Determining method and determining device for determining reliability of low-resolution map

Abstract

The invention relates to a determining method and a determining device for determining the reliability of a low resolution map in order to make the activation of at least one driving assistance system of an autonomous vehicle more reliable, said autonomous vehicle travelling on a road and comprising a navigation system comprising said map and providing at least one slope called map slope and at least one curvature called map curvature of the road around said vehicle, and a perception system providing a slope called measured slope and a trajectory called measured trajectory of the road around said vehicle, said method comprising the following steps: • determining (201) the trajectory of said road; • determining (202) a slope indicator; • determining (203) a trajectory indicator; • calculating (204) a correlation value; • determining (205) a reliability indicator of said map.

Description

Technical Field

[0001] This invention claims priority to French application 2102516, filed on March 15, 2021, the contents of which (text, drawings and claims) are incorporated herein by reference.

[0002] This invention relates to the field of driver assistance systems for autonomous vehicles. Specifically, this invention relates to determining the reliability of low-resolution maps to make the activation of at least one driver assistance system for an autonomous vehicle more reliable. Background Technology

[0003] “Vehicle” is understood to mean any type of vehicle (e.g., motor vehicle, motor bicycle, motorcycle, storage robot in a warehouse, etc.). “Autonomous driving” in “autonomous vehicle” is understood to mean any method that can assist the driving of a vehicle. The method may thus be intended to partially or completely guide the vehicle or provide any type of assistance to the natural person driving the vehicle. The method is thus covered by autonomous driving from level 0 to level 5 in the OICA (“Organisation International des Constructeurs Automobiles”) classification table.

[0004] Among the aforementioned driver assistance systems are known examples such as lane keeping assist devices, lane changing assist devices, adaptive cruise control devices, etc.

[0005] Vehicles including one of these devices include numerous sensors (e.g., cameras, RADAR, LIDAR, ultrasonic meters, accelerometers, inertial centrality sensors, position or location sensors, velocity sensors, acceleration sensors, etc.). Information processing is known to be performed by at least one computer mounted in the vehicle, enabling environmental perception. "Environment" is understood to refer to both the exterior and interior of the vehicle. This environmental perception enables the measurement of the slope of the road over which the vehicle travels and the determination of the trajectory of the road over which the vehicle travels. The determination of the road trajectory involves, for example, the determination of the coefficients of at least one polynomial (which models the road trajectory). For example, the trajectory is divided into multiple segments or sub-segments. Each segment or sub-segment is modeled by a polynomial.

[0006] In addition, the vehicle includes a navigation system, which comprises components for self-localization and a map. Different types of maps exist. High-resolution maps, called HD maps, represent lanes and road-related attributes (e.g., number of lanes, road curvature, road gradient, road signs, etc.). Furthermore, the navigation system (which connects to a server outside the vehicle via a telecommunications link) regularly updates the HD map. Thus, the road-related attributes are very up-to-date. Additionally, in HD maps, these attributes are well-positioned (at a resolution of 1 meter).

[0007] In simple resolution SD maps, the lanes are depicted less accurately, and the attributes are not always digitized well (missing, or with less precise positioning by a few meters). The maps are not updated very regularly. When the update is performed manually and annually based on the map provider's database, the data is on average six months back.

[0008] A verification apparatus for verifying high-resolution maps of motor vehicles is known from document FR3090547. An estimation method for estimating the accuracy of high-resolution maps is known from document CN111767354. A driver assistance system mounted on a vehicle is known from document FR3082349. A vehicle driver assistance device and a vehicle driver assistance method are known from document US2019 / 227545. A driver assistance system is known from document US2007 / 299606. A determination apparatus for determining the reliability of information used for driver assistance is known from document EP2282170.

[0009] HD maps are inherently reliable due to their regular updates and attribute accuracy. Therefore, it is known that vehicle ADAS systems incorporating HD maps use information from these maps to improve the functionality of these ADAS systems. For example, these functions can affect the longitudinal and lateral dynamics of the vehicle by activating driver assistance systems (e.g., modifying set speeds, modifying the vehicle's positioning within lanes, etc.). Unfortunately, the cost of installing HD maps is very high relative to the manufacturing cost of the vehicle. When the vehicle is equipped with SD maps, adjusting or modifying the activation or modification of settings for driver assistance systems (which directly affect the vehicle's direction and / or speed) is dangerous. SD maps are not sufficiently reliable.

[0010] For use by a driver assistance system, several devices are known for calculating the instantaneous reliability index of the map. Specifically, positioning accuracy provided by devices such as GPS (Global Positioning System) is used. However, these devices for calculating the instantaneous reliability index rely on the premise that the information used to calculate the index is continuously available, and remain largely conceptual in practice. Unfortunately, environmental perception requires a large amount of information to be processed and thus a computationally intensive task. Consequently, in certain usage scenarios (many vehicles around the autonomous vehicle, strong acceleration or deceleration, etc.) or weather conditions, some perception information is frequently missing or delayed. Furthermore, perception information from image processing is also frequently missing due to available visibility (e.g., a truck may obscure road signals at the edge of a lane). Therefore, the periodic calculation of the instantaneous reliability index of the map changes abruptly and randomly as the vehicle travels on the road. In a driver assistance system, this index is thus unstable, unreliable, and unusable for safety reasons. Summary of the Invention

[0011] The purpose of this invention is to overcome the above-mentioned problems, specifically to make the information representing lanes and attributes from SD maps more reliable.

[0012] To this end, a first aspect of the present invention relates to a determination method for determining the reliability of a low-resolution map to make activation of at least one driving assistance system for an autonomous vehicle traveling on a road and including a navigation system and a perception system, the navigation system including the map and providing at least one slope, referred to as map slope, and at least one curvature, referred to as map curvature, of the road surrounding the vehicle, the perception system providing a slope, referred to as measured slope, and a trajectory, referred to as measured trajectory, of the road surrounding the vehicle, the method being periodically updated as the vehicle travels on the road, and comprising the following steps:

[0013] • A trajectory, referred to as a map trajectory, is determined based on at least the aforementioned map curvature;

[0014] • Determine a slope index that indicates the consistency between the slope on the map and the measured slope. The slope index is a number between 0 and 1, where a value of 1 indicates strong consistency and a value of 0 indicates no consistency.

[0015] • Determine a trajectory index, which indicates the consistency between the map trajectory and the measured trajectory. The trajectory index is a number between 0 and 1, where a value of 1 indicates strong consistency and a value of 0 indicates no consistency.

[0016] • Calculate the correlation value based on the slope index and the trajectory index;

[0017] • Determine whether the reliability index of the map is reliable or unreliable. When the correlation value is greater than a predetermined threshold, the index is reliable.

[0018] Therefore, when good reliability is available, this means that the map data is essentially equal to the measured data. The navigation system thus positions the vehicle well on the map, and the map is locally reliable (within the perceived distance). Driver assistance systems that control the longitudinal and / or lateral dynamics of the vehicle can rely on data from the map.

[0019] The use of these two indicators provides predictive and reactive correlation value updates. The correlation value updates are predictive because the trajectory indicator is based on the long-range field of view of the sensing system (e.g., between 30 and 300 meters). The correlation value updates are reactive because the slope indicator is based on a field of view very close to the vehicle, and the slope measurement is only known where the vehicle is passing or has passed.

[0020] Furthermore, due to the overload computation, the reliability is also robust to random short errors in the detection and positioning. This robustness stems from the calculation of the correlation value and the determination of each level.

[0021] Advantageously, the reliability index of the map is determined according to four levels: strong, moderate, insufficient, or uncertain, which are defined based on the correlation value.

[0022] Therefore, when good reliability has been previously established, and then the correlation value decreases in the absence of a map track and / or a measured track or a map slope and / or a measured slope, it does not directly transition to insufficient reliability or uncertain reliability, which poses a risk of sudden discontinuation of driving assistance.

[0023] Advantageously, the slope index increases or decreases according to a calculated gradient, which depends on the deviation between the slope on the map and the measured slope.

[0024] Therefore, when good reliability has been previously established, and then the gradient index decreases, taking into account history, it will not directly transition to weak reliability or uncertain inconsistency, which carries the risk of sudden discontinuation of driving assistance.

[0025] Advantageously, the trajectory index increases or decreases according to a calculated gradient, which depends on the lateral deviation between the map trajectory and the measured trajectory.

[0026] Therefore, when good reliability has been previously established, and then the trajectory indicators decrease, taking into account history, it will not directly transition to weak reliability or uncertain inconsistency, which poses a risk of sudden discontinuation of driving assistance.

[0027] Advantageously, the correlation value is a weighted average between the slope index and the trajectory index.

[0028] Advantageously, the determination of the map trajectory is obtained based on the interpolation of the at least one map curvature.

[0029] A second aspect of the invention relates to an apparatus comprising a memory associated with at least one processor configured to implement a method according to a first aspect of the invention.

[0030] The present invention also relates to a vehicle that includes the device.

[0031] The present invention also relates to a computer program comprising instructions adapted to perform the steps of the method according to the first aspect of the present invention when the program is executed by at least one processor. Attached Figure Description

[0032] Other features and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments and the accompanying drawings, in which:

[0033] - Figure 1 An apparatus according to a specific embodiment of the present invention is illustrated schematically.

[0034] - Figure 2 A method for determining the reliability of a low-resolution map according to a specific embodiment of the present invention is illustrated schematically. Detailed Implementation

[0035] The invention is described below in a non-limiting application in the case of autonomous motor vehicles traveling on roads or lanes. Other applications (e.g., robots in storage warehouses or motorcycles on rural roads) are also possible.

[0036] Figure 1 An example of device 101 is shown, which may be included in the vehicle, in a network (“cloud”), or in a server. Device 101 may be used as a reference hereinafter. Figure 2A centralized device for at least some steps of the described method. In an embodiment, the device corresponds to an autonomous driving computer.

[0037] In this invention, device 101 is included in the vehicle.

[0038] The device 101 can take the form of a box including printed circuits, any type of computer, or a mobile phone (“smartphone”).

[0039] The apparatus 101 includes a random access memory 102 for storing instructions for implementing at least one step of the method as described above by a processor 103. The apparatus also includes a mass storage memory 104 for storing data to be retained after the method is implemented.

[0040] The device 101 may also include a digital signal processor (DSP) 105. The DSP 105 receives data and shapes, demodulates, and amplifies the data in a manner known per se.

[0041] The device 101 further includes an input interface 106 and an output interface 107, the input interface being used to receive data implemented by the method according to the invention, and the output interface being used to transmit data implemented by the method according to the invention.

[0042] Figure 2 A method for determining the reliability of a low-resolution map according to a specific embodiment of the present invention is illustrated schematically.

[0043] The method is implemented by device 101 in an autonomous vehicle, which includes at least one driving assistance system. The autonomous vehicle travels on a road and includes a navigation system and a perception system. The navigation system includes the map and provides at least one slope, referred to as map slope, and at least one curvature, referred to as map curvature, of the road surrounding the vehicle. The perception system provides a slope, referred to as measured slope, and a trajectory, referred to as measured trajectory, of the road surrounding the vehicle.

[0044] As the vehicle passes, the method updates periodically, for example, every 100ms, but other values ​​are also possible. In an embodiment, the update is also performed according to a detected event or according to a command from one of the driving assistance systems.

[0045] Step 201 (DetTrajc) is a determination step for determining a trajectory, called a map trajectory, of the road based on at least the map curvature.

[0046] In a preferred operating mode, the navigation system provides information about the area around the vehicle within a field of view, which, for example, covers an area from 3000 meters in front of the vehicle to -500 meters (and thus behind the vehicle). Within this field of view, the map is divided into segments. Each segment represents a section of the road on which the vehicle travels. The segments therefore include distances. The at least one map curvature corresponds to the curvature of each segment and the length of that segment. Connected end-to-end, all segments present the road on which the vehicle travels within the provided field of view.

[0047] Therefore, based on the curvature of each segment, a trajectory, referred to as a map trajectory, is determined for the road. This determination is performed, for example, by determining a polynomial calculated based on data of the at least one map curvature. In another example, this determination is obtained based on interpolation of the at least one map curvature.

[0048] Step 2020 (DetIp) is a determination step for determining a slope index, which indicates the consistency between the slope on the map and the measured slope. The slope index is a number between 0 and 1, where a value of 1 indicates strong consistency and a value of 0 indicates no consistency.

[0049] In the operating mode, when the absolute deviation and / or relative deviation between the map slope (i.e., the slope provided by the navigation system at the location where the vehicle is traveling) and the measured slope is less than a predetermined threshold, the slope index is set to 1; otherwise, the slope index is set to 0.

[0050] In the preferred operating mode, the slope index increases or decreases according to a calculated gradient, which depends on the deviation between the slope on the map and the measured slope. The gradient is a number between -1 and 1. The gradient value is selected based on the deviation. For example, when the deviation is less than a threshold, the gradient equals 0.002; otherwise, the gradient equals -0.002. Thus, in each cycle, the slope index is updated and therefore takes into account history, ensuring robustness against random measurement errors and / or errors and / or inaccuracies in the map.

[0051] In another operating mode, the gradient also depends on the change in the measured slope. In the presence of strong acceleration or deceleration, the measured slope is temporarily more prone to error. For example, the gradient is equal to 0.004 when the change in slope is less than another predetermined threshold, and when the deviation between the map slope and the measured slope is less than the predetermined threshold.

[0052] Step 203 (DetItraj) is a determination step for determining a trajectory index that indicates the consistency between the map trajectory and the measured trajectory. The trajectory index is a number between 0 and 1, where a value of 1 indicates strong consistency and a value of 0 indicates no consistency.

[0053] In the operation mode, when the absolute lateral deviation and / or relative lateral deviation between the map trajectory and the measured trajectory is less than a predetermined threshold, the slope index is set to 1; otherwise, the slope index is set to 0.

[0054] In the preferred operating mode, the trajectory index increases or decreases according to a calculated gradient, which depends on the lateral deviation between the map trajectory and the measured trajectory. The gradient is a number between -1 and 1. The gradient value is selected based on the deviation. For example, when the deviation is less than a threshold, the gradient equals 0.002; otherwise, the gradient equals -0.002. Thus, in each cycle, the trajectory index is updated and therefore takes history into account, ensuring robustness against random measurement errors and / or errors and / or inaccuracies in the map.

[0055] In another operating mode, for each segment, the gradient depends on a combination of deviations between the map trajectory and the measured trajectory. The smaller the deviation is than a threshold, the higher the gradient.

[0056] Step 204 (CalcCor) is a calculation step for determining the correlation value based on the slope index and the trajectory index. Advantageously, the correlation value is a weighted average between the slope index and the trajectory index. The correlation value is a number between 0 and 1. For example, the correlation value is equal to 40% of the slope index plus 60% of the trajectory index. The percentages provided are merely examples, and other values ​​are possible. Ideally, the percentage of the slope index is smaller than that of the trajectory index because the trajectory index is more robust. The percentage also depends on other parameters, such as at least one threshold for variation of the measured slope.

[0057] Step 205 (DetFia) is a determination step for whether the reliability index of the map is reliable or unreliable. The index is considered reliable when the correlation value is greater than a predetermined threshold. Advantageously, the determination of the map's reliability index is performed according to four levels: strong, moderate, insufficient, or uncertain, defined by the correlation value. Historical data is taken into account when the trajectory index and the slope index change according to the gradient. Thus, when the correlation value is close to 1, it means that the slope index and the trajectory index are close to 1 for a period of tens of seconds, indicating very good consistency between the map and the environment. The map is therefore reliable and will be reliable for the next several tens of seconds, even if the vehicle experiences strong acceleration / deceleration (which degrades the slope measurement) and even if there is steering wheel movement for avoidance (which interferes with the determination of the trajectory measured by the sensing device).

[0058] This invention is not limited to the embodiments described above as examples; it extends to other variations. For example, specific numerical values ​​for the gradient are provided. Other values ​​are also possible and, for example, depend on the refresh cycle of the process.

Claims

1. A determination method for determining the reliability of a low-resolution map to make activation of at least one driving assistance system for an autonomous vehicle traveling on a road and including a navigation system and a perception system, the navigation system including the map and providing at least one slope, referred to as map slope, and at least one curvature, referred to as map curvature, of the road surrounding the autonomous vehicle, the perception system providing a slope, referred to as measured slope, and a trajectory, referred to as measured trajectory, of the road surrounding the autonomous vehicle, the determination method being periodically updated as the autonomous vehicle travels on the road, and comprising the following steps: • Determine the trajectory of the road, referred to as the map trajectory, based on at least the map curvature (201); • Determine (202) a slope index, which indicates the consistency between the map slope and the measured slope, the slope index being a number between 0 and 1, with a value of 1 indicating strong consistency and a value of 0 indicating no consistency; • Determine (203) a trajectory index, which indicates the consistency between the map trajectory and the measured trajectory, the trajectory index being a number between 0 and 1, where a value of 1 indicates strong consistency and a value of 0 indicates no consistency. • Calculate the (204) correlation value based on the slope index and the trajectory index; • Determine whether the reliability index of the map (205) is reliable or unreliable, and the index is reliable when the correlation value is greater than a predetermined threshold.

2. The determination method according to claim 1, wherein, The reliability index of the map is determined (205) according to four levels: strong, medium, insufficient or uncertain, and the level is limited according to the correlation value.

3. The determining method according to claim 1 or 2, wherein, The slope index increases or decreases according to a calculated gradient, which depends on the deviation between the slope on the map and the measured slope.

4. The determining method according to claim 1 or 2, wherein, The trajectory index increases or decreases according to a calculated gradient, which depends on the lateral deviation between the map trajectory and the measured trajectory.

5. The determining method according to claim 1 or 2, wherein, The correlation value is a weighted average between the slope index and the trajectory index.

6. The determining method according to claim 1 or 2, wherein, The determination of the map trajectory is based on the interpolation of at least one map curvature.

7. An apparatus (101) including a memory (102) associated with at least one processor (103), the at least one processor being configured to implement the determination method according to any one of claims 1 to 6.

8. A vehicle comprising the device according to claim 7.

9. A computer program product, the computer program comprising a computer program including instructions adapted to perform the steps of the determining method according to any one of claims 1 to 6 when the computer program is executed by at least one processor (103).

Images