Method and device for range estimation of a lidar sensor

Abstract

Method for estimating the range of a lidar sensor (2), wherein objects (O1 to O4) in the vicinity of the lidar sensor (2) are detected for environment detection and distances of the objects (O1 to O4) to the lidar sensor (2) are determined, characterized in that - the lidar sensor (2) comprises several photodetector elements, each of which is assigned to a field of view sector, wherein a sensitivity (E1 to En) of the respective photodetector element is adapted to a light radiation received from the respective field of view sector, - determines whether the sensitivity (E1 to En) of one of the photodetector elements has been reduced by more than a predefined threshold during environmental detection, and - such a photodetector element with such reduced sensitivity (E1 to En) is disregarded in the range estimation if an object (O1 to O4) is detected in the field of view sector assigned to this photodetector element whose distance to the lidar sensor (2) is less than a specified distance value.

Description

[0001] The invention relates to a method for estimating the range of a lidar sensor according to the preamble of claim 1.

[0002] The invention further relates to a device for estimating the range of a lidar sensor according to the preamble of claim 3 and a method for automated driving operation of a vehicle.

[0003] A vehicle equipped with a lidar system is known from DE 10 2018 221 083 A1. The lidar system is designed to scan an environment with a light beam to acquire information about it. Furthermore, the lidar system is designed to detect both highly reflective objects in the near field and low-reflective objects in the far field. The lidar system includes several photodetectors configured to have different saturation probabilities. Multiple dimming filters are provided, arranged in a common filter matrix and forming a single filter component. Each dimming filter is positioned upstream of one of the photodetectors to reduce the saturation probability of the respective photodetector downstream of the dimming filter.

[0004] Furthermore, DE 10 2020 128 877 B3 discloses a method for determining a change in the range of a lidar sensor for a vehicle or robot. In a reference measurement, a reference noise level of infrared radiation received by the lidar sensor and a signal-to-noise ratio of infrared radiation reflected from the reference target and received by the lidar sensor are determined with a reference target located at a predetermined distance from the lidar sensor. In a driving measurement, a current noise level of infrared radiation received by the lidar sensor is determined during operation of the vehicle or robot. From this current noise level, a theoretical distance to a position where the reference target would have to be located if the signal-to-noise ratio were the same as in the reference measurement.Furthermore, a range change calculation determines a deviation between the specified distance and the theoretical distance, whereby the deviation corresponds to the change in the range of the lidar sensor compared to its range during the reference measurement.

[0005] From DE 10 2019 005 148 A1 a method for detecting fog using a lidar sensor is known, wherein reflections at a static retroreflector are evaluated.

[0006] From DE 10 2020 115 252 A1 a method for detecting contaminants on a protective screen of a lidar sensor is known, wherein it is provided that background noise is determined in a detection area of ​​the lidar sensor and in a sector of the detection area and from this a conclusion is drawn about a contamination and the location of the contamination.

[0007] The invention is based on the objective of - a novel method for estimating the range of a lidar sensor, - a novel device for range estimation of a lidar sensor and - to specify a novel method for the automated driving operation of a vehicle.

[0008] The problem is solved according to the invention by - a method for estimating the range of a lidar sensor, which has the features specified in claim 1, - a device for range estimation of a lidar sensor, which has the features specified in claim 3, and - a device for automated driving of a vehicle, which has the features specified in claim 6.

[0009] Advantageous embodiments of the invention are the subject of the dependent claims.

[0010] In an inventive method for estimating the range of a lidar sensor with several photodetector elements, each assigned to a field of view sector and whose sensitivity is adapted to light radiation received from the respective field of view sector, objects in the vicinity of the lidar sensor are detected by means of the lidar sensor and their distances to the lidar sensor are determined.

[0011] Furthermore, it is determined whether the sensitivity of a photodetector element has been reduced by more than a predefined threshold during environmental detection. Such a photodetector element with this reduced sensitivity is disregarded in the range estimation if an object is detected in the field of view sector assigned to this photodetector element whose distance to the lidar sensor is less than a predefined distance value.

[0012] For safe and stable automated, especially highly automated or autonomous, driving based on environmental data acquired by a lidar sensor, a reliable detection range estimate is required. This estimate indicates the maximum distance at which the lidar sensor can confidently detect objects. Based on this, the vehicle's speed is regulated, for example, so that the maximum permissible speed increases with increasing detection range. A problem with environmental perception using a lidar sensor is that reflections of sunlight from surrounding objects, such as vehicles, can falsely lead to shorter detection ranges, necessitating a reduction in the maximum driving speed. This is because the sunlight hitting the photodetector element causes an adaptation of its sensitivity.This sensitivity is adjusted to avoid oversaturation by ambient light. However, this also leads to a reduction in the detection range.

[0013] However, the present method makes it particularly advantageous to avoid this when the light source is not the sun, but rather reflections from objects in the vicinity. This is achieved very reliably by disregarding a photodetector element with a sensitivity reduced by more than the specified threshold in the range estimation if an object is detected in the field of view sector assigned to this photodetector element that is located in close proximity to the lidar sensor. The close proximity is limited by the specified distance value.

[0014] If, on the other hand, direct sunlight is responsible for the reduction in sensitivity, there is generally no object in the immediate detection range of the lidar sensor in the direction of the sunlight. Therefore, "glare" on the corresponding photodetector element of the lidar sensor caused by reflections can be almost completely ruled out, so that the corresponding photodetector element continues to be used and, consequently, the vehicle's speed may be reduced during automated driving mode.

[0015] This suppresses false-positive measurements, thereby increasing the stability of the detection range estimate when reflections from objects are caused by strong sunlight, and thus enabling higher vehicle speeds during automated driving. At the same time, impairments to the detection range caused by sunlight are still reliably detected. Since this effect also occurs with halogen brake lights located within the detection range of the lidar sensor, the present method also enables greater stability in the detection range estimate for vehicles with halogen brake lights ahead.

[0016] This means that by including additional object information, such as the position and speed of an object, from a tracker in the detection range calculation, the stability of this can be increased by detecting side effects, such as sun reflections on vehicle sides on lateral sides of the lidar sensor's detection area, and not including them in the detection range calculation.

[0017] According to one possible implementation of the method, the specified distance value is chosen within a range between 0 m and 50 m. Such a distance value has proven particularly useful for the reliable execution of the method, since reflection is highly likely when objects are located at this distance and light radiation originates from the direction of the objects.

[0018] In a device according to the invention for estimating the range of a lidar sensor with several photodetector elements, each assigned to a field of view sector and whose sensitivity is adjustable to light radiation received from the respective field of view sector, the lidar sensor is coupled to or includes a computing unit. The computing unit is configured to detect objects in the vicinity of the lidar sensor and to determine their distances to the lidar sensor based on data acquired by means of the lidar sensor.

[0019] The computing unit is still trained, - to determine whether the sensitivity of a photodetector element has been reduced by more than a predetermined threshold during environmental detection, and - to disregard such a photodetector element with such reduced sensitivity in the range estimation if an object has been detected in the field of view sector assigned to this photodetector element whose distance to the lidar sensor is less than a specified distance value.

[0020] Similar to the previously described method, the device prevents erroneously shorter detection ranges from being determined, thus avoiding the need to reduce the maximum driving speed. This suppresses false-positive measurements, thereby increasing the stability of the detection range estimate when reflections from objects are caused by strong sunlight, and thus enabling higher vehicle speeds during automated driving.

[0021] According to one possible embodiment of the device, the specified distance value lies in a range between 0 m and 50 m. Such a distance value has proven particularly useful for reliable operation of the device, since reflection is highly likely when objects are present at this distance and light radiation originates from the direction of the objects.

[0022] According to another possible embodiment of the device, the photodetector elements are each designed as an avalanche photodiode. Such avalanche photodiodes are particularly advantageous for detecting very low radiation powers and thus enable reliable detection even with such low radiation powers or low-reflective objects in the environment.

[0023] In the inventive method for automated, in particular highly automated or autonomous, driving operation of a vehicle depending on the results of an environment sensing carried out by means of at least one lidar sensor, a maximum driving speed of the vehicle is selected during the automated driving operation of the vehicle depending on a detection range of the at least one lidar sensor. The detection range is determined in a previously mentioned method for estimating the range of a lidar sensor.

[0024] This method reliably prevents erroneously lower detection ranges from being determined, thus avoiding the need to reduce the vehicle's maximum speed during automated driving. This suppresses false-positive measurements, thereby increasing the stability of the detection range estimate when reflections from objects are caused by strong sunlight, and ultimately enabling higher vehicle speeds during automated driving.

[0025] Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

[0026] This shows: Fig. 1. Schematic image of a vehicle's surroundings and Fig. 2 schematically a point cloud of the environment generated by means of a lidar sensor according to Fig. 1 and sensitivities of different field-of-view sectors assigned to photodetector elements.

[0027] Corresponding parts are marked with the same reference symbols in all figures.

[0028] In Fig. Figure 1 shows an image B of the environment of a vehicle 1, captured by a camera. Fig. Figure 2 shows a map with a point cloud PW generated by a lidar sensor 2 of the vehicle 1, representing a top view of the surroundings according to Fig. 1 and sensitivities E1 to En of different field-of-view sectors assigned to photodetector elements of the lidar sensor 2.

[0029] The sensitivity E1 to En of the photodetector elements, also referred to as sensitivity, determines a detection range of the lidar sensor 2 and, in the case of automated driving operation of the vehicle 1 depending on data acquired by the lidar sensor 2, a maximum driving speed of the vehicle 1. The maximum driving speed decreases as the detection range decreases.

[0030] Solar radiation contains wavelengths in its spectrum that are also detected and analyzed by the lidar sensor 2. Particularly under direct sunlight, this leads to the problem of oversaturation of the photodetector elements, such as avalanche photodiodes. To counteract this oversaturation and minimize noise, the threshold of the photodetector elements can be raised or their sensitivity (E1 to En) reduced. This means that the sensitivity (E1 to En) of the photodetector elements is adjusted to the light radiation received from the respective field of view sector in order to avoid oversaturation caused by ambient light by reducing the sensitivity (E1 to En). The sensitivity (E1 to En) can also be adjusted for the entire detection range of the lidar sensor 2 as a function of ambient brightness and darkness.However, if the sensitivity E1 to En is reduced, reflections of real objects O1 to O4, such as vehicles and other road users, are less or not perceived at all in this field of view sector, and the detection range of the lidar sensor 2 is reduced in this field of view sector.

[0031] Using this information, the present map can be generated, which depicts the sensitivities E1 to En of the photodetector elements in the field-of-view sectors within the detection range of the lidar sensor 2. This map is also referred to as a gain map, and the detection range of the lidar sensor 2 can be estimated, for example, by averaging the sensitivities E1 to En shown in the map. Further parameters as input data for a detection range estimation algorithm could include, for example, the degree of soiling of a front cover of the lidar sensor 2, or weather conditions such as fog, rain, spray, etc.

[0032] If, as shown in the map, the sensitivities E1 to En of photodetector elements in certain fields of view drop in value, i.e., the respective value is reduced by more than a predetermined threshold, it is analyzed whether this drop was caused by direct sunlight or by reflection from an object O1 to O4.

[0033] To detect reflections from objects O1 to O4, the point cloud PW is used to determine whether objects O1 to O4 are located in the immediate vicinity of the lidar sensor 2, for example, at a distance of up to 50 m from the lidar sensor 2. If one of the detected objects O1 to O4, such as object O1 in the illustrated embodiment, is located in one of the fields of view sectors where the drop in sensitivity E1 to En of the photodetector elements was detected, it is assumed that the drop results from the reflection of sunlight from this object O1. Such a drop is represented in the map by an area A. If an object O1 is located in a field of view sector with a drop in sensitivity E1 to En, this field of view sector and the photodetector elements associated with it should not be considered in the range estimation of the lidar sensor 2.

[0034] If, on the other hand, direct sunlight is the reason for the drop in sensitivity E1 to En, which is recognized conversely when no object O1 to O4 is present in the corresponding field of view sector when sensitivity E1 to En drops, the corresponding field of view sector is taken into account in the range estimation and the maximum driving speed of vehicle 1 is supplemented accordingly if necessary.

[0035] A maximum detection range is particularly important in the central field of view of the lidar sensor 2, as this influences the driving speed of the automated vehicle 1. Therefore, any interference effects in the range estimation must be avoided. Here, a significantly improved range estimation is achieved using the described procedure by detecting dips in the sensitivities E1 to En of the photodetector elements caused by sunlight reflections, for example, from vehicle sides and / or windows to the side of the so-called field of view, especially the central field of view, and by disregarding these dips in the calculation of the detection range.

[0036] In addition to sunlight, halogen lamps also have the property of causing oversaturation of the photodetector elements at close range. This can occur, for example, when another vehicle is in a traffic jam, perhaps up to 10 m in front of vehicle 1, and its brake lights are illuminated by halogen light sources. Here, too, the previously described procedure can be applied analogously, and photodetector elements with reduced sensitivity E1 to En can be disregarded in the range estimation for lidar sensor 2 if an object O1 to On is detected in the field of view sector assigned to that photodetector element, and its distance to lidar sensor 2 is less than a predefined distance value.

Claims

[1] Method for estimating the range of a lidar sensor (2), wherein the lidar sensor (2) is used to detect objects (O1 to O4) in its environment and the distances of the objects (O1 to O4) to the lidar sensor (2) are determined, characterized by , that - the lidar sensor (2) comprises several photodetector elements, each of which is assigned to a field of view sector, wherein a sensitivity (E1 to En) of the respective photodetector element is adapted to a light radiation received from the respective field of view sector, - determines whether the sensitivity (E1 to En) of one of the photodetector elements has been reduced by more than a predefined threshold during environmental detection, and - such a photodetector element with such reduced sensitivity (E1 to En) is disregarded in the range estimation if an object (O1 to O4) is detected in the field of view sector assigned to this photodetector element whose distance to the lidar sensor (2) is less than a specified distance value. [2] Method according to claim 1, characterized by , that the specified distance value is chosen in a range between 0 m and 50 m. [3] Device for estimating the range of a lidar sensor (2), wherein the lidar sensor (2) is coupled to or includes a computing unit configured to detect objects (O1 to O4) in the vicinity of the lidar sensor (2) and to determine their distances to the lidar sensor (2) for environmental detection based on data acquired by means of the lidar sensor (2), characterized by , - that the lidar sensor (2) comprises several photodetector elements, each of which is assigned to a field of view sector, wherein a sensitivity (E1 to En) of the respective photodetector element is adjustable to light radiation received from the respective field of view sector, - that the computing unit continues to be trained, - to determine whether the sensitivity (E1 to En) of one of the photodetector elements was reduced by more than a predetermined threshold during environmental detection, and - to disregard such a photodetector element with such reduced sensitivity (E1 to En) in the range estimation if an object (O1 to O4) has been detected in the field of view sector assigned to this photodetector element whose distance to the lidar sensor (2) is less than a specified distance value. [4] Device according to claim 3, characterized by, that the specified distance value lies in a range between 0 m and 50 m. [5] Device according to claim 3 or 4, characterized by that the photodetector elements are each designed as an avalanche photodiode. [6] Method for automated driving operation of a vehicle (1) depending on results of an environment sensing carried out by means of at least one lidar sensor (2), wherein a maximum driving speed of the vehicle (1) specified during the automated driving operation of the vehicle (1) is selected depending on a detection range of the at least one lidar sensor (2), wherein the detection range is determined in a method for range estimation of a lidar sensor (2) according to claim 1 or 2.

Images