Method for determining navigation data using a GNSS system

Abstract

Method for determining navigation data in a vehicle (104) with a GNSS system (102), wherein the method models a measurement noise of a vertical velocity of a vehicle (104) and the method comprises at least the following steps: a) Modeling a constant part of the measurement noise of the vertical velocity, reflecting the vehicle's driving situations (104) with a vertical dynamics; b) Modeling a time-varying part of the measurement noise of the vertical velocity, wherein the time-varying part is dependent on a vehicle state and reflects driving situations of the vehicle (104) with a higher vertical dynamics compared to the vertical dynamics in step a); c) Determining a modeled total measurement noise (R) neu) as a combination of the constant part of the measurement noise determined in step a) and the time-varying part of the measurement noise determined in step b); and d) Determining the navigation data taking into account the modeled total measurement noise (R) neu ).

Description

State of the art

[0001] The invention relates to a novel method for determining navigation data using a GNSS system.

[0002] The Kalman filter (also Kalman-Bucy filter, Stratonovich-Kalman-Bucy filter or Kalman-Bucy-Stratonovich filter) is a mathematical filter model for the iterative estimation of system states based on faulty input data, which are in particular sensor data.

[0003] The Kalman filter is used to estimate system variables that cannot be measured directly, while optimally reducing errors in observations. Based on the input data, the Kalman filter typically maintains an internal mathematical model as a constraint. This model is incorporated into the parameter estimation and takes into account dynamic relationships between system variables. For example, this mathematical model represents equations of motion that describe the relationship between changing positions and velocities or accelerations. Error-prone input data relating to these positions, velocities, and / or accelerations can thus be correlated within the Kalman filter. In this way, precise estimates can be generated from (error-prone) input data for navigation data (especially positions and velocities).

[0004] Kalman filters are particularly useful for the iterative estimation of system states based on observations that are typically subject to errors. In this context, Kalman filters have proven especially advantageous for applications where sensor information from various sensors needs to be combined (or fused) with model information. Furthermore, Kalman filters are frequently used in embedded systems because their computations are efficient, accurate, and robust. Microcontrollers can also perform Kalman filter computations with advantageous efficiency. Therefore, Kalman filters are often implemented on microcontrollers.

[0005] Kalman filters are primarily used to fuse data from various sensors that determine vehicle positions, thereby generating highly precise navigation data. Navigation data here refers specifically to position data, as well as data concerning speed and acceleration. Sensors whose data can be processed or fused using Kalman filters include GNSS sensors for determining navigation data (or localization) via GNSS satellites, inertial sensors, and, for example, wheel sensors and steering angle sensors used in motor vehicles to monitor the vehicle's movement via the chassis. These sensors thus generate a kind of precursor data for navigation data, which is used as input data in a Kalman filter to determine highly precise navigation data.By jointly considering this described data in a Kalman filter using the system models, parameters and / or system states stored in the Kalman filter, it becomes possible to generate highly precise navigation data from the described input data.

[0006] Kalman filters are used in particular to obtain highly precise navigation data for highly automated driving functions of motor vehicles and especially for autonomous driving functions.

[0007] In a Kalman filter, a continuous correction of internal parameters and system states is typically performed. The term "corrected parameters" is used here to refer to all corrections taking place within the Kalman filter. The quality of the internal parameters and / or system states in the Kalman filter is crucial for improving navigation data through the use of the Kalman filter.

[0008] However, there is a fundamental need for improvement regarding the precision and confidence of the navigation data obtained with Kalman filters, because for applications of highly automated and autonomous driving, very high precision and high confidence of the navigation data used is generally desirable.

[0009] For example, the vertical velocity in the vehicle's fixed coordinate system (e.g., ISO 8855) is supported only by GNSS and not by other sensors such as odometers or steering angle sensors. This can lead to drift in the navigation solution (position / speed / orientation) in terms of vertical velocity and altitude, resulting in increased errors. While virtual zero measurements of the vertical velocity are fed into the Kalman filter (so-called Non-Holonomic Constraints, NHC) to represent the actual vehicle dynamics, where the average vertical velocity in the fixed coordinate system is zero, this can lead to short-term fluctuations in vertical velocity depending on the vehicle's dynamics (e.g., driving over bumps, crests, concrete slabs, etc.). Therefore, the assumption of a zero measurement is more or less prone to error depending on the vehicle's dynamics. Disclosure of the invention

[0010] Based on this, a particularly advantageous method for determining navigation data using a GNSS system will be described.

[0011] The invention relates to a method for determining navigation data in a vehicle with a GNSS system, wherein the method models a measurement noise of a vertical speed of the vehicle and the method comprises at least the following steps: a) Modeling a constant part of the measurement noise of the vertical velocity that reflects a driving situation of the vehicle with a vertical dynamic; b) Modeling a time-varying part of the measurement noise of the vertical velocity, wherein the time-varying part is dependent on a vehicle state and reflects driving situations of the vehicle with a higher vertical dynamics compared to the vertical dynamics in step a); c) Determining a modeled total measurement noise as a combination of the constant part of the measurement noise determined in step a) and the time-varying part of the measurement noise determined in step b); and d) Determining the navigation data taking into account the modeled total measurement noise.

[0012] For the purposes of this application, the term "navigation data" can be understood to mean geographical positions and / or speeds and / or accelerations of the vehicle that can be determined using the method according to the invention.

[0013] Vertical dynamics are often referred to, particularly in relation to a Cartesian coordinate system, as dynamics in the Z-direction. Accordingly, the vertical direction is usually, and for the purposes of this application, also referred to as the Z-direction.

[0014] The constant component of the measurement noise in step a) reflects, in particular, driving situations of the vehicle with low vertical dynamics. This means that the vehicle moves little in the Z-direction. In other words, the vehicle is preferably moving on a level road surface and exhibits a smooth driving profile in the Z-direction.

[0015] In contrast, the term "higher vertical dynamics" as used in this application can be understood to mean that the vehicle moves more in the Z-direction than in the vertical dynamics of step a). In other words, the vehicle is traveling on an uneven road surface, for example, over bumps, unevenness, or crests. Such situations are represented in the time-varying part of the measurement noise, which is modeled according to step b). The modeling of the time-varying part of the measurement noise in step b) and the modeling of the constant part of the measurement noise in step a) are performed separately and are then combined in step c) to form the modeled total measurement noise R. neu combined. This total measurement noise R neuThe vertical dynamics are then taken into account in step d) in the Kalman filter to process (uncertainty-laden) input data for determining navigation data. The total measurement noise R determined in step c) neu is preferably considered as part of a system model in the Kalman filter.

[0016] The fundamental approach of the method described here is to improve the determination of navigation data, particularly with regard to position determination in the Z-direction. Position determination in the Z-direction is normally dependent on position determination in the XY directions due to the road topology. However, poor modeling of measurement noise in the Z-direction (which is technically more complex than modeling measurement noise in the XY directions) negatively impacts the overall processing of data for position determination. This negative impact can be reduced by improving the modeling of measurement noise in the vertical direction (Z-direction).

[0017] This improves the determination of navigation data and thus the overall position determination. Furthermore, the invention is based on the consideration that the assumption that the vertical velocity in a vehicle's fixed coordinate system is zero only holds true on average across all measurements. Depending on the vehicle's dynamics, for example, when driving over uneven surfaces, this can lead to short-term fluctuations in vertical velocity. Therefore, assuming a zero measurement depending on the vehicle's dynamics is more or less prone to error. The resulting error distribution is thus dependent on the vehicle's dynamics and is not normally distributed. The method according to the invention takes precisely this circumstance into account when determining navigation data and thus the vehicle's position.

[0018] According to one embodiment, the vehicle's vertical dynamics are acquired for modeling the measurement noise in steps a) and b) using at least one sensor. This sensor acquires parameters reflecting the vertical dynamics. The sensor unit may also include several individual sensors. Such individual sensors within a sensor unit can be arranged at different positions within the vehicle. This allows parameters acquired at different locations within the vehicle to be used for a particularly accurate determination of the vertical dynamics.

[0019] For capturing the vertical dynamics, velocities and rotation rates are recorded, for example, but this is not the only possible measurement. Preferably, the at least one sensor or sensor unit can be an inertial measurement unit (IMU), which is typically installed in a vehicle. The recorded vertical dynamics are preferably used for modeling the measurement noise in steps a) and b). Preferably, the same sensors or sensor unit are used for capturing the vertical dynamics in both steps a) and b).

[0020] According to a further embodiment, for the execution of step b), mutually correlated parameters of the recorded vertical dynamics are removed, leaving mutually uncorrelated parameters, and the time-varying part of the measurement noise is determined taking into account the uncorrelated parameters of the recorded dynamics. The advantage here can be seen in a reduction of the overall complexity of the method.

[0021] Furthermore, it can be advantageous if, according to one embodiment in step b), the time-varying part of the measurement noise is modeled taking into account at least one scaling factor, wherein the at least one scaling factor contains information about the vehicle. In particular, the vertical dynamics detected by the sensors or the sensor unit are modeled with the at least one scaling factor. The at least one scaling factor preferably serves as a so-called tuning factor to incorporate the various pieces of information about the vehicle into the modeling of the time-varying part of the measurement noise. A non-restrictive example of information about the vehicle could be the type or type of vehicle. For example, a passenger car behaves differently with regard to vertical dynamics when driving over a bump than a truck when driving over the same bump.The scaling factors thus improve the determination of navigation data and, consequently, the overall position determination. The scaling factor can, for example, include information regarding the spring constants of the vehicle, which is considered as a spring-mass system.

[0022] The time-varying component of the measurement noise is preferably determined according to an alternative or supplementary embodiment, taking map data into account. This improves the accuracy of the position determination.

[0023] According to a further embodiment, parameters are read from the map data along with the vehicle's position, which are then taken into account when determining the varying part of the measurement noise. This improves the determination of the navigation data in a complementary or alternative way, similar to the embodiment already described above.

[0024] In order to consider only the dynamic part of the vertical acceleration, according to an advantageous embodiment a proportion of gravity is detected and the vertical dynamics in steps a) and b) are compensated by the proportion of the detected gravity.

[0025] According to one embodiment, steps a) and b) only consider vertical dynamics directed orthogonally to the vehicle's direction of travel. Thus, for example, when the vehicle is traveling uphill or downhill, absolute vertical movement is deliberately not included in the modeling of the measurement noise. In other words, only the relative movement of the vehicle orthogonal to the road surface is used for the modeling in steps a) and b).

[0026] In another embodiment, a Kalman filter is used to perform step d). Kalman filters have proven particularly suitable for determining navigation data using GNSS systems.

[0027] Furthermore, a computer program product, comprising instructions that, when executed by a computer, cause the computer to execute the described procedure, is to be described.

[0028] The computer program product is preferably installed on the described control unit.

[0029] The following will be described in more detail: a computer-readable storage medium comprising instructions which, when executed by a computer, cause it to execute the described procedure or the steps of the described procedure.

[0030] The invention and its technical context are explained in more detail below with reference to the figures. The figures show preferred embodiments, to which the invention is not limited. It should be noted in particular that the figures, and especially the size relationships shown in the figures, are only schematic. They show: Fig. 1: a schematically represented block diagram of the method according to the invention; and Fig. 2: a diagram showing a comparison of a high-precision reference velocity with, on the one hand, an associated variance according to the prior art and, on the other hand, an associated variance according to the method according to the invention; Fig. 3: a diagram showing a comparison between measurement noise based on an error of the vertical velocity according to the prior art and according to the method according to the invention.

[0031] In the figures, identical or equivalent components are always marked with the same reference symbols.

[0032] In Fig. Figure 1 schematically illustrates the process according to the invention by means of a block diagram.

[0033] In step a), a constant portion of the vertical velocity measurement noise is modeled, reflecting vehicle driving situations with vertical dynamics. This constant portion specifically represents a driving situation where the vehicle is on a flat and "smooth" road surface. In other words, the vehicle experiences low vertical dynamics or has a low vertical velocity, for example, because it is driving on a flat and level road surface without bumps.

[0034] In step b), a time-varying portion of the vertical velocity measurement noise is then modeled. This time-varying portion depends on the vehicle's state and reflects driving situations with a higher vertical dynamic compared to the vertical dynamics in step a). This portion specifically represents driving situations where the vehicle is on an uneven road surface with multiple bumps.

[0035] This step is based on the idea that a Kalman filter-typical model of measurement noise using only constant normally distributed parameters, such as the constant part of the vertical dynamics, is prone to errors because it does not cover both dynamic and non-dynamic situations. Depending on the dynamics, this leads to increased modeling errors and consequently to errors in the navigation solution. Step b) and the modeling of the time-varying part of the measurement noise allow the dynamic component to be included in the calculation of the overall measurement noise, thus leading to a more accurate determination of navigation data.

[0036] Thus, in step c) a modeled total measurement noise is determined as a combination of the constant part of the measurement noise determined in step a) and the time-varying part of the measurement noise determined in step b).

[0037] In step d) of the inventive method, navigation data are subsequently determined taking into account the modeled total measurement noise, which are therefore more accurate than navigation data determined according to the prior art and without considering the time-varying part.

[0038] The procedure can be used, for example, and in relation to Fig. 1 in a GNSS system 102 in a vehicle 104. The vertical dynamics, in particular the vertical velocity of the vehicle 104, are recorded, for example, by means of sensors 106 or a sensor unit and made available to the method. The GNSS system 102, the vehicle 104, and the sensors 106 are shown only as examples and schematically in the form of rectangles in the Fig. 1 shown.

[0039] Out of Fig. Figure 2 shows a diagram that compares a highly accurate vertical reference velocity v ref with associated variance V SdT according to the state of the art and a variance V neu according to the method of the invention.

[0040] As shown in the diagram Fig. As can be seen in section 2, the variance V SdT according to the state of the art, regardless of the value of the vertical reference velocity v refa constant value, which is inaccurate for determining the navigation data and makes it erroneous. The displayed variance V neu According to the inventive method, in which the time-varying part of the measurement noise is also included, depending on the value of the vertical reference velocity v ref a different value, which allows for a more accurate determination of navigation data and thus improved positioning of the vehicle.

[0041] In Fig. Figure 3 shows a diagram that compares a measurement noise R SdT based on a vertical velocity error according to the state of the art and a measurement noise R neu according to the method of the invention.

[0042] As can be seen from the graph, the measurement noise R SdTAccording to the state of the art, i.e., without including the time-varying part in the determination of the total measurement noise, a significantly larger amplitude is observed than the measurement noise R. neu according to the method according to the invention.

[0043] This amplitude leads to errors when determining navigation data according to the state of the art. The amplitude of the measurement noise R neu However, according to the inventive method, in which the time-varying part of the measurement noise of the vertical speed is included, it is significantly lower and thus shows that an additional use of the time-varying part of the measurement noise of the vertical speed according to the inventive method leads to an improved determination of the navigation data.

Claims

[1] Method for determining navigation data in a vehicle (104) with a GNSS system (102), wherein the method models a measurement noise of a vertical velocity of the vehicle (104) and the method comprises at least the following steps: a) Modeling a constant part of the measurement noise of the vertical velocity, which reflects a driving situation of the vehicle (104) with a vertical dynamics; b) Modeling a time-varying part of the measurement noise of the vertical velocity, wherein the time-varying part is dependent on a vehicle state and reflects driving situations of the vehicle (104) with a higher vertical dynamics compared to the vertical dynamics in step a); c) Determining a modeled total measurement noise (R) neu) as a combination of the constant part of the measurement noise determined in step a) and the time-varying part of the measurement noise determined in step b); and d) Determining the navigation data taking into account the modeled total measurement noise (R) neu ). [2] Method according to claim 1, wherein the vertical dynamics of the vehicle (104) is detected for the modeling of the measurement noise in steps a) and b) using at least one sensor (106), in particular of the vehicle (104), wherein the at least one sensor (106) detects the parameters reflecting the vertical dynamics. [3] Method according to claim 2, wherein for the modeling of the time-varying part of the measurement noise in step b) mutually correlated parameters of the detected vertical dynamics are removed, so that mutually uncorrelated parameters remain and the time-varying part of the measurement noise is determined taking into account the remaining uncorrelated parameters of the detected dynamics. [4] Method according to claim 3, wherein in step b) the time-varying part of the measurement smoke is modeled taking into account at least one scaling factor, wherein the at least one scaling factor contains information about the vehicle (104). [5] Method according to one of the preceding claims, wherein the time-varying part of the measurement noise is determined taking into account map data. [6] Method according to claim 5, wherein parameters are read from the map data using a position of the vehicle (104) which are taken into account when determining the varying part of the measurement noise. [7] Method according to one of the preceding claims, wherein a portion of the gravity is detected and the vertical dynamics in steps a) and b) are compensated by the portion of the detected gravity. [8] Method according to one of the preceding claims, wherein in steps a) and b) only a vertical dynamic directed orthogonally to a direction of travel of the vehicle (104) is used. [9] A method according to any of the preceding claims, wherein a Kalman filter is used for the performance of step d), wherein the total measurement noise (R) modeled in step c) neu ) is taken into account in the Kalman filter. [10] Control unit comprising a processor adapted / configured to perform the method according to any one of claims 1 to 9. [11] Computer program product comprising instructions which, when the computer program product is executed by a computer, cause it to execute the method according to any one of claims 1 to 9. [12] Computer-readable storage medium comprising instructions which, when executed by a computer, cause it to perform the steps of the method according to any one of claims 1 to 9.

Images