Method for determining an integrity region of a parameter estimation
Through the environmental characteristic positioning method based on sensor data, the integrity area of vehicle position parameter estimation is determined, which solves the problem of inaccurate positioning of autonomous vehicles when GNSS signals are poor in urban environments, and the accurate and reliable estimation of vehicle position is achieved.
Patent Information
- Application Number
- CN202380050798.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-07-12
- Filing Date
- 2023-06-23
- Publication Date
- 2025-02-14
AI Technical Summary
In the field of autonomous driving, it is difficult for the prior art to accurately and reliably determine the vehicle's own position in urban environments, especially under conditions of poor GNSS signal.
The integrity area of vehicle positioning parameter estimates is determined by an environmental feature positioning method based on sensor data. The method includes determining the basic integrity information and at least one additional integrity information, combining the information to determine the integrity area, thereby providing reliable information about positioning accuracy.
This method can provide accurate and reliable estimates of vehicle location in environments with poor GNSS signals, ensure the credibility of positioning accuracy, and is suitable for at least partially automated or autonomous vehicles.
Smart Images
Figure CN119452272A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for determining an integrity region for parameter estimation for positioning a vehicle based on sensor data obtained for positioning according to environmental features. In addition, it relates to a computer program for carrying out the method, a machine-readable storage medium storing the computer program, and a control device for a motor vehicle, wherein the control device is designed to carry out the method. In particular, the method can be used in the field of at least partially automated or autonomous driving. Background Art
[0002] One of the most important challenges of autonomous driving is to determine the own position of the autonomous vehicle as accurately and reliably as possible. Autonomous vehicles usually have sensors, such as inertial sensors, wheel sensors, environmental sensors, GNSS sensors, optical and / or acoustic sensors, with the help of which the vehicle can estimate its own position.
[0003] It is known from GNSS positioning (GNSS sensors as primary data source) that information about the (expected) estimated accuracy of the determined own position can also be output. In this context, for example, the confidence level of the determined own position can be expressed by a so-called "protection level" ("PL" for short). In this context, the PL can describe a statistical error limit, the calculation of which is usually based on statistical considerations and, if necessary, additionally on appropriate coordination of the estimation algorithm.
[0004] The concept of providing protection levels is widespread, especially in the aviation sector. However, solutions developed in the aviation sector cannot be easily transferred to autonomous driving applications. In particular, problems such as urban canyons and their influence on satellite signals are issues that do not arise in aviation applications.
[0005] Furthermore, improvements to at least partially automated or autonomous driving aim to replace GNSS signals as the primary data source for positioning by sensor data from environmental sensors. In particular, primary GNSS positioning should be replaced by primary positioning based on known environmental features (so-called feature positioning), in particular for positioning tasks in specific areas (e.g. urban areas).
[0006] It is desirable for such feature-based localization methods to provide as reliable information as possible about their (expected) estimated accuracy. Summary of the invention
[0007] According to claim 1, a method is proposed for determining an integrity region for parameter estimation for localizing a vehicle based on acquired sensor data, the sensor data being used for localization based on environmental features, wherein the integrity region describes a region in which the estimated parameters are located with minimal probability, wherein the method comprises at least the following steps:
[0008] a) determine the basic integrity information,
[0009] b) determining at least one additional integrity information,
[0010] c) determining an integrity area using the base integrity information and the at least one first additional integrity information.
[0011] To perform the method, steps a), b) and c) can, for example, be performed at least once and / or repeatedly in the order described. In addition, steps a), b) and c), in particular steps a) and b), can be performed at least partially in parallel or simultaneously. The method can, for example, be performed by a control device as described herein. The vehicle can, for example, be a motor vehicle, such as a car. The vehicle is preferably designed for at least partially automatic or autonomous driving operation. The method advantageously helps to provide a protection level calculation for feature location.
[0012] An integrity region describes a region in which an estimated parameter (value) lies with the minimum probability (actual). Here, an estimated parameter (value) essentially describes a (single, in particular instantaneous) estimation result of the parameter estimation. In other words, this means in particular that an integrity region describes a region in which the actual value or true value of the estimated parameter lies with the minimum probability. Such an integrity region can also be referred to as a so-called "protection level".
[0013] The minimum probability is usually a predefined minimum probability. The minimum probability is preferably 90%, particularly preferably 95% or even 99%.
[0014] Preferably, the integrity region is a protection level. A protection level here generally describes a (spatial, in particular two-dimensional or three-dimensional) region in which an estimated parameter (value) is (actually) located with minimal probability. In this case, an estimated parameter (value) essentially describes a (single, in particular instantaneous) estimation result of a parameter estimation. In other words, this means, in particular, that the protection level describes a region in which the actual value or true value of the estimated parameter is located with minimal probability.
[0015] 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 minimal probability. In this case, the estimated value of the parameter is usually located in the middle or center of the confidence interval or confidence range.
[0016] The minimum probability that the true or actual value of the estimated parameter actually lies within the protection level is much higher than in the "usual" integrity region. In this case, the minimum probability is usually more than 99.99%, particularly preferably more than 99.999% or even more than 99.9999%. In the case of a protection level, the minimum probability cannot be expressed as a percentage, but rather as a possible error within a certain time interval. For example, the protection level can be defined so that the parameter under consideration is outside the protection level at most once every 10 years. For example, the protection level can be expressed as a unitless probability or ratio, i.e. the probability of an error within a time interval.
[0017] The method is preferably used to determine an integrity region of a parameter estimate of a driving operation parameter of a motor vehicle. The driving operation parameter is typically a safety-critical or safety-relevant parameter of the driving operation of the motor vehicle. The driving operation parameter is preferably a (safety-critical or safety-relevant) parameter of a driving operation of a motor vehicle that is operated (or operated) at least partially automatically or even autonomously.
[0018] Driving operation parameters are here in particular understood to be parameters which contribute to describing the driving operation of the motor vehicle in space or the operation of the motor vehicle in space. In particular, the driving operation parameters contribute at least to describing the own movement and / or the own position of the motor vehicle. The driving operation parameters can be, for example, the (own) position, (own) speed, (own) acceleration or position (or orientation) of the motor vehicle. The driving operation parameter is preferably the own position of the motor vehicle.
[0019] The method is preferably used to determine an integrity region that describes the integrity of the estimate of the vehicle's own position. In other words, this means in particular that the parameter is preferably the vehicle's own position. The method can (therefore) be used, for example, to determine an integrity region for the position estimate of the vehicle's position. In this case, the integrity region can describe a region in which the vehicle's estimated own position is (actually) located with minimal probability. As an alternative or in addition to estimating the vehicle's own position, the method can also be used to estimate the vehicle's own speed, orientation, own movement, etc.
[0020] Parameter estimation may in principle comprise one or more methods for estimating the (same) parameter. For example, parameter estimation may comprise at least two different methods, such as a first method and a second method for estimating the parameter that is different from the first method. Preferably a method for estimating the parameter is used that may also provide and / or determine integrity information about the integrity of the estimate.
[0021] This method may preferably be used when the vehicle is in a particular area where, for example, GNSS reception is expected to be relatively poor. "GNSS" stands for Global Navigation Satellite System. For example, this method may be performed when the vehicle is located in an urban area.
[0022] According to an advantageous design proposal, the positioning according to environmental features is performed as relative positioning to known features in the vehicle's surroundings. For example, known features can be obtained from a digital map of the vehicle's surroundings. In particular, in conjunction with the method described herein, feature-based positioning (feature positioning) is used as the main positioning method. This "main" positioning method can in principle be supplemented by auxiliary positioning methods, such as inertial navigation or (simplified) GNSS positioning. For example, the advantage of this application can be that the vehicle can be equipped with a relatively inexpensive GNSS receiver.
[0023] According to another advantageous design, the sensor data is obtained from a camera sensor, a video sensor, a radar sensor or a lidar sensor. These sensors are particularly advantageous for performing feature-based or characteristic-based vehicle positioning. One or more sensors can be arranged in or on the vehicle.
[0024] According to another advantageous design, at least one result of the parameter estimation and the basic integrity information are determined by a filter. For example, a Kalman filter or a particle filter can be used as a filter. Preferably, a Kalman filter can be used. The Kalman filter can read sensor data suitable for feature-based positioning. In addition to the estimated positioning result, the Kalman filter usually also outputs at least one random information (for example in the form of a covariance matrix), which mathematically describes the reliability of the estimation result.
[0025] According to another advantageous embodiment, it is provided that the basic integrity information is determined based on a mathematical model and / or describes a random measure. In particular, the basic integrity information describes a variance of a (characteristic or characteristic) positioning and / or an information quality estimate.
[0026] For example, the basic integrity information may be determined based on at least one variance and / or with the aid of data from a covariance matrix.The corresponding mathematical or stochastic information may generally be provided, for example, by a Kalman filter on the respective position result.
[0027] According to a further advantageous embodiment, it is provided that at least one additional integrity information item describes information about the number and / or distribution (and / or density) of detectable features. Preferably, at least the characteristic or the number of features is included as additional integrity information item in the calculation of the integrity area. For example, a Fisher distribution (with the number of features as a parameter) can be formed or used, in particular in order to scale the basic integrity information item (e.g. variance / information quality estimate).
[0028] In particular, the at least one additional integrity information may (also) describe at least one information about the distribution of the detectable feature.An example of a detectable feature distribution may also be, for example, periodically recurring structures, such as pillars on a highway.
[0029] Advantageously, the integrity area may be determined using a random measure (as basic integrity information) and information about the characteristic distribution (as additional integrity information).
[0030] One or more of the following information may be used as advantageous (further) additional integrity information:
[0031] Vehicle system status (e.g. wheel speed, running duration);
[0032] The number of plausible hypotheses for matching between feature measurements and feature maps;
[0033] similarity values of feature measures and feature maps, in particular in a way that is evaluated after matching (e.g. Hausdorff metric);
[0034] Periodic structures in the map (e.g. guardrail posts spaced 1.5m apart increase the probability of selecting the wrong optimum in feature matching);
[0035] Additional monitoring outputs (e.g. strong fluctuations in similar values);
[0036] State variables (e.g. time, initialization phase, situation estimation, etc.);
[0037] Movement information (e.g., position, orientation, speed).
[0038] According to another advantageous embodiment, it is provided that the integrity region is determined as a protection level. The protection level generally describes the region (in space, in particular in two or three dimensions) in which the estimated parameter (value) with the minimum probability (actual) lies. Before providing the protection level, a coordinate transformation can be performed, in particular in order to output the protection level in a desired form.
[0039] The protection level can advantageously provide information about the error of the position estimate for feature localization. Furthermore, the protection level can have a specification of how often the position estimate is less than a target value. Here, it can also be defined how likely it is that the position error is greater than a defined target value without the protection level indicating this.
[0040] Optionally, the method may be combined with existing GNSS positioning algorithms. For example, these may help determine the integrity area according to step c).
[0041] According to another aspect, a computer program for executing the method proposed here is proposed. In other words, this particularly relates to a computer program (product) comprising instructions which, when the program is executed by a computer, cause the computer to execute the method described here.
[0042] According to another aspect, a machine-readable storage medium is provided, on which the computer program provided herein is saved or stored. The machine-readable storage medium is typically a computer-readable data carrier.
[0043] According to another aspect, a control device for a motor vehicle is provided, wherein the control device is designed to perform the method described herein. The control device may, for example, include a computer or a controller capable of executing instructions to perform the method. To this end, the computer or the controller may, for example, execute the described computer program. For example, the computer or the controller may access the described storage medium in order to be able to execute the computer program.
[0044] The control device can be, for example, a component of a movement and position sensor, which is or can be arranged in particular in or on a (motor) vehicle, or can be connected to such a sensor for exchanging information. In this context, it can be provided that the GNSS sensor and / or the positioning device (including the filter explained above) and / or the control device are components of the movement and position sensor. Furthermore, it can be provided (alternatively) that the control device includes the movement and position sensor, which in this case can include, for example, a GNSS sensor and / or a positioning device.
[0045] Details, features and advantageous embodiments of the methods explained in this article may also be present in the computer program and / or storage medium and / or control device proposed here, and vice versa. In this regard, full reference is made to the embodiments for more detailed characterization of the features. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] The solution proposed here and its technical environment are explained in more detail below with reference to the accompanying drawings. It should be noted that the present invention is not limited to the embodiments shown. In particular, unless otherwise explicitly stated, some aspects of the facts explained in the accompanying drawings can also be extracted and combined with other drawings and / or other components and / or findings of this specification. The accompanying drawings schematically show:
[0047] Figure 1 An exemplary process of the method proposed here is shown.
[0048] Figure 2 One example of the vehicle control device described herein is shown. DETAILED DESCRIPTION
[0049] Figure 1An exemplary flow of the method proposed here is schematically shown, the method for determining an integrity region for parameter estimation for positioning a vehicle based on acquired sensor data, the sensor data being used for positioning according to environmental features, wherein the integrity region describes the region in which the estimated region is located with the minimum probability. The order of steps a), b) and c) shown in blocks 110, 120 and 130 is exemplary and can be traversed at least once in the order shown, for example, to perform the method.
[0050] In block 110, base integrity information is determined according to step a). In block 120, at least one additional integrity information is determined according to step b). In block 130, an integrity zone is determined using the base integrity information and the at least one first additional integrity information according to step c).
[0051] Figure 2 An example of a control device 1 described here for a motor vehicle 2 is shown schematically, wherein the control device 1 is designed to carry out the method described here.
[0052] The positioning may be performed as a relative positioning to known features in the vehicle's surroundings based on the environmental features. For example, the known features may be obtained from a digital map of the area surrounding the vehicle. For example, a rough positioning of the vehicle relative to the digital map may be performed during initialization and / or using additional sensors such as GNSS sensing mechanisms and / or inertial sensors (e.g., inertial measurement units, referred to as IMUs).
[0053] Here, the sensor data can preferably be obtained from at least one environmental sensor. For example, the sensor data can be obtained from a camera sensor, a video sensor, a radar sensor or a lidar sensor.
[0054] At least one result of the parameter estimation and the basic integrity information can be determined by a filter. For example, the data can be determined by means of a Kalman filter or a particle filter. The filter can be a component of a positioning device of the vehicle, which can transmit the data to the control device.
[0055] The basic integrity information may be determined based on a mathematical model and / or describe a random measure. For example, the basic integrity information may be determined based on at least one variance and / or with the aid of data of a covariance matrix. The corresponding mathematical or random information may usually be provided, for example, by a Kalman filter for the corresponding position result.
[0056] The at least one additional integrity information may describe information about the number and / or distribution of the detectable features. In particular, the at least one additional integrity information may describe at least one information about the distribution of the detectable features. Advantageously, the integrity area may be determined using the random measure (as base integrity information) and the information about the distribution of the features (as additional integrity information).
[0057] The integrity area can be determined as a protection level. This represents a particularly advantageous option for enabling the use of the method for modern systems for at least partially automated and / or autonomous driving.
Claims
1. A method for determining an integrity region for parameter estimation for positioning a vehicle based on acquired sensor data, the sensor data being used for positioning according to environmental features, wherein the integrity region describes a region in which the estimated parameters are located with minimum probability, wherein the method comprises at least the following steps: a) determine the basic integrity information, b) determining at least one additional integrity information, c) determining the integrity area using the basic integrity information and the at least one first additional integrity information. 2 . The method according to claim 1 , wherein the positioning according to environmental features is performed as a relative positioning with respect to known features in the surrounding environment of the vehicle.
3. The method according to claim 1 or 2, wherein the sensor data is obtained from a camera sensor, a video sensor, a radar sensor or a lidar sensor.
4. The method according to any of the preceding claims, wherein at least one result of the parameter estimation and the underlying integrity information are determined by a filter.
5. The method according to any of the preceding claims, wherein the basic integrity information is determined based on a mathematical model and / or describes a random measure.
6. The method according to any of the preceding claims, wherein the at least one additional integrity information describes information about the number and / or distribution of detectable features.
7. The method according to any of the preceding claims, wherein the integrity area is determined as a protection level.
8. A computer program for performing the method according to any one of claims 1 to 7.
9. A machine-readable storage medium having stored thereon the computer program according to claim 8.
10. A control device (1) for a motor vehicle (2), wherein the control device (1) is designed to carry out the method according to any one of claims 1 to 7.