Method for detecting a degradation of a lidar sensor
Patent Information
- Application Number
- EP2023714767
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-05-30
- Filing Date
- 2023-03-30
- Publication Date
- 2025-04-09
AI Technical Summary
Current methods for detecting lidar sensor degradation are inadequate for ensuring reliable and safe operation of automated vehicles and robots, as they fail to accurately assess the impact of environmental influences and aging on sensor performance.
The method involves emitting lidar pulses and tracking reflections to determine a failure rate, which indicates the frequency of missed reflections over time, allowing for the assessment of sensor degradation and adaptation of vehicle operation accordingly, including reducing driving speed as sensor degradation increases.
This approach enables reliable detection of lidar sensor degradation, ensuring safe and efficient operation of automated vehicles by adapting driving styles and minimizing restrictions on automated operation, thereby enhancing traffic safety.
Smart Images
Figure 1.1
Abstract
Description
[0001] Method for detecting degradation of a lidar sensor
[0002] The invention relates to a method for detecting degradation of a lidar sensor.
[0003] The invention further relates to a method for operating a vehicle and / or robot.
[0004] DE 102018 008903 A1 discloses a method for determining the visibility of a beam-based sensor in a vehicle for environmental detection. The beam-based sensor emits sensor beams of known intensity, and the sensor detects reflections of the sensor beams and evaluates the reflected intensity. A measured distance and the reflected intensity are combined, and it is determined whether the sensor's visibility is reduced compared to its maximum visibility.
[0005] The invention is based on the object of providing a novel method for detecting a degradation of a lidar sensor and a novel method for operating a vehicle and / or robot.
[0006] The object is achieved according to the invention by a method for detecting a degradation of a lidar sensor, which has the features specified in claim 1, and by a method for operating a vehicle and / or robot, which has the features specified in claim 6.
[0007] Advantageous embodiments of the invention are the subject of the subclaims.
[0008] In the method for detecting degradation of a lidar sensor, according to the invention, lidar pulses are emitted by the lidar sensor, and reflections of the emitted lidar pulses are detected by the lidar sensor. Furthermore, an object from which lidar pulses are reflected is detected in the vicinity of the lidar sensor, with the detected object being tracked over several time intervals. Taking into account the distance of the tracked object from the lidar sensor and the geometry of the tracked object, it is determined which lidar pulses would have to be reflected when tracking the object. Furthermore, a failure rate is determined, which indicates how often an expected reflection is not detected within a specified period of time. Based on the failure rate and the distance to the tracked object, a conclusion is drawn about degradation of the lidar sensor.
[0009] In addition to determining the current degradation of the lidar sensor, for example, influenced by environmental factors or contamination, the present method also enables the reliable determination of aging and loss of individual transmitter-receiver pairs of the lidar sensor. This information enables particularly reliable and safe operation of an automated, particularly highly automated or autonomous vehicle and / or robot, since the information can be used to adapt the current driving style of the vehicle and / or robot to the degradation of the lidar sensor, and to determine the aging state of the lidar sensor.
[0010] In one possible embodiment of the method, a reduction in the range of the lidar sensor is determined as degradation. This allows the current driving style of the vehicle and / or robot to be adapted to the current range of the lidar sensor. For example, the driving speed of the vehicle and / or robot can be reduced with increasing degradation of at least one lidar sensor.
[0011] In another possible embodiment of the method, the failure rate is determined for each receiver pixel of the lidar sensor. This allows the information acquired by receiver pixels located outside of a degraded area to continue to be reliably used for the automated operation of the vehicle, thus minimizing any limitations on automated operation.
[0012] In another possible embodiment of the method, to detect a defect and / or aging of the respective receiver pixel, a comparison of the determined failure rate for each receiver pixel with the determined failure rates of neighboring receiver pixels is performed depending on the distance to the tracked object. This can further increase the reliability of the method.
[0013] In another possible embodiment of the method, the range of the lidar sensor is determined from the return intensity of the lidar pulses reflected from an object and from the distance of the lidar sensor to the object. This type of range determination is particularly simple, reliable, and accurate.
[0014] In the method according to the invention for operating a vehicle and / or robot, an environment of the vehicle and / or robot is detected by means of at least one lidar sensor and, depending on data detected by means of the lidar sensor, an automated, in particular highly automated or autonomous driving operation of the vehicle and / or robot is carried out, taking into account a degradation of the at least one lidar sensor detected by means of a previously described method.
[0015] Due to the reliable detection of the degradation of the lidar sensor, the method enables equally reliable operation of the automated vehicle and thus increases road safety.
[0016] In one possible embodiment of the method, in automated ferry operation, a driving speed of the vehicle and / or robot is reduced with increasing degradation of the at least one lidar sensor, so that, depending on the degradation of the lidar sensor, safe operation of the vehicle and thus high traffic safety can always be enabled.
[0017] Embodiments of the invention are explained in more detail below with reference to drawings.
[0018] Showing:
[0019] Fig. 1 schematically shows a receiver of a lidar sensor when detecting an object at a first distance,
[0020] Fig. 2 schematically shows the receiver according to Figure 1 when detecting the object at a second distance which is reduced compared to the first distance, Fig. 3 schematically shows a receiver of a lidar sensor when detecting an object and
[0021] Fig. 4 schematically shows a failure rate of a receiver of a lidar sensor as a function of a distance to a detected object.
[0022] Corresponding parts are provided with the same reference numerals in all figures.
[0023] Figure 1 schematically illustrates a receiver 1 of a lidar sensor detecting an object O at a first distance d. Figure 2 shows the receiver 1 according to Figure 1 detecting the object O at a second distance d that is smaller than the first distance d. The distance d is illustrated in more detail in Figure 4.
[0024] Lidar sensors scan their surroundings with infrared light and record a distance d and information about backscattered light for each measuring point or receiver pixel 1.1 to 1.n. This information is, for example, the intensity of the backscattered light. Due to the measuring principle, the number of measuring points or receiver pixels 1.1 to 1.n per unit area decreases with increasing distance d, since the receiver 1 of the lidar sensor measures in polar coordinates.
[0025] Consequently, as shown schematically in Figure 1, fewer receiver pixels 1.1 to 1.n receive a signal on an object O at a large distance d, for example 150 m, than when detecting an object O according to Figure 2, in which the object O has a significantly smaller distance d from the receiver 1. If objects O have been measured multiple times and tracked over time, a comparison can be made as to how often individual receiver pixels 1.1 to 1.n fail, even though a signal is expected at the receiver pixel 1.1 to 1.n due to existing object dimensions.
[0026] Figure 3 shows a receiver 1 of a lidar sensor detecting an object O. Object O is configured such that it would have to be detected by receiver pixels 1.3 to 1.5, 1.12 to 1.14, 1.21 to 1.23, and 1.30 to 1.32. Figure 4 shows a failure rate a for different receiver pixels 1.1 to 1.n of a lidar sensor as a function of the distance d from a detected object O.
[0027] If an object O is first detected at a first time t_0 and then continuously tracked over subsequent time points t_k up to a further time point t_m, a failure rate a can be determined for each receiver pixel 1.1 to 1.n of the lidar sensor based on a known geometry. The failure rate a can be combined with the distance d of the object O between the first time point t_0 and the further time point t_m. The failure rate a as a function of the distance d can thus be determined for each receiver pixel 1.1 to 1.n.
[0028] A comparison of the determined failure rate a as a function of the distance between different receiver pixels 1.1 to 1.n allows conclusions to be drawn about potentially defective or aged receiver pixels 1.1 to 1.n. Possible temporary interference, such as rain or a dirty windshield of the lidar sensor, can be identified by different temporal accumulations.
[0029] If the failure rate a is determined in the laboratory as a function of the distance d at the beginning of a lidar sensor's life cycle, these values can be used as a basis for determining temporary or permanent changes.
[0030] Figure 3 shows an example of a discrepancy between receiver pixel 1.13 and the defective receiver pixel 1.22 during the detection of object O. A failure rate a of 100%, shown in Figure 4, corresponds to a non-existent measurement, i.e., receiver 1 of the lidar sensor has not detected object O.
[0031] Determined failure rates a as a function of the distance d allow the following conclusions to be drawn. A relationship between the maximum range of the lidar sensor and the failure rate a is obtained according to where the factor ß can be determined based on laboratory measurements or previous sensor measurements and depends on the failure rate a.
[0032] This results in, for example, the max = 1.25 ■ d(tfc) with a(d(t fe )) = 25% and ß = f(a) = (25%) = 1.25. (2)
[0033] Differences in failure rate curves between two different
[0034] Receiver pixels 1.1 to 1.n can be permanently affected by defective or aged electronic components and / or optical components, and / or temporarily by a dirty lidar sensor windshield or external interference, such as rain or fog. This is illustrated in Figure 4 for the failure rate a(1.22) of the defective receiver pixel 1.22 and the failure rate a(1.13) of the neighboring receiver pixel 1.13.
Claims
Patent claims Method for detecting degradation of a lidar sensor, wherein - lidar pulses are emitted via the lidar sensor, - reflections of the emitted lidar pulses are detected by means of the lidar sensor, - an object (O) from which lidar pulses are reflected, is detected in an environment of the lidar sensor, - the detected object (O) is tracked over several time intervals, - taking into account a distance (d) of the tracked object (O) to the lidar sensor and a geometry of the tracked object (O), it is determined which lidar pulses would have to be reflected on the object (O) when tracking it, - a failure rate (a) is determined, which indicates how often an expected reflection is not detected within a given period of time, and - based on the failure rate (a) and the distance (d) to the tracked object (O), a degradation of the lidar sensor is inferred. Method according to claim 1, characterized in that a reduction in the range of the lidar sensor is determined as degradation. Method according to claim 1 or 2, characterized in that the failure rate (a) is determined for each receiver pixel (1.1 to 1.n) of the lidar sensor. Method according to claim 3, characterized in that, to determine a defect and / or aging of the respective receiver pixel (1.1 to 1.n), a comparison of the determined The failure rate (a) per receiver pixel (1.1 to 1.n) is determined using the failure rates (a) of neighboring receiver pixels (1.1 to 1.n) as a function of the distance (d) to the tracked object (O). A method according to one of the preceding claims, characterized in that the range of the lidar sensor is determined from the back-reflectance intensity of the lidar pulses reflected from an object (O) and from the distance (d) of the lidar sensor to the object (O). A method for operating a vehicle and / or robot, wherein - the environment of the vehicle and / or robot is detected using at least one lidar sensor and - Depending on data acquired by means of the lidar sensor, an automated, in particular highly automated or autonomous, ferry operation of the vehicle and / or robot is carried out, taking into account a degradation of the at least one lidar sensor detected by means of a method according to one of the preceding claims. Method according to claim 6, wherein in automated ferry operation the driving speed of the vehicle and / or robot is reduced with increasing degradation of the at least one lidar sensor.