Method and Apparatus for Providing a High-Resolution Digital Map

Abstract

A method for providing a high-resolution digital map includes locating a device and providing sensor data at a located position during a test drive of the located device. The method further includes ascertaining detection indicators for at least one object detected based on the provided sensor data at the located position, and adding at least one additional layer to the high-resolution digital map. The at least one additional layer includes the ascertained detection indicators for the at least one detected object.

Claims

1. A method for providing a high-resolution digital map, comprising: locating a device; providing sensor data at a located position during a test drive of the located device; ascertaining detection indicators for at least one object detected based on the provided sensor data at the located position; and adding at least one additional layer to the high-resolution digital map, the at least one additional layer including the ascertained detection indicators for the at least one detected object.

2. The method according to claim 1, wherein the at least one additional layer includes an additional layer for each sensor type configured to generate the sensor data.

3. The method according to claim 2, wherein the ascertained detection indicators are determined per the sensor type.

4. The method according to claim 1, wherein the ascertained detection indicators include at least one (i) a distance at which the at least one detected object was detected, and (ii) a probability with which the at least one detected object was detected.

5. The method according to claim 1, wherein: the ascertained detection indicators are related to defined weather conditions, or data relating to the defined weather conditions are stored in the high-resolution digital map.

6. The method according to claim 1, further comprising: ascertaining the detection indicators while performing the test drive or after performing the test drive.

7. The method according to claim 1, wherein the ascertained detection indicators are determined in the device or in the cloud.

8. The method according to claim 1, further comprising: creating locating data based on a simultaneous localization and mapping (“SLAM”) algorithm during the locating of the device using the provided sensor data.

9. The method according to claim 8, further comprising: ascertaining an estimated trajectory.

10. The method according to claim 1, wherein the sensor data is generated by at least one of the following types of sensors including radar, lidar, camera, and ultrasound.

11. An electronic system for creating a high-resolution digital map, comprising: at least one sensor device configured to provide sensor data during a test drive; a locating device configured to locate a device; a detection device configured to ascertain detection indicators of at least one object detected from the provided sensor data; and an addition device configured to add at least one additional layer to the high-resolution digital map, wherein the at least one additional layer includes the ascertained detection indicators for the at least one detected object.

12. The method according to claim 1, wherein a computer program product comprises a program code configured to cause an electronic device to perform the method when the computer program product is run on the electronic device.

13. The method according to claim 1, wherein the device is a vehicle.

Description

IN THE DRAWINGS

[0030] FIG. 1 shows a basic sequence of an embodiment of the proposed method; and

[0031] FIG. 2 is a block diagram of a proposed electronic system for carrying out the proposed method.

DESCRIPTION OF EMBODIMENTS

[0032] A core idea of the present invention lies in particular in an expansion of a high-resolution digital map (high-definition map, HD map) such that both qualitative and quantitative evaluation of sensor models is made possible using the high-resolution digital map. High-resolution digital maps play an essential role in the field of autonomous driving.

[0033] It is proposed that the invention should provide an improved high-resolution digital map 300 for an at least partially automated vehicle. In particular, the extension of such a high-resolution digital map 300 comprises a further additional layer 310a . . . 310n, which has what are known as sensor-specific confidence or detection indicators, referred to as key performance indicators (KPI).

[0034] These detection indicators represent performance indicators, on the basis of which the progress or the degree of fulfillment can be measured with respect to important objectives or critical success factors. They are defined from the requirements for the sensor system or also from further processing stages, such as an object detection. In order to use these performance indicators as a basis for a comparison with the generated sensor data from the simulation, it is necessary to determine tolerances within which the result of the sensor models is to be regarded as acceptable, since it is to be expected that the real sensor data (from real test drives) and the synthetic sensor data (from simulation drives using sensor models) differ to a certain degree. For this purpose, test drives are carried out using real sensors, and the threshold values relating to location quality are calculated on the basis of the resulting measurement data and marked in the high-resolution digital map 300.

[0035] Hereinafter, the term “automated vehicle” is used synonymously with the terms “fully automated vehicle,” “autonomous vehicle” and “partially autonomous vehicle.”

[0036] The steps for generating the at least one additional layer 310a . . . 310n are explained in more detail below. At the beginning, the current position of a device, in particular in the form of an automated vehicle, is ascertained. For this purpose, a locating device can preferably be used for providing GNSS or GPS data, which is used for locating a measurement vehicle (not shown). In the context of a test drive, the measurement vehicle is used to carry out a sensor-based detection of environment data by means of a sensor device, it being possible, for example, for a radar sensor and/or a lidar sensor and/or a camera and/or an ultrasound sensor to be used as sensor devices. In this way, a position of the device can be accurately determined, it being possible for any suitable locating method to be used for this purpose.

[0037] After the detection of the individual sensor data, the respective detection indicators can be calculated for at least one object detected from the sensor data. For this purpose, the current position of the vehicle is used to transform the calculated values into a high-resolution digital map 300. Calculating the detection indicators may be performed for one, a plurality or all of the sensor types used by the automated measurement vehicle. In this case, it can be provided for a separate additional layer 310a . . . 310n to be added to the high-resolution digital map 300 for each sensor type used.

[0038] In this case, for example a static object (e.g., traffic lights, traffic signs, buildings, etc.) can be ascertained using the detected sensor data. For example, the following may be specified as a detection indicator: the distance (e.g., 50 m) at which a static object is detected by the sensor device and/or the probability with which the static object has been detected by the sensor device, etc. Alternatively or additionally, the detected sensor data can be used to ascertain a predicted trajectory by means of an SLAM algorithm known per se. In this way, an additional layer 310a . . . 310n is created that provides sensor-specific properties as an extension of the high-resolution digital map 300.

[0039] It can optionally be provided to create the identified detection indicators for defined weather conditions (e.g., clear vision with sunshine, night, fog, rain, etc.) or to store data relating to the weather conditions as additional information within the additional layer 310a . . . 310n in the high-resolution digital map 300.

[0040] The high-resolution digital map 300 improved in this way can advantageously serve as a basis for the validation of sensor models in order to thereby verify that the developed sensor models are modeled sufficiently precisely and with correct properties. In this case, the observed scenarios are extracted from the real test drives in order to reproduce these in the simulation. The scenario can thus be simulated in the virtual surroundings using the developed sensor models. By calculating the detection indicators in the simulation, a comparison can be carried out between the detection indicators which were ascertained using real and using synthetically based sensor data.

[0041] As a result, a discrepancy of the performance between the real sensor system and the sensor models is not only visible, but also quantified. Furthermore, the method set out makes it possible not only to validate the sensor models but also to use validated position-dependent models in the simulation.

[0042] FIG. 1 is a highly schematic illustration of a basic sequence of a proposed method.

[0043] In a step 100, a device, in particular a vehicle, is located.

[0044] In a step 110, sensor data are provided at a located position during a test drive of the device.

[0045] In a step 120, detection indicators for at least one object detected by means of the provided sensor data are ascertained at the located position.

[0046] In a step 130, at least one additional layer 310a . . . 310n is added to the high-resolution digital map 300, the at least one further additional layer 310a . . . 310n including the detection indicators for the at least one detected object.

[0047] As a result, an extended high-resolution digital map 300 for a device, for example for an at least partially automated vehicle, is thereby provided, by means of which sensor models can be efficiently simulated or validated in a subsequent simulation process. Outlay for validating the sensor models can advantageously be minimized in this way.

[0048] In the following, the proposed sequence is examined in greater detail on the basis of an example. For this purpose, what is known as the “simultaneous localization and mapping” (SLAM) algorithm is suitable, as an example, in order to explain the described methodology.

[0049] The SLAM algorithm is characterized in that it can be used with all (e.g., three) sensor types (e.g., radar, lidar, camera). The objective of the SLAM algorithm is to simultaneously create a consistent high-resolution digital map 300 of the surroundings, and to estimate its own position within this high-resolution digital map 300. Consequently, a test drive results in both a high-resolution digital map of the surroundings and a predicted trajectory.

[0050] In order to evaluate the SLAM algorithm, for example the absolute error of the trajectory estimate relative to the GNSS data can be calculated as a confidence indicator. Since it is to be assumed that this deviation is different for each sensor type, the results of all sensor types are stored and added as a new layer 310a . . . 310n of the high-resolution digital map 300. This allows the current evaluation result of the system to be recorded in the high-resolution digital map 300.

[0051] This has the advantage that, for subsequent examinations, such as a resimulation of a test drive or the comparison after a system update, the calculated confidence indicators of the last system status are stored in a position-dependent manner, and a comparison can thus be carried out efficiently without recalculating the results.

[0052] FIG. 2 is a schematic block diagram of a proposed electronic system 200 for ascertaining a high-resolution digital map 300.

[0053] At least one sensor device 220a . . . 220n for providing sensor data which are recorded during a performed test drive is visible. A locating device 210 for locating the device is functionally connected to the sensor device 220a . . . 220n. A detection device 230 for ascertaining the detection indicators, described in more detail above, for at least one object detected by means of the provided sensor data is connected functionally to the locating device 20. An addition device 240 for adding at least one additional layer 310a . . . 310n to the high-resolution digital map 300 having a plurality of layers 310a . . . 310n, which contain specific information, such as road edges, traffic signs, buildings, etc., is connected functionally to the detection device 230. The at least one additional layer 310a . . . 310n comprises the detection indicators for the at least one detected object.

[0054] Advantageously, the proposed system 200 can be arranged in the automated measurement vehicle or in the cloud. In the first case, the additional layer 310a . . . 310n is ascertained directly when the sensor data is detected. In the second case, the additional layer 310a . . . 310n can be ascertained during the test drive or after its completion.

[0055] A validation of the sensor models can take place as a special application using the high-resolution digital map 300 according to the invention. In this way, it can be demonstrated that the sensor models perform the locating by means of the high-resolution digital map 300 in a similarly successful way to real sensor devices using real sensor data. In this way, the sensor devices can be efficiently tested by means of simulation methods. As a result, this can contribute in practice to an effective validation of automated vehicles with the aid of simulation processes.

[0056] Advantageously, the proposed method can be implemented as a software which runs, for example, on the apparatus 200. A simple adaptability of the method is supported in this way.

[0057] A person skilled in the art will suitably modify and/or combine the features of the invention without departing from the essence of the invention.