Method for Providing Information on the Reliability of Parameter Estimation of Parameters for the Operation of a Vehicle - Patent application

Abstract

The present invention relates to a method for providing information (1, 2) on the reliability of parameter estimation of parameters for the operation of a vehicle (3), the method comprising at least the steps of: a) calculating a completeness range (1) for the parameter estimation, describing a range within which the estimated parameter lies with minimum probability; b) calculating a validity index (2) describing the validity of the completeness range (1) calculated in step a); and c) providing the completeness range (1) calculated in step a) and the validity index (2) calculated in step b).

Description

[Technical Field]

[0001] The present invention relates to a method for providing information on the reliability of parameter estimation of parameters for the operation of a vehicle, a computer program for implementing the method, a machine-readable storage medium, and a control device for a vehicle, the invention being particularly suitable for use in connection with autonomous driving. [Background technology]

[0002] Prior art One of the most important tasks in autonomous driving is to determine the ego-position of an autonomous vehicle as accurately and reliably as possible. An autonomous vehicle is typically equipped with multiple sensors, such as inertial sensors, wheel sensors, environmental sensors, GNSS sensors, optical sensors, and / or acoustic sensors, by means of which the ego-position can be estimated. In this context, it is useful if information regarding the (expected) estimation accuracy of the calculated ego-position is also output. In this context, for example, the reliability of the calculated ego-position can be indicated by a so-called "protection level" (abbreviated as "PL"). In this case, the PL can describe a statistical error limit, the calculation of which is typically based on statistical considerations and, if necessary, additionally, on an appropriate adjustment of the estimation algorithm. Summary of the Invention [Problem to be solved by the invention]

[0003] The concept of providing a protection level is widespread, especially in the aviation industry. However, the solutions developed there are not easily transferable to the field of autonomous driving. In particular, gaps between buildings and their influence on satellite signals are issues that do not arise in aviation applications. Therefore, an improved method for calculating a protection level that is as reliable as possible is desirable, which is capable of delivering reliable results, especially in difficult environments, such as urban areas. [Means for solving the problem]

[0004] Disclosure of the Invention The present specification provides a method for providing information on the reliability of parameter estimation of parameters for the operation of a vehicle (motor vehicle) according to claim 1, comprising at least a) calculating completeness bounds for the parameter estimates, describing the ranges within which the estimated parameters lie with minimum probability; b) calculating a validity index describing the validity of the completeness coverage calculated in step a); c) providing the completeness extent calculated in step a) and the validity indicator calculated in step b); A method including:

[0005] Steps a) through c) can be performed at least one or more times sequentially in the order described to implement the method. Furthermore, steps a) and b) can be performed at least partially in parallel or can be performed simultaneously.

[0006] The parameters may be safety-related parameters and / or parameters related to or necessary for at least partially automated and / or autonomous driving behavior. The parameters may provide information related to the driving behavior of the vehicle and / or may contribute to or be suitable for describing the driving behavior of the vehicle. Basically, one or more (various) parameters may be estimated. The parameters may be, for example, the ego position, the ego speed, the ego acceleration, etc. Preferably, the parameters relate to the ego position of the vehicle.

[0007] The parameters may be driving behavior parameters of the vehicle (automobile). Driving behavior parameters may be understood in this specification as parameters that contribute to describing the spatial driving behavior of the vehicle or the maneuvering of the vehicle in space. In particular, the driving behavior parameters contribute to describing at least the ego-motion and / or ego-position of the vehicle. The driving behavior parameters may be, for example, the (ego) position, the (ego) speed, the (ego) acceleration, or the attitude (or orientation) of the vehicle. Preferably, the driving behavior parameter is the ego-position of the vehicle. The method (therefore) may be used, for example, in connection with position estimation of the vehicle position. In this case, the completeness range may describe a range within which the estimated ego-position of the vehicle is (actually) present with minimum probability.

[0008] The information to be provided regarding the reliability of the parameter estimation may include, for example, at least two pieces of information. These pieces of information preferably include at least one integrity range for the parameter estimation and a validity indicator. Providing at least these two pieces of information advantageously contributes to improving the reliability of the determination of the vehicle's own position (especially an at least partially automated or autonomously driven vehicle). A further advantage of the method can be found in that, in addition to the integrity information in the form of an integrity range, validity information in the form of a validity indicator is provided. This provides a means of enhancing the information provided by the integrity range without having to change the integrity range itself. This advantageously allows the design of the integrity range determination for use in various contexts to remain unchanged or to be maintained independently from the validity information in the form of a validity indicator. Therefore, the integrity range can also be output as mathematically defined statistical error information or error limits. This offers advantageous usability, especially in terms of integrated interfaces, since the original meaning / definition of the integrity range does not need to be changed. This also allows for improved means for modular design. Therefore, the provided completeness range can also potentially be used independently of the validity index, which can advantageously broaden its applicability.

[0009] In step a), a completeness range for the parameter estimation is calculated, which describes the range in which the estimated parameter (value) lies (actually) with the least probability. The estimated parameter (value) here essentially describes the (individual, in particular current) estimation result of the parameter estimation. This means, in other words, that the completeness range describes the range in which the real or actual value of the estimated parameter lies with the least probability. Such a completeness range can also be referred to as a so-called "protection level".

[0010] The minimum probability is usually a predetermined minimum probability, preferably 90%, particularly preferably 95%, or may be 99%.

[0011] Preferably, the integrity range is a protection level. In this case, the protection level generally describes a (spatial, in particular two-dimensional or three-dimensional) range in which the estimated parameter (value) is (actually) present with the least probability. The estimated parameter (value) here basically describes the (individual, in particular current) estimation result of the parameter estimation. This means, in other words, that the protection level in particular describes a range in which the real or actual value of the estimated parameter is present with the least probability.

[0012] In other words, the protection level describes in particular the confidence interval or (spatial) confidence range in which the true value of the estimated parameter lies with the smallest probability, where the parameter estimate is usually in the center or centre of the confidence interval or range.

[0013] The minimum probability that the real or actual value of the estimated parameter actually lies within the protection level is much higher than in the "normal" completeness range. The minimum probability here is typically greater than 99.99%, and particularly preferably greater than 99.999% or even greater than 99.9999%. Furthermore, the minimum probability cannot be expressed as a percentage in the protection level, but can be expressed as an error that can occur within a given time interval. For example, the protection level can be defined such that the parameter in question deviates from the protection level at most once every 10 years. The protection level can be expressed, for example, as a unitless probability or rate, i.e., the probability of an error occurring over a time interval.

[0014] Parameter estimation may essentially involve one or more techniques for estimating the (same) parameters. Preferably, at least one technique for estimating parameters is used that can additionally provide and / or determine completeness information regarding the completeness of the estimation or reliability information regarding the reliability of the estimation result. For example, parameter estimation can be performed using a Kalman filter. In addition to the (original) estimation result, the Kalman filter usually also provides reliability information regarding the parameter estimate (e.g., a covariance matrix) that can be used to calculate the completeness range here.

[0015] In step b), a validity index is calculated, which describes the validity of the integrity range calculated in step a). The validity index may be binary (valid / not valid) or may include various "digital intermediate states". Possible "intermediate states" may, for example, indicate that a certain integrity risk number (e.g., 1e-3 for urban scenarios, 1e-5 for highways) is applicable at that time. Furthermore, the validity index may, for example, include a value (numerical value) describing the validity or reliability of the calculated integrity range. The calculation may, for example, include a combination of various (input) information, in particular various sensor information and / or status information (e.g., from vehicle safety-related systems and / or vehicle diagnostic devices). For example, the unavailability of at least one (input) information may have a (particularly negative) impact on the validity index.

[0016] The validity index can be calculated, for example, depending on vehicle-side conditions or information and / or vehicle-external conditions or information and / or information about the integrity range. For example, the vehicle-side conditions and / or vehicle-external conditions can contribute to a greater or lesser degree of reliability of the calculated integrity range. The vehicle-side conditions or information can be, for example, conditions or information about a predetermined driving situation, for example, a longitudinal acceleration and / or a lateral acceleration exceeding a defined limit, and / or conditions or information about a predetermined (safety-relevant) vehicle state and / or system state of a vehicle system, for example, an at least partial system failure and / or a sensor failure. The vehicle-external conditions or information can be, for example, conditions or information about weather conditions and / or environmental influences such as sensor shadows. The corresponding conditions or information can be associated, for example, with a value (numerical value) that can be introduced into the calculation of the validity index if the corresponding condition exists. If multiple conditions or information exist, these can be, for example, multiplied and / or added together or logically combined with each other to calculate the validity index.

[0017] In step c), the integrity range calculated in step a) and the validity index calculated in step b) are provided. In this case, the calculated integrity range and the calculated validity index can preferably be provided together. In this context, these two pieces of information can be provided, for example, in the form of a value pair. This information can be provided, for example, to a system of a vehicle configured for localization of the vehicle. Examples of such systems can be vehicle motion and position sensors and / or control devices for at least partially automated or autonomous driving of the vehicle. Such a system can, for example, determine whether, and possibly with what weight, corresponding parameter estimates should be used for position determination depending on the calculated integrity range and / or the calculated validity index.

[0018] According to an advantageous configuration, it is proposed that the determination of the integrity range according to step a) and / or the determination of the validity index according to step b) is performed on the basis of sensor data of at least one sensor of the vehicle. The at least one sensor may for example comprise a GNSS (Global Navigation Satellite System) sensor, which is typically able to receive and, optionally after (pre-)processing, provide GNSS data (e.g. GNSS correction data and / or GNSS position data). Alternatively or additionally, the at least one sensor may comprise at least one environmental sensor. The at least one environmental sensor may for example be a camera, a radar sensor, a LiDAR sensor and / or an ultrasonic sensor. Furthermore, other (or alternative) sensor data and / or input data may also be taken into account when calculating the completeness range according to step a) and / or when calculating the validity index according to step b), for example sensor data relating to vehicle parameters (e.g. speed, acceleration, steering angle, etc.) and / or input data describing the vehicle state, for example input data describing whether additional components, such as a trailer, bicycle carrier, etc., are connected to the vehicle.

[0019] According to a further advantageous configuration, it is proposed that the completeness range in step a) is calculated based on statistical reliability information for the parameter estimates. The statistical reliability information can be, for example, the variances and / or residuals of the parameter estimates. Furthermore, the reliability information can (optionally) be information calculated depending on the variances and / or residuals and / or (other) indicators of the reliability of the estimates. As reliability information, in particular residuals, variances, covariances, etc. are taken into account. For example, the reliability information can be formed in the form of or from a covariance matrix.

[0020] In particular in this context, the completeness range may be, for example, a confidence interval. A confidence interval (also called a credible range or a credible and prediction range) is a statistical interval that indicates the precision of a state estimate of a parameter (e.g., a mean value). The confidence interval indicates a range that (relative to a confidence level) would contain the true state of the parameter if a random experiment were repeated infinitely with some probability.

[0021] According to a further advantageous configuration, it is proposed to determine the effectiveness index using at least one environmental information relating to the environment around the vehicle, which may in particular be provided in the form of data from environmental sensors of the vehicle, which may be, for example, a camera, a radar sensor, a LiDAR sensor and / or an ultrasonic sensor.

[0022] According to a further advantageous configuration, it is proposed that the effectiveness index is calculated using at least one state information relating to the state of the vehicle or the state of a vehicle component. The state information may, for example, comprise information on the occurrence of a predetermined driving situation, for example information on longitudinal and / or lateral acceleration exceeding a particularly defined limit. The state information may, for example, comprise information on the occurrence of a predetermined (safety-relevant) vehicle state and / or information on the occurrence of a system state of a vehicle system, for example information on the occurrence of an at least partial system and / or sensor failure.

[0023] According to a further advantageous configuration, it is proposed to calculate the validity index using the completeness range calculated in step a), so that the completeness range can have an influence on the validity index, for example.

[0024] According to a further advantageous configuration, it is proposed to provide the validity indicator as a value from a predetermined value set. Thus, for example, predetermined validity stages can be set and it can be calculated which of these validity stages is present at the time or in the corresponding completeness range. The value set can be, for example, 2, and therefore the validity indicator is binary.

[0025] According to a further aspect, a computer program for carrying out the methods described herein is also proposed, in other words a computer program product comprising instructions for causing a computer to carry out the methods described herein when the program is executed by said computer.

[0026] According to a further aspect, a machine-readable storage medium having stored thereon a computer program as described herein is also proposed. Typically, a machine-readable storage medium is a computer-readable data carrier.

[0027] According to a further aspect, a control device for a vehicle (motor vehicle) is also proposed, which is configured to perform the method described herein. The control device may for example be a device (or computer) for localization of the vehicle.

[0028] The control device may include, for example, an integrity module for calculating the integrity range and a validity module for calculating the validity index. The modules can essentially be realized as physical components of the control device or as parts of a computer program. The integrity module here may be configured, for example, to calculate the integrity range depending on at least one reliability information. The reliability information may be obtained, for example, from a Kalman filter estimation. The validity module may be configured, for example, to calculate the validity index depending, in particular, on vehicle-side conditions or information and / or vehicle-external conditions or information and / or the integrity range. For this purpose, the validity module may include a combiner that can combine various information, such as vehicle-side conditions or information and / or vehicle-external conditions or information and / or information on the integrity range, to the validity index or (pre-)process and subsequently combine this information to the validity index. The various information may include, for example, sensor data, at least one environmental information, and / or at least one state information, and / or information on at least one integrity range.

[0029] Furthermore, a vehicle equipped with the control device described herein can also be provided, which is essentially a passenger vehicle, preferably an at least partially automated and / or autonomously operating passenger vehicle, in particular an autonomous passenger vehicle.

[0030] The details, features and advantageous configurations discussed in relation to the method may also occur correspondingly in the computer program, storage medium, control device and / or vehicle presented herein, and vice versa, and in this respect the embodiments of the present specification are to be referred to in their entirety in order to characterize each feature in more detail.

[0031] The solutions presented herein and their technical environments will be described in more detail below with reference to the drawings. It should be noted that the present invention is not limited to the illustrated embodiments. In particular, unless otherwise explicitly indicated, partial aspects of the factual content illustrated in the drawings may be extracted and combined with other drawings and / or other components and / or knowledge in this specification. The drawings schematically show: [Brief explanation of the drawings]

[0032] [Figure 1] FIG. 2 illustrates an exemplary sequence of the methods described herein. [Figure 2] FIG. 2 illustrates an exemplary structure of a control device as described herein. [Figure 3] 1 is an exemplary diagram illustrating a vehicle equipped with a control device as described herein. DETAILED DESCRIPTION OF THE INVENTION

[0033] 1 shows a schematic representation of an exemplary sequence of the method described herein. The method is used to provide information 1, 2 about the reliability of parameter estimation of parameters for the operation of a vehicle 3. The order of method steps a), b) and c) indicated by blocks 110, 120 and 130 is usually set in a regular operating sequence. In particular, steps a) and b) can be performed at least partly in parallel and can also be performed simultaneously.

[0034] In block 110, according to step a), a completeness range 1 for the parameter estimation is calculated, where the completeness range 1 describes the range in which the estimated parameter lies with minimum probability. In block 120, according to step b), a validity index 2 is calculated that describes the validity of the completeness range 1 calculated in step a). In block 130, according to step c), the completeness range 1 calculated in step a) and the validity index 2 calculated in step b) are provided.

[0035] 2 shows a schematic diagram of an exemplary structure of a control device 8 described herein. The control device 8 is configured to perform the methods described herein. The control device 8 includes, for example, an integrity module 9 that calculates an integrity range 1 (here, a protection level) and a validity module 10 that calculates a validity index 2. Each module here can essentially be realized as a physical component of the control device 8 or as part of a computer program.

[0036] The integrity module 9 is exemplarily configured to calculate the integrity range 1 depending on at least one reliability information 5. The reliability information 5 can be obtained, for example, from a Kalman filter estimation. In this context, the reliability information 5 can be, for example, in the form of or formed from a covariance matrix. The Kalman filter estimation can be based on sensor data of vehicle sensors. Thus, Fig. 2 shows that in step a) the integrity range 1 can be calculated based on statistical reliability information 5 via parameter estimation, and possibly also shows an example of how this can be calculated.

[0037] The validity module 10 is exemplarily configured to calculate the validity index depending in particular on vehicle-side conditions or information and / or vehicle-external conditions or information and / or information on the integrity range 1. For this purpose, the validity module 10 comprises, for example, a combiner 11, by means of which various information, such as vehicle-side conditions or information and / or vehicle-external conditions or information and / or information on the integrity range 1, can be combined with the validity index 2 or (pre-)processed information can be combined with the validity index 2 subsequent to processing. Such various information can, for example, comprise sensor data 4, at least one environmental information 6, and / or at least one state information 7. Of course, further information can also be taken into account, which can be provided by sensors 12 (see FIG. 3 ) and / or the integrity module 9.

[0038] 2 thus shows that the effectiveness index 2 can be calculated using at least one environmental information 6 relating to the environment around the vehicle 3, and possibly also shows examples of how this can be calculated. Figure 2 therefore further shows that the effectiveness index 2 can be calculated using at least one state information 7 relating to the state of the vehicle 3 or the state of a component of the vehicle 3, and possibly also shows examples of how this can be calculated. Figure 2 therefore further shows that the effectiveness index 2 can be calculated using the completeness range 1 calculated in step a), and possibly also shows examples of how this can be calculated.

[0039] 3 shows a schematic illustration of an exemplary vehicle 3 equipped with a control device 8 as described herein. The vehicle 3 further comprises a number of sensors 12 that can provide data or information to the control device 8, for example. The information can be, for example, environmental information 6 relating to the environment around the vehicle 3 and / or status information 7 relating to the state of the vehicle 3. Thus, FIG. 3 also shows an example of how the integrity range 1 can be calculated according to step a) and possibly how the validity index 2 can be calculated according to step b) based on sensor data 4 of at least one sensor 12 of the vehicle 3. In particular, the validity index 2 can be provided as a value from a predetermined value set.

Claims

1. 1. A computer-implemented method for providing information (1, 2) about the reliability of parameter estimation of parameters for the operation of a vehicle (3), comprising at least a) calculating a completeness bound (1) for the parameter estimation, describing the bounds within which the estimated parameters lie with minimum probability; b) calculating a validity index (2) describing the validity of the completeness scope (1) calculated in step a); c) providing the completeness extent (1) calculated in step a) and the validity index (2) calculated in step b); The method includes:

2. 2. The method according to claim 1, wherein the step of calculating the completeness range (1) according to step a) and / or the step of calculating the validity index (2) according to step b) is performed based on sensor data (4) of at least one sensor (12) of the vehicle (3).

3. 3. The method according to claim 1 or 2, wherein the completeness bound (1) in step a) is calculated based on statistical confidence information (5) on the parameter estimates.

4. 4. The method according to claim 1, wherein the effectiveness index (2) is calculated using at least one environmental information (6) relating to the environment surrounding the vehicle (3).

5. 5. The method according to claim 1, wherein the validity index (2) is calculated using at least one state information (7) relating to a state of the vehicle (3) or a state of a component of the vehicle (3).

6. The method according to any one of claims 1 to 5, wherein the validity index (2) is calculated using the completeness range (1) calculated in step a).

7. 7. The method according to any one of the preceding claims, wherein the validity indicator (2) is provided as a value from a predefined set of values.

8. A computer program for carrying out the method according to any one of claims 1 to 7.

9. A machine-readable storage medium storing the computer program according to claim 8.

10. A control device (8) for a vehicle (3), configured to implement a method according to any one of claims 1 to 7.

Images