal and obtain the estimated acceleration of the moving body, an acceleration error estimation circuit that feedback-outputs the average value of the calculated input and the further corrected traveling direction acceleration signal are integrated. The derivative of the difference between the value and the value obtained by integrating the moving body speed estimated value is input, an average value of the differential input is obtained, and the value obtained by integrating the traveling direction acceleration signal is further corrected to obtain the speed estimated value. A speed error estimation circuit that feedback-outputs the average value of the obtained input to obtain, and a movement alignment detection circuit having ,
An acceleration estimation unit that outputs the estimated acceleration of the moving object output from the movement alignment detection circuit and a distance estimation value obtained by integrating the speed estimation value ;
A moving body acceleration / distance estimation circuit comprising: In a positioning device that obtains the navigation position of a moving object by a positioning navigation unit that outputs a positioning navigation speed and a positioning navigation position by positioning navigation, and an inertial navigation unit that outputs an inertial navigation speed and inertial navigation position by inertial navigation.
1)
1-1) A scale factor calculation circuit that inputs a moving body speed detection signal obtained along with the movement of the moving body and obtains a coefficient for converting the moving body speed detection signal into the moving body speed as a scale factor;
1-2) When a plurality of predetermined conditions are satisfied, information including the positioning navigation speed and the positioning navigation position, which are outputs from the positioning navigation section, and the inertial navigation speed and the inertial navigation, which are outputs from the inertial navigation section. A first Kalman filter that selects at least one of the information consisting of the position and outputs the selected navigation information and the scale factor calculation circuit output to the scale factor calculation circuit; ,
The first Kalman filter and the scale factor calculation circuit constitute a closed loop, and the scale factor calculation circuit receives the output from the first Kalman filter as an input and updates the scale factor to obtain the updated scale factor. A scale factor / speed estimation unit that adjusts the mobile body speed detection signal based on the output and outputs a mobile body speed estimation value;
2)
2-1) A moving body acceleration signal is input, and the moving body is determined to be stationary based on a plurality of signals for detecting the stationary state of the moving body, and the bias value included in the traveling direction acceleration signal when the stationary state is determined. A static alignment detection circuit for detecting
2-2) Using the differential of the difference between the moving body speed estimated value and the value obtained by integrating the traveling direction acceleration signal with the bias value corrected by the output of the stationary alignment detection circuit as an input, the average value of the differential input is In order to further correct the traveling direction speed signal and obtain the estimated acceleration of the moving body, an acceleration error estimation circuit that feedback-outputs the average value of the obtained input, and the further corrected traveling direction speed signal Taking the derivative of the difference between the integrated value and the value obtained by integrating the moving body speed estimated value as an input, the average value of this differential input is obtained, and the value obtained by integrating the traveling direction acceleration signal is further corrected to estimate the speed. A moving body alignment circuit having a speed error estimation circuit that feedback-outputs the average value of the obtained input to obtain a value ,
An acceleration estimating unit that outputs the estimated acceleration of the moving object outputted from the mobile alignment detection circuit, the distance estimate obtained by integrating the velocity estimate, and
3) The positioning navigation speed and positioning navigation position, which are outputs of the positioning navigation section, and the inertial navigation speed and inertial navigation position, which are outputs of the inertial navigation section, are input, and an error estimation value is calculated based on the inputs. A second Kalman filter that calculates and outputs the estimated error value,
The inertial navigation unit inputs the estimated acceleration of the moving object output from the acceleration estimating unit, the estimated distance value of the moving object, and the estimated error value output from the second Kalman filter, and estimates the moving object. A movement characterized in that an inertial navigation speed and an inertial navigation position of a moving object are obtained by calculation based on an acceleration, a distance estimated value of the moving object, and an error estimated value, and the obtained inertial navigation speed and inertial navigation position are output. Body positioning device. In a positioning method for determining a navigation position of a moving body by a positioning navigation calculation step that outputs a positioning navigation speed and a positioning navigation position by positioning navigation, and an inertial navigation calculation step that outputs an inertial navigation speed and inertial navigation position by inertial navigation,
Inputting a moving body speed detection signal obtained as the moving body moves;
A scale factor calculation step for obtaining a coefficient for converting the moving body speed detection signal to the moving body speed as a scale factor, and the positioning navigation speed and the positioning navigation position that are outputs of the positioning navigation step when a plurality of predetermined conditions are satisfied. And at least one of the information consisting of the inertial navigation speed and the inertial navigation position, which is the output of the inertial navigation step, and the selected navigation information and the scale factor calculation step output And a first Kalman filter calculation step that outputs to the scale factor calculation step, and the first Kalman filter calculation step and the scale factor calculation step constitute a closed loop, and the scale factor calculation step From the first Kalman filter computation step Update the an input force above scale factor, and the scale factor and speed estimation step of outputting the mobile speed estimation value by adjusting the mobile speed detection signal on the basis of the updated the scale factor,
Inputting a moving direction acceleration signal of the moving object;
A stationary alignment detection step of determining stationary of the moving body based on a plurality of signals for detecting stationary of the moving body, and detecting a bias value included in the traveling direction acceleration signal at the time of the stationary determination ;
Using the difference between the value obtained by integrating the traveling direction acceleration signal corrected in the bias value obtained in the stationary alignment detection step and the estimated moving body speed as an input, the average value of the differential input is obtained, An acceleration error estimating step of feedback-outputting the average value of the obtained input for further correcting the traveling direction acceleration signal to obtain the estimated acceleration of the moving body, a value obtained by integrating the further corrected traveling direction acceleration signal, In order to obtain an estimated speed value by further correcting the value obtained by integrating the traveling direction acceleration signal by obtaining an average value of the differential input with the derivative of the difference from the value obtained by integrating the moving body speed estimated value as an input. A speed error estimation step that feedback-outputs the average value of the obtained inputs, and a distance estimation that integrates the estimated acceleration of the moving object and the speed estimation value by correction by the speed error estimation step. And moving the alignment detection step of outputting a value,
An acceleration estimation step for outputting an estimated acceleration of the moving object output in the moving alignment detection step and a distance estimation value by integrating the speed estimation value of the moving object ;
Input the positioning navigation speed and positioning navigation position, which are the outputs of the positioning navigation step, and the inertial navigation speed and inertial navigation position, which are the outputs of the inertial navigation step, and calculate the error estimates by calculating the inputs. And a second Kalman filter calculation step for outputting the obtained error estimated value,
In the inertial navigation calculation step, the estimated acceleration output from the acceleration estimation step, the distance estimation value, and the error estimation value output from the second Kalman filter calculation step are input, and the estimated acceleration and distance estimation value are input. And a step of outputting the obtained inertial navigation speed and inertial navigation position by calculating the inertial navigation speed and inertial navigation position of the mobile object by calculation using the error estimation value and the error estimation value.

Images