POSITIONING DEVICE
By integrating satellite and behavioral detection sensors to compare and select the most accurate reference point, the positioning device minimizes errors in position estimation, enhancing overall accuracy.
Patent Information
- Application Number
- DE112018003443
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2017-07-07
- Filing Date
- 2018-06-22
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2038-06-22
AI Technical Summary
Existing positioning technologies, such as satellite navigation and autonomous navigation, suffer from errors in position estimation, particularly when satellite navigation data is unreliable, leading to increased errors when combined with tight coupling methods.
A positioning device that estimates the trajectory using both satellite information and behavioral detection sensors, comparing satellite positioning errors with autonomous positioning errors to determine the reference point, thereby reducing overall estimation errors.
The proposed method reduces positioning errors by using satellite information as a reference only when its accuracy is higher than autonomous navigation, ensuring more accurate trajectory estimation.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
TECHNICAL AREA
[0001] The present disclosure relates to a positioning device that reduces positioning errors. STATE OF THE ART
[0002] The technology of a Global Navigation Satellite System (GNSS) receiver, or a receiver of a global navigation satellite system, which receives navigation signals transmitted by the navigation satellites included in a GNSS to determine a current position based on the navigation signals received by the GNSS receiver, is well known. A technique for successively determining position changes of a moving object using this technology can also be referred to as satellite navigation.
[0003] A technique for estimating the trajectory of a moving object based on detection values acquired by various behavior detection sensors that detect the behavior of the moving object, and estimating the position based on a reference point and the trajectory, is also widely known. This technique can also be referred to as autonomous navigation. A technique that combines satellite navigation and autonomous navigation to successively determine position changes is also known.
[0004] JP 2017-003395 A discloses a system that determines the position of a vehicle using a combination of satellite navigation and autonomous navigation. JP 2017-003395 A determines whether satellite navigation measurement data can be used. If the measurement data cannot be used, positioning is switched from satellite navigation to autonomous navigation. In other words, if satellite navigation measurement data is permitted in JP 2017-003395 A, satellite navigation is performed. One of the conditions for determining whether satellite navigation measurement data can be used is that a captured number of navigation satellites is greater than or equal to a number that enables positioning, and the positioning accuracy of the captured navigation satellites is greater than or equal to a predetermined accuracy.
[0005] For example, as described in JP 2013-113789 A and JP 2017-009294 A, a method for position estimation by integrating pre-estimation data, such as pseudoranges obtained from navigation signals, and the magnitude of position change estimated by behavior detection sensors is also known. In this description, this method is referred to as tight coupling. Tight coupling is a type of autonomous navigation because it utilizes the detection values from behavior detection sensors.
[0006] In satellite navigation, the most common method determines the position of a moving object by simultaneously detecting navigation signals from four or more satellites and simultaneously solving equations in which the three-dimensional coordinates and a clock error of the moving object are unknown.
[0007] Furthermore, a method disclosed in JP 2011-209268 A is known as satellite navigation. JP 2011-209268 A discloses a method that utilizes the amount of Doppler frequency shift of navigation signals. The method described in JP 2011-209268 A calculates a vehicle speed vector from the amount of Doppler frequency shift and calculates a vehicle trajectory from the speed vector. The difference between the distance from a point on the vehicle trajectory to each satellite and a pseudorange is used as an evaluation value, and the vehicle trajectory is translated to minimize the evaluation value. This method is also a type of satellite navigation because, while it uses navigation signals, it does not use detection values from behavior detection sensors.
[0008] When position estimation is performed using tight coupling, the position estimation errors are smaller than in autonomous navigation, which does not use navigation signals. Therefore, when a position estimation method is performed using tight coupling as autonomous navigation, performing satellite navigation, if satellite navigation measurement data can be used as in JP 2017-003395 A, can increase the position estimation errors.
[0009] Another positioning device is known from JP 2006-177772 A. The positioning device is carried by a user and includes: a radio wave-using positioning device for receiving an external radio wave and calculating a current position from the reception result by calculation; an autonomous positioning device for integrating the direction of movement and the amount of movement from the initial position to obtain a current position at an initial position;and a control device for controlling the radio wave-using positioning device and the autonomous positioning device, a control device for setting the position obtained from the positioning results as the current position, and an initial position setting device for setting the positioning result more accurately estimated from the positioning results of the radio wave-using positioning device and the autonomous positioning device to the initial position of the autonomous positioning device for each positioning of the radio wave-using positioning device.;
[0010] It is an object of the present disclosure to provide a positioning device capable of reducing errors in position estimation.
[0011] The problem is solved by the subject matter of the main claim. Advantageous further developments are specified in the subclaims.
[0012] The autonomous positioning unit estimates the trajectory based on satellite information—that is, the information prior to calculating the position of the moving object in satellite navigation and the detection values of the behavior detection sensors—rather than simple autonomous navigation. Consequently, the accuracy is higher than with simple autonomous navigation, so the errors can be smaller than those in the position of the moving object determined by satellite navigation.
[0013] The satellite positioning error is compared with the autonomous positioning error. Since the satellite positioning error is comparatively smaller than the autonomous positioning error, the position of the moving object detected by the satellite positioning unit is used as the reference point for the autonomous positioning unit. In this way, the satellite positioning error is compared with the autonomous positioning error to determine whether the position of the moving object detected by the satellite positioning unit should be used as the reference point, thus reducing position estimation errors. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The above and other objects, features, and advantages of the present disclosure will become more apparent from the following detailed description with reference to the accompanying drawings. In the drawings: Fig. 1 is a configuration diagram of a positioning device in an embodiment, Fig. 2 is a flowchart showing the process performed by the positioning device in Fig. 1 executed processing shows, Fig. 3 is a flowchart showing the processing by the positioning device in Fig. 1 parallel to processing in Fig. 2 shows, Fig. 4 is a flowchart showing detailed processing to Fig. 3 to calculate a Doppler velocity, Fig. 5 is a flowchart showing detailed processing to Fig. 3 to determine an initial value from an azimuth estimation formula, Fig. 6 is a flowchart showing the processing after the execution of Fig. 3 shows, Fig. Fig. 7 is a diagram showing an example of a variation of an error in the implementation of the first embodiment, Fig. Fig. 8 is a flowchart showing processing in a second embodiment, which instead of Fig. 6 is executed, shows Fig. Fig. 9 is a diagram showing an example of a variation of an error when the second embodiment is implemented, and Fig. 10 is a diagram showing an example of a variation of an error when a third embodiment is implemented. DETAILED DESCRIPTION OF AN EMBODIMENT
[0015] In the following, exemplary embodiments are described with reference to the drawings. Fig. The positioning device 100 shown in Fig. 1 is mounted together with a GNSS receiver 2, a vehicle speed sensor 3, a gyro sensor 4, an acceleration sensor 5 and a displacement position sensor 6 on a vehicle 1, which is a moving object.
[0016] The GNSS receiver 2 receives navigation signals transmitted by a navigation satellite Si of a global navigation satellite system (GNSS). "i" is the number of the navigation satellite. The GNSS is GPS, for example. A navigation signal is superimposed on a carrier wave and transmitted by the navigation satellite as a radio wave. A radio wave transmitted by a navigation satellite is referred to below as a GNSS radio wave. The GNSS receiver 2 demodulates a received GNSS radio wave to extract a navigation signal.
[0017] From the extracted navigation signal, a pseudorange pi, a Doppler shift amount D i , a satellite position (X si , Y si , Z si), a satellite status, navigation data, etc. are determined. The navigation data includes, for example, the satellite number of the navigation satellite, an ephemeris, which represents orbit information of the navigation satellite, and the time at which the navigation satellite transmitted the radio wave.
[0018] The satellite position (X si , Y si , Z si ) of each navigation satellite Si is calculated based on the ephemeris of the navigation satellite Si and the time at which the GNSS radio wave was transmitted. The pseudorange ρ i is calculated by multiplying the time difference between the time at which the navigation satellite Si transmitted the GNSS radio wave and the time at which the GNSS receiving device 2 received the GNSS radio wave, ie a radio wave travel time, by the speed of light.
[0019] The Doppler shift amount D iis the frequency difference between the frequency of the carrier wave of the radio wave transmitted by the navigation satellite Si and the frequency of the carrier wave of the received GNSS radio wave. The carrier frequency of the radio waves transmitted by the navigation satellite Si is determined in advance. This frequency is stored in advance in a predetermined storage unit contained in the GNSS receiving device 2.
[0020] The GNSS receiver 2 outputs these values to the positioning device 100 at a fixed cycle, along with the S / N of the received navigation signal and the time at which the navigation signal was received. The information output by the GNSS receiver 2 to the positioning device 100 is information before the position is calculated in satellite navigation. Hereinafter, the information before the position is calculated in satellite navigation is referred to as satellite information. The fixed cycle in which the GNSS receiver 2 outputs satellite information is, for example, between 200 milliseconds and 400 milliseconds. There are multiple navigation satellites. The GNSS receiver 2 acquires satellite information from all navigation signals that can be demodulated from GNSS radio waves and outputs all acquired satellite information to the positioning device 100.
[0021] The vehicle speed sensor 3 detects the wheel speed of the vehicle 1. The vehicle speed sensor 3 outputs a signal indicating the wheel speed to the positioning device 100.
[0022] The gyro sensor 4 and the acceleration sensor 5 are inertial sensors. The gyro sensor 4 detects the rotational angular velocities around the yaw, pitch, and roll axes of the vehicle 1 and outputs a signal indicating the detected rotational angular velocities to the positioning device 100. The gyro sensor 4 functions as a yaw rate sensor because it detects the rotational angular velocity around the yaw axis, i.e., the yaw rate.
[0023] The acceleration sensor 5 detects the acceleration generated in the front-to-rear direction of the vehicle 1. In addition to the acceleration in the front-to-rear direction, accelerations in the width direction and vertical direction of the vehicle 1 can also be detected. The acceleration sensor 5 outputs a signal indicative of the detected acceleration to the positioning device 100.
[0024] The shift position sensor 6 detects the shift position of the vehicle 1 and outputs a signal indicating the shift position to the positioning device 100. Based on the signal output by the shift position sensor 6, it is determined whether the direction of travel of the vehicle 1 is forward or reverse.
[0025] The vehicle speed sensor 3, the gyro sensor 4, the acceleration sensor 5 and the gear position sensor 6 are behavior detection sensors 7 that output signals indicating the movement, that is, the behavior of the vehicle 1.
[0026] The positioning device 100 is a computer including a CPU, a ROM, a RAM, etc. (not shown). The CPU executes programs stored in a non-volatile, tangible storage medium such as a ROM, while utilizing a temporary storage function of a RAM. Thus, the positioning device 100 performs the functions of a satellite information acquisition unit 101, a vehicle information acquisition unit 102, a sensor correction value estimation unit 103, a satellite positioning unit 104, an azimuth estimation unit 105, an orbit calculation unit 106, an error estimation unit 107, and a position calculation unit 108. In executing these functions, processes corresponding to the programs stored in the non-volatile, tangible storage medium are executed. Part or all of the functions performed by the positioning device 100 can be configured as a hardware device using one or more ICs or the like. <Overview of processing by positioning device 100>
[0027] The following provides an overview of the processing by the positioning device 100. Part of the processing performed by the positioning device 100 is described in Fig. 2 and the flow charts shown below are described in detail.
[0028] The satellite information acquisition unit 101 acquires the satellite information from the GNSS receiving device 2 in a satellite information acquisition cycle and stores the acquired satellite information in a satellite information storage unit 110. The satellite information acquisition cycle corresponds to the cycle in which the GNSS receiving device 2 outputs satellite information. The satellite information storage unit 110 is a writable storage medium and can be volatile or non-volatile. For example, a RAM can be used for the satellite information storage unit 110.
[0029] The vehicle information acquisition unit 102 acquires the signals detected by the behavior detection sensors 7 in a sensor value acquisition cycle. The sensor value acquisition cycle is shorter than the cycle in which the GNSS receiver 2 outputs navigation signals, e.g., several tens of milliseconds. The vehicle information acquisition unit 102 stores the acquired signals in a vehicle information storage unit 111. The vehicle information storage unit 111 is a writable storage medium and may be volatile or non-volatile. For example, a RAM may be used for the vehicle information storage unit 111. The same storage medium as the satellite information storage unit 110 may be used for the vehicle information storage unit 111. The information stored in the vehicle information storage unit 111 is referred to as vehicle information.
[0030] The sensor correction value estimation unit 103 corrects the detection values output from the behavior detection sensors 7 and includes a yaw rate correction value estimation unit 103a, a pitch rate correction value estimation unit 103b, and a vehicle speed correction value estimation unit 103c.
[0031] The yaw rate correction value estimation unit 103a estimates a correction value for correcting a rotational angular velocity around the yaw axis (ie, a yaw rate) detected by the gyro sensor 4 (hereinafter, yaw rate correction value). The yaw rate correction value is estimated as follows.
[0032] An azimuth difference about the yaw axis is determined from the difference between two movement azimuths of the vehicle 1, which are determined by the satellite positioning unit 104 at the time of positioning. Using the positioning times of the two movement axes t1 and t2, the difference between an integrated value of the angular velocity about the yaw axis of the gyro sensor 4 at time t1, i.e., an integrated yaw rate value, and an integrated yaw rate value at time t2 is calculated.
[0033] If the azimuth difference is considered a true value, the difference between the azimuth difference and the difference between the integrated yaw rate values is an error in the integrated yaw rate value between times t1 and t2. A value obtained by dividing the difference between the azimuth difference and the integrated yaw rate value between times t1 and t2 by t1 - t2 is thus a yaw rate correction value per unit time. The yaw rate correction value estimation unit 103a periodically updates the yaw rate correction value.
[0034] The pitch rate correction value estimation unit 103b estimates a correction value for correcting a rotational angular velocity around the pitch axis (ie, a pitch rate) detected by the gyro sensor 4 (hereinafter, "pitch rate correction value"). The pitch rate correction value is estimated as follows.
[0035] The difference between two pitch angles of the vehicle 1 (hereinafter referred to as the pitch angle difference), which is determined by the satellite positioning unit 104 at the time of positioning, is determined. Secondly, at the two positioning times t1 and t2, a pitch angle estimate at time t1 and a pitch angle estimate at time t2 are estimated from an integrated value of the angular velocity around the pitch axis, i.e., an integrated value of the pitch rate successively output by the gyro sensor 4. For a pitch angle estimate, an integrated value for the pitch rate is not directly used as a pitch angle estimate. A pitch angle estimate at the current time is also estimated using a value for the pitch rate integrated in the past. This is because the pitch angle is influenced by the change in the acceleration of the vehicle 1 and therefore varies greatly.
[0036] Specifically, a pitch angle estimate is obtained as follows. At the time of sensor data acquisition, a time-increasing straight line of an integrated pitch rate value is calculated using the least squares method, using the last integrated pitch rate value and the past integrated pitch rate value. The current time is substituted into the increasing straight line to obtain a pitch angle estimate at the current time. This processing is performed at each of times t1 and t2 to determine a pitch angle estimate at time t1 and a pitch angle estimate at time t2.
[0037] The difference between the difference value between the pitch angle estimate at time t1 and the pitch angle estimate at time t2 and the pitch angle difference calculated from the pitch angles determined by the satellite positioning unit 104 is an error in the integrated pitch rate value between times t1 and t2. A value obtained by dividing the error by t1 - t2 is a correction value for the pitch rate per unit time. The pitch rate correction value estimation unit 103b periodically updates the pitch rate correction value.
[0038] The vehicle speed correction value estimation unit 103c estimates a vehicle speed correction value. The vehicle speed correction value is a value by which the wheel speed detected by the vehicle speed sensor 3 is multiplied. A vehicle speed estimate is calculated by multiplying a detection value of the vehicle speed sensor 3 by the tire circumferential length and the vehicle speed correction value.
[0039] The vehicle speed correction value is estimated as follows. The azimuth estimation unit 105 determines a speed vector from satellite information. A method for determining a speed vector in the azimuth estimation unit 105 will be described later. A traveling direction component of the speed vector of the vehicle 1 is the vehicle speed. A vehicle speed before correction is calculated by multiplying a wheel speed detected by the vehicle speed sensor 3 by the tire circumference. A coefficient by which the vehicle speed before correction is converted to a vehicle speed obtained from a speed vector determined by the azimuth estimation unit 105 is a vehicle speed correction value.
[0040] The satellite positioning unit 104 sequentially calculates the current position of the vehicle 1 using the satellite information acquired by the satellite information acquisition unit 101 and stored in the satellite information storage unit 110. A method for calculating the current position in the satellite positioning unit 104 is called satellite navigation. Various methods for calculating a current position using satellite information can be used. For example, the satellite positioning unit 104 can use a method described in JP 2011-209268 A. Another method can be used to simultaneously acquire navigation signals from four or more satellites and simultaneously solve equations in which the three-dimensional coordinates of the vehicle 1 and a clock error are unknown.The coordinates of the vehicle 1 calculated by the satellite positioning unit 104 are referred to as satellite positioning coordinates Pvs.
[0041] The azimuth estimation unit 105 successively updates the moving azimuth of the vehicle 1 using the vehicle information stored in the vehicle information storage unit 111 or the satellite information stored in the satellite information storage unit 110. For example, for an azimuth φ, the moving azimuth of the vehicle 1 is estimated from the Doppler shift amounts D, which are satellite information. An azimuth estimation method will be described below with reference to Fig. 3 and Fig. 6 described.
[0042] The trajectory calculation unit 106 successively calculates a movement trajectory of the vehicle 1 from the vehicle speed, the azimuth φ, and the pitch angle. The vehicle speed, the azimuth φ, and the pitch angle are calculated either from the satellite information or the vehicle information. When calculating from the vehicle information, the vehicle speed is calculated from the wheel speed, the tire circumference, and the vehicle speed correction value stored in the vehicle information storage 111. The vehicle speed can also be estimated from the Doppler shift amounts D, which are satellite information. For the azimuth φ, the movement azimuth is used, which is successively estimated using the azimuth estimation unit 105.
[0043] When the vehicle speed, the azimuth φ and the pitch angle determined from the vehicle information are used for the calculation of the trajectory, values corrected by the yaw rate correction value, the pitch rate correction value and the vehicle speed correction value estimated by the sensor correction value estimation unit 103 are used.
[0044] The error estimation unit 107 sequentially estimates a position error calculated by the satellite positioning unit 104 (hereinafter, satellite positioning error Es) and a position error during autonomous navigation (hereinafter, autonomous positioning error Ea). Details of the processing of the error estimation unit 107 are described in the description of the flowchart.
[0045] The position calculation unit 108 successively calculates the current position of the vehicle 1. The current position of the vehicle 1 is determined by extending the trajectory from a reference point. Since the trajectory is calculated from the vehicle information, the position is determined by autonomous navigation. Thus, the position calculation unit 108 functions as a unit for autonomous positioning. Since the vehicle information mainly consists of detection values from the inertial sensors, autonomous navigation is also called inertial navigation. The satellite positioning device 104 described above performs positioning without using the vehicle information.
[0046] The reference points in the present embodiment include a real reference point Br and a virtual reference point Bv. It is possible to use only the real reference point Br as the reference point without using the virtual reference point Bv to successively calculate the current position of the vehicle 1.
[0047] The satellite positioning coordinates Pvs are used for the real reference point Br. However, even when calculating the satellite positioning coordinates Pvs, the real reference point Br is not updated if the satellite positioning error Es is greater than or equal to the autonomous positioning error Ea. <Fluss der Verarbeitung durch Positioniervorrichtung 100>
[0048] The positioning device 100 performs the Fig. 2 and those shown in Fig. 3 is carried out in parallel. While, for example, the processing in Fig. 2 is executed, the processing is in Fig. 3 is executed by interrupt processing. First, the Fig. The processing shown in Figure 2 is described.
[0049] The positioning device 100 periodically performs the Fig. 2. A cycle in which Fig. 2 is the cycle of satellite information acquisition. Fig. 2, step S2 (hereinafter, “step” is omitted) is executed by the satellite information acquisition unit 101, steps S4 to S10 and S14 are executed by the satellite positioning unit 104, and steps S12 are executed by the yaw rate correction value estimation unit 103a and the pitch rate correction estimation unit 103b.
[0050] In S2, the satellite information is acquired by the GNSS receiving device 2. In S4, the satellite information acquired in S2 is stored in the satellite information memory 110. Part of the satellite information can be calculated by the positioning device 100 based on navigation signals or navigation data.
[0051] In S6, it is determined whether the signal quality of a navigation signal is high. The signal quality determination is performed for all detected navigation signals. Various criteria can be used to determine signal quality.
[0052] For example, the signal quality determines whether the signal quality is high: (Determination condition J1) the S / N is greater than or equal to a given value, (Determination condition J2) the remainder of the pseudo-distance ρ i is less than or equal to a criteria distance, (Determination condition J3) the elevation angle θ i is greater than or equal to a criteria angle, and (Determination condition J4) a combination of determination conditions J1 to J3 or similar.
[0053] The remainder of the pseudorange ρ i in the determination condition J2, the difference between the distance between the satellite position (X si , Y si , Z si ) of the navigation satellite Si and the satellite positioning coordinates Pvs(X v , Y v , Z v ) of the vehicle and the pseudo-distance ρ i . For the calculation of the residual, the satellite positioning coordinates Pvs (X v , Y v , Z v) of the vehicle is required. The satellite positioning coordinates Pvs of the vehicle are calculated using the same method as a later-described method for calculating the satellite positioning coordinates Pvs in S10. To calculate the remainder of the pseudorange ρ i The satellite positioning coordinates Pvs are calculated beforehand.
[0054] If the remainder of the pseudorange ρ i is large, multipath influence or similar is taken into account and thus determined that the signal quality is not high. In determination condition J1, if the S / N is greater than or equal to the specified value, it is determined that the signal quality is high. In determination condition J3, if the elevation angle θ i is greater than or equal to the criteria angle, determines that the signal quality is high.
[0055] In S8, it is determined whether the number of high-quality navigation signals detected in S6 is four or more. If the number is four or more, the process proceeds to S10.
[0056] In S10, the current vehicle coordinates (X v , Y v , Z v ) from the pseudorange ρ i and the satellite position recorded in S2 (X si , Y si , Z si ) of the navigation satellite Si. The coordinates are called satellite positioning coordinates Pvs.
[0057] The satellite positioning coordinates Pvs are determined by solving four or more simultaneous equations in which the satellite positioning coordinates Pvs(X v , Y v , Z v ) and a clock bias are unknown. The satellite positioning coordinates Pvs(X v , Y v , Z v) can be calculated using different methods.
[0058] In S12, the yaw rate and pitch rate are calculated using the satellite positioning coordinates Pvs calculated in S10 and the previously calculated satellite positioning coordinates Pvs. These are used as the true values, and the differences between the true values and the yaw rate and pitch rate obtained from the detection values of gyro sensor 4 are used as the new yaw rate correction value and the new pitch rate correction value, respectively.
[0059] In S14, 1 is added to k. The initial value of k is 0, i.e., k represents the number of calculations of the satellite positioning coordinates Pvs.
[0060] The following Fig. 3. The Fig. The flowchart shown in Figure 3 is executed repeatedly during the sensor value acquisition cycle. The parameters represent the values at the last updated time, unless otherwise stated.
[0061] In S20, the vehicle information is acquired from the behavior detection sensors 7. In S22, the vehicle information acquired in S20 is stored in the vehicle information storage 111.
[0062] In S24, the amount of relative azimuth angle change is calculated by multiplying a corrected yaw rate, which is determined by adding a yaw rate correction value to a yaw rate determined from a detection value of the gyro sensor 4 obtained in S20, by a time elapsed between the previous execution of S24 and the current time. The amount of relative azimuth angle change becomes a previous relative azimuth angle φ. t(-1) gyro added to determine the relative azimuth angle φ t gyro to update. The relative azimuth angle φ t gyrois a relative azimuth angle at time t with respect to the azimuth of movement of vehicle 1 at a reference time. Thus, the relative azimuth angle φ t gyro a reference azimuth. In S24, a relative pitch angle is also updated in the same way as the relative azimuth angle φ t gyro.
[0063] In S26, it is determined whether the satellite positioning coordinates Pvs have been calculated since the last execution of S26 until the current execution of S26. The satellite positioning coordinates Pvs were calculated in S10. However, to calculate the satellite positioning coordinates Pvs, satellite information must be acquired, and the cycle in which the satellite information is acquired is longer than the cycle in which the vehicle information is acquired. In addition, the satellite positioning coordinates Pvs are not calculated even when satellite information is acquired unless four or more high-quality navigation signals have been acquired. In addition, calculation is also required after acquiring navigation signals, which takes time. Therefore, the determination in S26 may be NO.
[0064] If the determination in step 26 is YES, the process continues with step 28. In step 28, a Doppler velocity is calculated. The Doppler velocity is the vehicle's velocity vector (Vx, Vy, Vz) or the magnitude of the velocity vector. Since the velocity vector (Vx, Vy, Vz) is calculated using the Doppler shift magnitudes D, this is referred to as a Doppler velocity.
[0065] The detailed processing in S28 is in Fig. 4. In Fig. 4, in S282, the direction Ri(θi, φi) of each navigation satellite Si is calculated from the satellite positioning coordinates P calculated in S10 VS (X v , Y v , Z v ) and the satellite position stored in S2 in the satellite information memory 110 (X si , Y si , Z si ) of the navigation satellite Si. The direction Ri of each navigation satellite Si is determined by an elevation angle θ iwith respect to the horizontal direction and an azimuth angle φi with respect to the north direction when the navigation satellite Si is viewed from the vehicle 1.
[0066] In S284 the velocity vector (Vxs i , Vys i , Vzs i ) of each navigation satellite Si from time series data of the satellite position (X si , Y si , Z si ) of the navigation satellite Si, calculated in S4, using a method that uses a derivation of the Kepler equation.
[0067] In S286 the speed Vsat i each navigation satellite Si in the vehicle direction by Vsat i = Ri[Vxs i , Vys i , Vzs i ] T from the direction Ri of the navigation satellite Si calculated in S282 and the velocity vector (Vxs) calculated in S284 i , Vys i , Vzs i) of the navigation satellite Si. T represents the transformation of the matrix.
[0068] In S288, four or more simultaneous equations are formulated from formula (1), and the simultaneous equations are solved. Thus, the velocity vector (Vx, Vy, Vz) of vehicle 1 and a clock shift Cbv can be calculated. t a clock contained in the GNSS receiving device 2. [Formula 1] (Vsat0+D0⋅CIFVsat1+D1⋅CIF⋮Vsati+Di⋅CIF)=(−cosθ0sinϕ0cosθ0cosϕ0sinθ0−1− cosθ1sinϕ1cosθ1sinϕ1sinθ1−1⋮⋮⋮⋮−cosθisinϕicosθisinϕisinθi−1)(VxVyVzCbv)
[0069] In formula (1) D i is the Doppler shift amount, C is a speed of light, and F is the frequency of the carrier wave of radio waves transmitted from the navigation satellite Si. Formula (1) is disclosed in JP 2013-113789 A, and therefore, a derivation method of formula (1) is not described.
[0070] If there are four or more navigation satellites from which GNSS radio waves have been received, the simultaneous equations described above can be solved in formula (1). However, even if GNSS radio waves have been received, GNSS radio waves that do not indicate high signal quality are excluded. Therefore, if GNSS radio waves have been received from four or more navigation satellites that have been determined to have high-quality navigation signals in S8, the vehicle's velocity vector is calculated from formula (1).
[0071] When the velocity vector is obtained, the three-dimensional movement azimuth of the vehicle 1 is obtained. Thus, in S290, the azimuth angle determined from the velocity vector and the pitch angle determined from the velocity vector at which the vehicle 1 is traveling are used as the initial azimuth value φ 0and updated as the initial pitch angle value. The previous processing in Fig. 4 is performed by the azimuth estimation unit 105.
[0072] In S292, the vehicle speed correction value is updated from the speed of the vehicle 1 in the traveling direction, which is determined from the speed vector, and the vehicle speed, which can be calculated from the wheel speed detected by the vehicle speed sensor 3. S292 is executed by the vehicle speed correction value estimation unit 103c.
[0073] The description returns to Fig. 3. If S28 has been executed, the Fig. 6 is executed. Before describing the position update processing, S30 and subsequent steps will be described.
[0074] If it is determined in S26 that the satellite positioning coordinates Pvs have not been calculated, the process switches to S30. In S30, it is determined whether initial values can be updated. Specifically, this determination is to determine whether three or more azimuth estimation formulas shown in formula (2) can be formulated using navigation signals after the initial azimuth value φ has been calculated. 0 was updated. [Formula 2] Vs'=V'wheelGxticos(ϕ0+ϕtgyro)+VtwheelGytisin(ϕ0+ϕtgyro)−Cbv0−At
[0075] Formula (2) is an equation disclosed in JP 2013-113789 A as a formula for calculating the final velocity vector. In formula (2), Vs is a velocity in the satellite direction, t is a time, V wheel is a detection value of the vehicle speed sensor 3, φ 0is an azimuth angle of the vehicle 1 in the direction of travel at an initial time (ie an initial azimuth value), φ gyro is an integrated value of the azimuth angle of the vehicle 1 in the direction of travel, ie a relative azimuth angle, Cbv 0 is a clock shift at the initial time, A is the slope of change of the clock shift over time, and Gx and Gy are respectively an x- and a y-component of a line-of-sight vector from the vehicle to the navigation satellite Si.
[0076] In the present embodiment, the previously described velocity vector of vehicle 1 was calculated in S28. If the velocity vector can be calculated, the azimuth of vehicle 1 can be determined. However, to calculate the velocity vector (Vx, Vy, Vz) of vehicle 1 in S28, the satellite positioning coordinates Pvs are required. Therefore, the velocity vector (Vx, Vy, Vz) of vehicle 1 cannot be calculated unless high-quality GNSS radio waves can be received from four or more navigation satellites.
[0077] For example, in an urban area with many high-rise buildings, the time during which the velocity vector (Vx, Vy, Vz) cannot be calculated may be long. Even in an environment where four or more GNSS radio waves with high signal quality can be received, S26 is only YES in a fixed cycle of 200 milliseconds to 400 milliseconds. The velocity vector (Vx, Vy, Vz) of vehicle 1 may only be calculated in the cycle where S26 is YES.
[0078] On the other hand, the detection values of the behavior detection sensors 7 can be detected in a fixed cycle regardless of the driving environment, namely in a cycle shorter than the cycle in which the velocity vector (Vx, Vy, Vz) can be calculated, for example, every 20 ms.
[0079] The azimuth estimation and the speed estimation are performed based on the detection values of the behavior detection sensors 7 between the calculation of the speed vector (Vx, Vy, Vz) of the vehicle 1 and the next calculation of the speed vector (Vx, Vy, Vz) of the vehicle 1. The output values for the azimuth estimation and the speed estimation are determined in S32.
[0080] In S32, the Fig. 5 is performed. S32 is executed by the azimuth estimation unit 105. In Fig. 5, in S322, the velocity vector (Vxsi, Vysi, Vzsi) of each navigation satellite Si is calculated from the time series data of the satellite position (Xsi, Ysi, Zsi) of the navigation satellite Si calculated in S4.
[0081] In S324, the Doppler shift amount D i, which is contained in the satellite information acquired in S2 and stored in S4, is inserted into the formula (3) to determine the relative velocity Vr i of the vehicle with respect to the navigation satellite Si. [Formula 3] Vri=−Di⋅CIF
[0082] In S326, the current position P(X v , Y v , Z v ) of vehicle 1. When S326 is executed, four or more high-quality navigation signals were not received because S26 was set to NO. The position P(X v , Y v , Z v ) is used to calculate the line of sight vector (Gx i , Gy i , Gz i ) from vehicle 1 to navigation satellite Si in the next S328.
[0083] Because the navigation satellite Si is located at a distance, the current position used to determine the angle between the navigation satellite Si and the vehicle 1 may have low accuracy. Therefore, it may be unnecessary to have received four or more high-quality signals.
[0084] For example, in S326, using four or more navigation signals including all signals that cannot be determined to be of high quality, the current satellite positioning coordinates Pvs of the vehicle 1 are calculated as in S10, and the satellite positioning coordinates Pvs are stored as the current position P(X v , Y v , Z v ) of vehicle 1.
[0085] The position P(X v , Y v , Z v ) of the vehicle 1 can be determined by a method with lower accuracy than the position determination using the pseudorange ρ iDepending on the estimation accuracy permitted in a system or similar, a position error of vehicle 1 in the range of several hundred meters leads to a speed estimation error of 1 m / s or less, which is not a significant problem. For example, the position can be determined from a map or similar or the position P(X v , Y v , Z v ) of the vehicle 1 can be determined from information such as a past positioning history or a beacon.
[0086] Since the accuracy may be low, the position update processing described later ( Fig. 6) here the last updated position P of vehicle 1 as position P(X v , Y v , Z v ) of vehicle 1.
[0087] In S328 the line of sight vector (Gx i , Gyi, Gz i) from vehicle 1 to the navigation satellite Si. The x-component, y-component, and z-component of the line-of-sight vector are calculated from formula (4). [Formula 4] Gxii=1ρti(Xisi−Xtv),Gyti=1ρti(Ytsi−Ytv),Gzti=1ρti(Ztsi−Ztv)
[0088] In formula (4) ρ t i is the pseudorange of the navigation satellite Si at time t and (X t si , Y t si , Z t si ) the satellite position of the navigation satellite Si at time t. These are stored in S4. (X t V , Y t V , Z t V ) is the current position of vehicle 1 at time t and is calculated in S326.
[0089] In S330, the relative velocity Vr calculated in S324 i , the line of sight vector calculated in S328 (Gx i , Gy i , Gz i) and the velocity vector calculated in S322 (Vxs i , Vys i , Vzs i ) of the navigation satellite Si is inserted into the formula (5) to calculate the satellite direction velocity VS i of vehicle 1 in the direction of the navigation satellite Si. [Formula 5] Vsti=DiifC+GxtiVxsti+GytiVysti+Gzt,Vzsti
[0090] The first term on the right side of formula (5) means the relative velocity Vr i of the vehicle 1 with respect to the navigation satellite Si, and the second to fourth terms mean the speed of the navigation satellite Si toward the vehicle 1. The sum of these terms means the speed of the vehicle 1 toward the navigation satellite Si, and thus the formula (5) holds.
[0091] In S332, the satellite direction velocity Vs calculated in S330 i , the relative azimuth angle φ updated in S24gyro and the line of sight vector calculated in S328 (Gx i , Gy i , Gz i ) into formula (2). This results in three unknown parameters, φ 0 , Cbv 0 and A in formula (2).
[0092] Thus, three or more equations are formulated by taking the speed of the satellite direction V Si , the relative azimuth angle φ gyro and the line of sight vector (Gx i , Gy i , Gz i ) into formula (2). Then, simultaneous equations of the three or more equations are solved. As a result, the initial azimuth value φ 0 , which is an unknown parameter in formula (2).
[0093] The unknown parameters φ 0 , Cbv 0and A are also the same at different times after the initial time. Therefore, it is not necessary to formulate three equations simultaneously. If the total number of equations formulated at multiple time points is three or more, the unknown parameters can be determined. Even if the number of observed satellites at each of the three time points (t0, t1, t2) is one, for example, the initial azimuth value φ 0 determined using data from the observed navigation satellites.
[0094] In S332, an initial value for the pitch angle is also determined. One way to determine the initial value of the pitch angle is the same as for the initial azimuth value. That is, the initial value of the pitch angle is determined by formulating three or more equations, in which Gx and Gy are replaced by Gy and Gz, and θ in formula (2) is replaced by φ.
[0095] If the Fig. 3 is completed, the processing shown in Fig. 6 shown position update processing is executed. In Fig. 3, S20 and S22 are the processing of the vehicle information acquisition unit 102, and S24 and the subsequent steps are the processing of the azimuth estimation unit 105.
[0096] The following Fig. 6. In Fig. 6, in S41, an estimated azimuth φ e by estimating the azimuth in which vehicle 1 is moving. If a velocity vector was determined by executing step S28, the estimated azimuth is determined from the velocity vector. An estimated pitch angle is also determined from the velocity vector.
[0097] When a line-of-sight vector from the vehicle 1 to the navigation satellite Si has been calculated by executing S32, an azimuth in which the vehicle 1 is located is determined from the line-of-sight vector and the position of the navigation satellite Si, ie the estimated azimuth φ e .
[0098] On the other hand, if the determination in S30 is NO and S32 is not executed, the estimated azimuth φ e from the last initial azimuth value φ 0 and the relative azimuth angle φ gyro , updated in S24. Furthermore, the estimated pitch angle is calculated from the last initial value of the pitch angle and the relative pitch angle updated in S24.
[0099] In S42 j k in α + j k β is supplemented with 1, which is not 1 or more. The initial value of j kis 0. k is incremented by 1 each time the satellite positioning coordinates Pvs are calculated, as described in S14. Thus, each time the satellite positioning coordinates Pvs are calculated, a new j k generated. The initial value of j k is 0, and j k is currently t k 0. The time t k is the time at which the navigation signals used to calculate the satellite positioning coordinates P VS (t k ) were received. j k is executed every time Fig. 6 increased by 1. j k means how often the processing in Fig. 6 after time t k was carried out. The meanings of α and β are described in S49 and S51.
[0100] In the following S43 an autonomous navigation position P D (t) is updated. The autonomous navigation position P D(t) is updated by returning to a previous autonomous navigation position P D (t-1) in the direction of the estimated azimuth φ e and the estimated pitch angle calculated in S41, this time a distance calculated by multiplying the last estimated value of the vehicle speed by the update cycle of the autonomous navigation position P D obtained is added. The trajectory is updated by processing in S43.
[0101] In S44, it is determined whether the satellite positioning coordinates Pvs have been updated between the previous execution of S44 and the current execution of S44. If the determination in S44 is NO, the process proceeds to S52 without executing S45 through S51. If the determination in S44 is YES, the process proceeds to S45.
[0102] In S45, a satellite positioning error Es is calculated. The satellite positioning error Es can be calculated using various methods. In the present embodiment, the remainder of the pseudorange ρ i (hereinafter a pseudorange residual) as satellite positioning error It uses.
[0103] In step S46, an autonomous positioning error Ea is calculated. The autonomous positioning error Ea is a value obtained by adding an autonomous navigation error to the satellite positioning error Es at the real reference point Br used as the autonomous navigation reference point. The autonomous navigation error is a value obtained by multiplying the duration of autonomous navigation by an autonomous navigation error per unit time. The autonomous navigation error may be a value obtained by multiplying a travel distance in autonomous navigation by an autonomous navigation error per unit distance.
[0104] In the present embodiment, the error of autonomous navigation per unit time or the error of autonomous navigation per unit distance when the estimated azimuth φ e based on the Doppler shift amount D iis estimated, set smaller than if the estimated azimuth φ e is determined from the detection value of the gyro sensor 4. The error of autonomous navigation per unit time or the error of autonomous navigation per unit distance is set in advance by measurement with actual devices or the like.
[0105] The time at which the estimated azimuth φ e based on the Doppler shift amount D i is the time at which S28 is executed or at which S32 is executed. On the other hand, the estimated azimuth φ e determined from the detection value of the gyro sensor 4 when the determination in S30 is NO.
[0106] In S47, it is determined whether the satellite positioning error Es calculated in S45 is smaller than the autonomous positioning error Ea calculated in S46. If the determination is NO, the process goes to S52, and if YES, the process goes to S48.
[0107] In step S48, the reliability of the satellite positioning coordinates Pvs is determined. The reliability of the satellite positioning coordinates Pvs can be calculated using various methods. For example, the reliability of the satellite positioning coordinates Pvs is determined using a preset correspondence between (1) the number of navigation signals used to calculate the satellite positioning coordinates Pvs, (2) the S / N of the navigation signals, (3) the pseudorange residual, and (4) the elevation angle θ. i of the navigation satellite Si or a combination of (1) to (4) and the reliability.
[0108] In S49, an offset term α included in a virtual reference point calculation formula in Formula (7) is updated. Formula (7) shows an equation representing a virtual reference point derivative function F{P} shown in Formula (6) in the present embodiment. [Formula 6] Bv(tk+jkΔt)=F({P},tk+jkΔt) [Formula 7] F{P}=(1−α−jkβ)P(tk)+(α+jkβ)Br(tk)0≤1−α−jkβ≤1, 0≤α+jkβ≤1
[0109] In formulas (6) and (7) k is in a time t k an integer, and the time t k is the time at which the navigation signals used to calculate the satellite positioning coordinates Pvs were received.
[0110] In formula (6), Δt is a cycle in which the position P of vehicle 1 is updated by autonomous navigation. Hereinafter, this cycle is referred to as the autonomous navigation cycle. The autonomous navigation cycle is equal to the sensor value acquisition cycle.
[0111] Since Δt is the autonomous navigation cycle, and j k the number of executions of the processing in Fig. 6 after updating the satellite positioning coordinates Pvs means j kΔt is an elapsed time between the update of the satellite positioning coordinates Pvs and the present time.
[0112] Formula (6) expresses the following. The virtual reference point Bv(t k + y k Δt) at a time t k + y k Δt is a coordinate given by the virtual reference point derivative function F{P} at time t k + y k Δt is represented.
[0113] Formula (7) expresses the following: The function F{P} defined by the virtual reference point derivative at time t k + y k The coordinates shown in Δt are values resulting from the addition of the last satellite positioning coordinates P VS (tk) at this time and the one at time t k determined position P(t k ) of vehicle 1, multiplied by weighting factors.
[0114] Since j kis an integer greater than or equal to 0, the value of -j decreases k β at each increase of j k . In Fig. 6 the movement amount ΔP D of vehicle 1 is calculated in S53 described later. Thus, -j k β is a decrease term whose value depends on the number of calculations of the movement amount ΔPD decreases. On the other hand, j k β is an increase term whose value varies according to the frequency of calculating the movement amount ΔP D increased. (1 - α - j k β) including -j k β, which is a decrease term, is used as the first factor calculation formula, and (α + j k β) including j k β, which is an increase term, is used as the second factor calculation formula.
[0115] In the calculation formula for the virtual reference point (1 - α - j k β), with which the position P(t k) of vehicle 1, and (α + j k β), with which the satellite positioning coordinates Pvs(tk) are multiplied, are called weighting factors. The virtual reference point Bv(t k + y k Δt) after time t k is a position in a by (1 - α - j k β) and (α + j k β) determined relationship between the position P(t k ) of vehicle 1 at time t k and the satellite positioning coordinates P VS (t k ) at time t k .
[0116] The virtual reference point Bv(t k + y k Δt) approaches the satellite positioning coordinates P VS (t k ) with increasing number of calculations of the movement amount ΔP D .
[0117] In the calculation formula of the virtual reference point, the offset term α is independent of j kconstant. The larger the offset term α, the closer the virtual reference point Bv(t k + y k Δt) at the satellite positioning coordinates P VS (t k ), if j k = 0.
[0118] The reason for calculating the virtual reference point Bv(t k + y k Δt) is to prevent the position of vehicle 1 from changing discontinuously on a larger scale. The lower the accuracy of the satellite positioning coordinates P VS (t k ), the more the position of vehicle 1 changes discontinuously on a larger scale. Therefore, the more reliable the satellite positioning coordinates P VS (t k ), the closer the virtual reference point Bv(t k + y k Δt) to the satellite positioning coordinates P VS (t k ) can be brought.
[0119] In S49, the value of α is increased according to the reliability of the satellite positioning coordinates Pvs. Specifically, a correspondence between the reliability of the satellite positioning coordinates Pvs and α is determined in advance, and the value of α is updated based on the reliability determined this time and the correspondence. The correspondence between the reliability of the satellite positioning coordinates Pvs and α is a relationship in which the larger α is, the greater the reliability of the satellite positioning coordinates Pvs is.
[0120] In S50, the reliability of the movement amount is determined. The movement amount reliability means the reliability of the movement amount ΔP calculated in S53. D, which will be described later. During autonomous navigation, errors accumulate. That is, the longer the period in which the position P of the vehicle 1 is updated by autonomous navigation, the greater the error. Therefore, in S50, the longer the period in which the position P of the vehicle 1 is updated only by autonomous navigation without updating the satellite positioning coordinates Pvs, the lower the reliability of the movement amount. As the period in which the satellite positioning coordinates Pvs are not updated, the value of j updated in S42 can be used. k be used.
[0121] The higher the movement amount reliability in S50 determined in S51, the smaller the value to which β is updated. This is because the need to set the virtual reference point Bv(t k + y kΔt) to bring the satellite positioning coordinates Pvs(tk) close to the satellite positioning coordinates Pvs(tk) early on, the lower the higher the reliability of the movement amount.
[0122] Specifically, in S51, an updated β is determined from the preset relationship in which the higher the reliability of the movement amount is, the smaller the β is stepwise or continuously, the smaller the reliability of the movement amount becomes, and the reliability of the movement amount is determined in S50.
[0123] In S52 the position P(t k ) of vehicle 1 at time tk and the one from the at time t k satellite positioning coordinates P calculated from received navigation signals VS (t k ) into the last formula for calculating the virtual reference point Bv(t k + y k Δt) at time t k + y k Δt is used.
[0124] In S53 the movement amount ΔPD , by which the vehicle 1 moves from the time tk to the present time t k + y k Δt is calculated from formula (8). As can be seen from formula (8), the amount of movement ΔP D with the autonomous navigation position P calculated in S43 D calculated. [Formula 8] ΔPD=PD(tk+jkΔt)−PD(tk)
[0125] In S54, the current position P of the vehicle 1 is updated using the formula (9). [Formula 9] P=ΔPD+Bv(tk+jkΔt)=PD(tk+jkΔt)−PD(tk)+Bv(tk+jkΔt)
[0126] As shown in formula (9), the virtual reference point B V (t k + y k Δt) is used as a reference point, which is the base point of the movement amount ΔP D and the virtual reference point Bv(t k + y k Δt) approaches with increasing j k the satellite positioning coordinates P VS (t k). Consequently, as described in JP 2017-009 294 A, it is possible to prevent a trajectory representing the position P of the vehicle 1 from changing gradually.
[0127] In S55, the position P of vehicle 1 determined in S54 and an error in the position P are output to a device that executes an application with the position P of vehicle 1 and the error in the position P. The error is a smaller value of the satellite positioning error Es and the autonomous positioning error Ea calculated in S44 and S46. In addition to the position P of vehicle 1 and the error, the last azimuth, the last azimuth error, the last yaw angle, the last pitch angle, the last vehicle speed, etc., can also be output together.
[0128] In Fig. 6, S41 is processing of the azimuth estimation unit 105, S43 is processing of the orbit calculation unit 106, S45 and S46 are processings of the error estimation unit 107, and the rest is processing of the position calculation unit 108. <Zusammenfassung der Ausführungsform>
[0129] Since in the present representation the initial azimuth value φ 0 and the initial pitch angle value using the Doppler shift amount D i , i.e. the satellite information, the estimated azimuth φ eand the estimated pitch angle can be obtained with high accuracy. This increases the accuracy of the trajectory and thus reduces the increase in error over time or during autonomous navigation with the trajectory as the movement distance increases. This also allows the increase in error caused by not updating the reference point when determining the satellite positioning coordinates Pvs to be further reduced.
[0130] In the present embodiment, when updating the satellite positioning coordinates P VS(S44: YES) The satellite positioning error Es is compared with the autonomous positioning error Ea. If it is determined that the satellite positioning error Es is smaller than the autonomous positioning error Ea, S48 and subsequent steps are executed to update the virtual reference point derivative function F{P}. The virtual reference point Bv is used as the reference point for calculating the position P of vehicle 1 in formula (9).
[0131] If the satellite positioning error Es is smaller than the autonomous positioning error Ea, the updated satellite positioning coordinates Pvs are used as the reference point for calculating the position P of the vehicle 1. In other words, even if new satellite positioning coordinates Pvs are obtained, if the satellite positioning error Es is greater than the autonomous positioning error Ea, the satellite positioning coordinates Pvs are not used as the reference point for calculating the position P of the vehicle 1. This can reduce errors in position estimation.
[0132] Fig. Figure 7 shows an example of the error variation in the present embodiment. In Fig. In Figure 7, the horizontal axis represents time. As vehicle 1 travels, the travel distance increases over time. Thus, the horizontal axis can be considered the travel distance.
[0133] The vertical axis in an upper diagram of Fig. Figure 7 illustrates the possibility of satellite positioning. The possibility of satellite positioning corresponds to the number of navigation signals that can be used for positioning. Whether a navigation signal may be used for positioning can be determined by a criterion for determining the signal quality. A dashed line in the upper graph of Fig. 7 indicates a threshold of satellite positioning possibility at which satellite positioning can be determined as possible.
[0134] The vertical axis in a lower diagram of Fig. Figure 7 illustrates the error in the detected position. A solid line represents an error in the present embodiment. A dashed line represents an error in a comparative example. The comparative example here is a positioning method that continuously updates a reference point with the satellite positioning coordinates Pvs when the satellite positioning possibility exceeds the threshold.
[0135] Since the possibility of satellite positioning between times t0 and t10 does not exceed the threshold, autonomous navigation is performed without updating the reference point. In the description of the Fig. 7 to 10, autonomous navigation means that the autonomous navigation continues without updating the reference point, and satellite positioning means that the autonomous navigation is performed while updating the reference point to the latest satellite positioning coordinates Pvs.
[0136] Since autonomous navigation continues from t0 to t10, the error increases linearly at a constant slope. Since the satellite positioning possibility exceeds the threshold at t10, the satellite positioning coordinates Pvs are updated. This reduces the error. After that, the reference point is continuously updated to the satellite positioning coordinates Pvs obtained by satellite positioning. Therefore, the error does not increase over time from t10 to t12. In the comparison example, the reference point is continuously updated with the satellite positioning coordinates Pvs obtained by satellite positioning until t20. However, even when the threshold is exceeded, the error of the satellite positioning coordinates Pvs is not constant and varies due to various factors.
[0137] In the example of Fig. 7, from t12 to t13, satellite positioning remains possible, but the satellite positioning error Ea is larger than at time t12. Although the autonomous navigation error also increases from t12 to t13, the autonomous positioning error Ea is smaller than the satellite positioning error Es indicated by the dashed line. Therefore, in the present embodiment, autonomous navigation is performed from t12 to t13, although satellite positioning is possible.
[0138] Similarly, autonomous navigation is performed from t14 to t20 even though satellite positioning is possible. Therefore, the error of the present embodiment at time t20, when satellite positioning becomes impossible, is smaller than that of the comparative example.
[0139] At t30, satellite positioning becomes possible, and the system temporarily switches to satellite positioning. However, between t31 and t32, when the satellite positioning error is large, autonomous navigation is performed even though satellite positioning is possible. From t32 to t34, satellite positioning is performed, but at t34, the system switches back to autonomous navigation.
[0140] Even if satellite positioning is possible, the positioning device 100 of the present embodiment does not update the reference point to the satellite positioning coordinates Pvs when the satellite positioning error Es is larger than the autonomous positioning error Ea, so that the error can be made smaller than that of the comparative example. <Zweite Ausführungsform>
[0141] A second embodiment will be described below. In the description of the second and subsequent embodiments, elements with the same reference numerals as those used previously are identical to the elements with the same reference numerals in the previous embodiment(s), unless otherwise noted. When only a part of the configuration is described, the previously described embodiment(s) may be applied to the other part of the configuration.
[0142] The second embodiment differs from the first embodiment in that the Fig. 8 is added in S56 depending on the position calculation unit 108.
[0143] In the second embodiment, even if it is determined in the determination in S47 that the satellite positioning error Es is smaller than the autonomous positioning error Ea, S48 and the subsequent steps are not executed immediately and the determination in S56 is further performed.
[0144] In S56, it is determined whether a state where the satellite positioning error Es is smaller than the autonomous positioning error Ea has continued for more than a continuation determination time. The purpose of this determination is to consider that the satellite positioning error Es may be unstable immediately after satellite positioning becomes possible. The reason why the satellite positioning error Es may be unstable immediately after satellite positioning becomes possible is, for example, the following reason. Specifically, in a state where satellite positioning is not possible, the number of observable satellites may be small due to obstacles such as buildings, and immediately after the number of observable satellites increases, observation may only be temporarily possible due to multipath effects.Since the determination in S56 is performed for this purpose, the continuation determination time is a time for determining that satellite positioning is stable in a possible state after satellite positioning becomes possible. The continuation determination time is determined empirically by measuring with actual equipment. The continuation determination time is set to 1 second, for example.
[0145] If the determination in S56 is NO, the process proceeds to S52, as if the determination in S47 is NO. However, if the determination in S56 is YES, S48 and subsequent steps are executed to update the real reference point Br to the latest satellite positioning coordinates Pvs.
[0146] Fig. 9 shows an example of the error variation in the second embodiment. A lower diagram of Fig. 9 corresponds to the lower diagram of Fig. 7, where the horizontal axis represents the time or travel distance and the vertical axis represents the error. An upper graph of Fig. 9 is identical to the upper graphic of Fig. 7. In Fig. 7 and Fig. 9 means times marked with the same numerical value represent the same time. A dot-dash line in the lower graph of Fig. 9 represents the solid line in the lower graph of Fig. 7.
[0147] In Fig. 7, the transition from autonomous navigation to satellite positioning occurs at t10. However, in the second embodiment, autonomous navigation continues until t11, at which point the continuation determination time has expired because satellite positioning became possible at t10.
[0148] In Fig. 7, the system temporarily switches to satellite positioning between t30 and t31. However, the time between t30 and t31 is shorter than the continuation determination time, so that in the second embodiment, autonomous navigation is not terminated at t30. In Fig. 7 switches to satellite positioning at t32. In the second embodiment, autonomous navigation continues until t35.
[0149] The autonomous positioning error Ea is calculated with respect to the satellite positioning error Es at the time of switching to autonomous navigation. Thus, in the first embodiment, as shown at t34, when the satellite positioning error Es temporarily becomes small, the autonomous positioning error Ea calculated after t34 becomes small. However, the satellite positioning error Es temporarily becomes small at t34, and the satellite positioning error Es is not in a stable state. This satellite positioning error Es has a small value, but its reliability is low. Consequently, the reliability of the autonomous positioning error Ea calculated after t34 is also low.
[0150] In the second embodiment, if a state in which the satellite positioning error Es is smaller than the autonomous positioning error Ea continues for longer than the continuation determination time, the system switches to satellite positioning. Thus, autonomous navigation continues until t35 without switching to satellite positioning.
[0151] The error in the solid line at t35 in Fig. 9 is larger than the error at t35 in Fig. 7, which is also indicated by the dash-dot line for comparison in Fig. 9. However, it can be said that the value shown by the solid line in Fig. 9 displayed errors are more reliable in the error value. <Dritte Ausführungsform>
[0152] The third embodiment differs from the first embodiment in the method for calculating the satellite positioning error Es calculated in S45. In the first embodiment, the pseudorange residual is used as the satellite positioning error Es. On the other hand, in the third embodiment, a value obtained by additionally multiplying the pseudorange residual by a precision degradation factor is used as the satellite positioning error Es.
[0153] The precision degradation factor is a value determined based on at least one of the following criteria: the magnitude of variation in the pseudorange residual, the S / N ratio of the navigation signals, and the dilution of precision (DOP). DOP is also referred to as the precision degradation rate. The magnitude of variation in the pseudorange residual can be expressed, for example, by the variance of the pseudorange residual.
[0154] The relationship between the magnitude of the variation in the pseudorange residual, the S / N ratio of the navigation signals and / or DOP, and the precision degradation factor is set in advance. This relationship is, for example, a table in which the precision degradation factor is determined from the magnitude of the variation in the pseudorange residual, the S / N ratio of the navigation signals and / or DOP.
[0155] The greater the variation magnitude of the pseudorange residual, the greater the positioning error of satellite navigation. Furthermore, the greater the S / N of the navigation signals and / or DOP, the greater the positioning error of satellite navigation. Therefore, in the third embodiment, the precision degradation factor is determined based on at least one of the following factors: the variation magnitude of the pseudorange residual, the S / N of the navigation signals, and DOP. A value obtained by multiplying the pseudorange residual by the precision degradation factor is used as the satellite positioning error Es.
[0156] Fig. 10 shows an example of the error variation in the third embodiment. A lower diagram of Fig. 10 corresponds to the lower diagram of Fig. 7, where the horizontal axis represents time or travel distance and the vertical axis represents error. An upper diagram of Fig. 10 is identical to the upper diagram of Fig. 7. In Fig. 7 and Fig. 10 means times marked with the same numerical value represent the same time. A dot-dash line in the lower diagram of Fig. 10 represents the solid line in the lower diagram of Fig. 7. A dashed line in the lower diagram of Fig. 10 represents the dashed line in the lower diagram of Fig. 7, ie the satellite positioning error Es in the first embodiment, and a dash-dot-dot line in the lower diagram of Fig. 10 illustrates the satellite positioning error Es in the third embodiment.
[0157] The satellite positioning error Es of the third embodiment is a value obtained by multiplying the pseudorange residual by the precision degradation factor. As shown by the dashed line and the dash-dot-dot line in Fig. 10, therefore, the satellite positioning error Es of the third embodiment (the dashed-dotted line) is larger than the satellite positioning error Es of the first embodiment (the dashed line) in which the pseudorange residual is directly used as the satellite positioning error Es.
[0158] In the third embodiment, autonomous navigation is more likely to continue than in the first embodiment. Thus, autonomous navigation continues until t33 in the third embodiment. Satellite positioning is performed from t33 to t34, and autonomous navigation is performed from t34. The satellite positioning error Es at t34 is larger than the satellite positioning error Es of the first embodiment, and thus the error of the third embodiment indicated by the solid line after t34 is larger than the error of the first embodiment indicated by the dashed line. However, it can be said that the error indicated by the solid line in Fig. 10 displayed errors are more reliable in the error value.
[0159] It should be noted that a flowchart or flowchart processing in the present application contains sections (also referred to as steps), each represented, for example, as S2. Furthermore, each section can be divided into several subsections, while several sections can be combined into a single section. Furthermore, each of the sections thus configured can also be referred to as a device, module, or means. <Erste Modifikation>
[0160] Autonomous navigation can be performed without using the virtual reference point Bv. S48 to S52 in Fig. 6 are omitted. In S53, a trajectory is determined after determining the last satellite positioning coordinates Pvs. In S54, the last satellite positioning coordinates Pvs are used as a reference point, and a position shifted from the reference point by the trajectory is set as the current position P. <Zweite Modifikation>
[0161] The position P of vehicle 1 can be determined using quadratic coordinates. In this case, the pitch angle and pitch rate are unnecessary. The position of vehicle 1 can be determined taking the roll angle into account. A roll angle, a roll velocity, and a roll rate correction value can be calculated in the same way as for a yaw angle, a yaw rate, and a yaw rate correction value. <Dritte Modifikation>
[0162] In the embodiments described above, the moving object is a car. The moving object may be something other than a car.
Claims
[1] Positioning device comprising: a satellite positioning unit (104) configured to obtain a position of a moving object determined by satellite navigation based on navigation signals received by a GNSS receiving device (2); an autonomous positioning unit (108) configured to to estimate a trajectory of the moving object based on satellite information, which is information before the position of the moving object is calculated by satellite navigation, and detection values from behavior detection sensors (7), and to estimate the position of the moving object based on a reference point and the trajectory of the moving object; an error estimation unit (107) configured to estimate a satellite positioning error as an error in the position of the moving object determined by the satellite positioning unit and to estimate an autonomous positioning error as an error in the position of the moving object estimated by the autonomous positioning unit; and an azimuth estimation unit (105) configured to estimate a movement azimuth of the moving object based on Doppler shift amounts of the navigation signals, wherein the autonomous positioning unit is further configured to: setting the position of the moving object obtained by the satellite positioning unit as the reference point in response to the satellite positioning error being smaller than the error of autonomous positioning; to estimate the trajectory of the moving object based on the azimuth of movement of the moving object and a speed of the moving object; to obtain a detection value of a yaw rate sensor included in the behavior detection sensors; estimate the trajectory of the moving object based on the azimuth of movement of the moving object estimated by the azimuth estimation unit when the azimuth estimation unit estimates the azimuth of movement of the moving object; and estimate an amount of azimuth angle change of the moving object based on the detection value of the yaw rate sensor and determine the moving azimuth of the moving object based on the amount of azimuth angle change of the moving object and a past reference azimuth when the azimuth estimation unit does not estimate the moving azimuth of the moving object, and the error estimation unit is further configured to: calculate the autonomous positioning error by adding an autonomous navigation error to the satellite positioning error at the reference point; and to set the autonomous navigation error so that it is smaller in a situation where the autonomous positioning unit estimates the trajectory of the moving object based on the moving azimuth of the moving object estimated by the azimuth estimation unit than in a situation where the autonomous positioning unit determines the moving azimuth of the moving object based on the detection value of the yaw rate sensor. [2] The positioning device according to claim 1, wherein the autonomous positioning unit is configured to set the position of the moving object obtained by the satellite positioning unit as the reference point in response to the satellite positioning error being smaller than the autonomous positioning error for a time longer than a preset continuation determination time period. [3] Positioning device according to claim 1 or 2, wherein the error estimation unit is configured to calculate the satellite positioning error by multiplying a pseudorange residual by a precision degradation factor, and wherein the precision degradation factor is determined based on at least one of the elements, magnitude of variation in the pseudorange residual, S / N of the navigation signals, or precision degradation rate.
Citation Information
Patent Citations
Position estimating device and program
JP2011209268A
Solid-state image pickup device, image pickup device and signal reading method
JP2013009294A
Speed estimation device and program
JP2013113789A
Vehicle positioning system
JP2017003395A
Positioning system and position information system using the same
JP2006177772A
