Driving path model maintenance

Abstract

This disclosure relates to driving path model maintenance. Specifically, it relates to maintaining a driving path model during vehicle operation. The driving path model is configured to provide environmental perception of a current segment of a road to one or more driving automation system features, and includes at least one set of multiple lane model segments, each approximating a corresponding lane segment of a lane, and includes forward multiple lane model segments and rear multiple lane model segments. To maintain the driving path model, a curvature profile is derived for the distance ahead of the vehicle, indicating the curvature of the current segment of the road ahead. Based on the curvature profile, segments with approximately constant curvature changes are identified. Each forward multiple road segment is then maintained to include lane model segments based on the identified segments, and each set of rear multiple lane model segments is maintained by reassigning the lane model segments to each corresponding rear multiple segment.

Description

Technical Field

[0001] This disclosure generally relates to trajectory planning based on one or more driving automation system features of a vehicle, and more precisely to a driving path model and its lane model segments that approximate the route the vehicle is traveling on the road. Background Technology

[0002] Modern vehicles typically offer at least some level of driving automation. Environmental perception along the road route the vehicle is traveling on is necessary to enable driving automation. To acquire this environmental perception, the vehicle determines and maintains a driving path model using one or more driving automation system features; that is, the driving path model is initially generated and then updated as the vehicle travels along the road. The driving path model approximates the vehicle's travel along one or more lanes of the road using multiple lane segments for each of one or more lanes. Existing methods for determining and maintaining driving path models rely on approximating the road using lane model segments of constant length. The constant length of the lane model segments is typically chosen to ensure that the driving path model provides environmental perception of the road route along a predetermined length behind and in front of the vehicle at a given point in time, where multiple segments allow for real-time or at least near-real-time processing of the driving path model. However, depending on the road geometry, constant-length lane segments may not be able to approximate the road at the level of accuracy required to enable safe driving automation. Therefore, reducing the constant length of the lane model segments to accommodate all conceivable road geometries may conflict with processing requirements, as the driving path model needs to be processed by the vehicle in real-time or at least near-real-time to enable the vehicle to perform safe lateral and / or longitudinal guidance.

[0003] Therefore, the purpose of this disclosure is to maintain a driving path model that can accurately approximate possible road geometries while ensuring processing of the driving path model within a time frame suitable for safe lateral and / or longitudinal guidance of the vehicle. Summary of the Invention

[0004] To achieve this objective, this disclosure provides a method for maintaining a driving path model during vehicle operation. The driving path model is configured to provide environmental perception of a current segment of a road to one or more driving automation system features of the vehicle, and includes at least one plurality of lane model segments. Each set of plurality of lane model segments approximates a corresponding lane segment of the corresponding lane in the current segment of the road. Each plurality of lane model segments includes a forward plurality of lane model segments preceding the vehicle's current position and a rear plurality of lane model segments following the vehicle's current position. The method includes, when the vehicle's current position changes: deriving a curvature profile of the forward distance ahead of the vehicle's current position from one or more map datasets, the curvature profile indicating the curvature of at least one lane of the current segment of the road ahead of the vehicle; identifying segments with approximately constant curvature variation along the forward distance based on the curvature profile; maintaining each set of forward multiple lane model segments to include lane model segments based on the identified segments with approximately constant curvature variation; maintaining each set of rear multiple lane model segments by reassigning each set of forward multiple lane model segments to each corresponding rear multiple lane model segment based on the vehicle's current position; and providing a driving path model to one or more driving automation system features configured to perform lateral and / or longitudinal control of the vehicle based on the driving path model.

[0005] This disclosure also provides a corresponding vehicle control unit and a vehicle including the vehicle control unit. Attached Figure Description

[0006] Examples of this disclosure will be described with reference to the following figures, wherein the same reference numerals refer to the same elements.

[0007] Figures 1A to 1D A vehicle and driving path model based on the vehicle's current location traveling along a road is shown as an example according to this disclosure.

[0008] Figures 2A to 2D A driving path model for an exemplary road is shown, based on the curvature of the road and the corresponding curvature changes, according to an example of this disclosure.

[0009] Figures 3A to 3D A driving path model for an exemplary road is shown, based on the curvature of the road and the corresponding curvature changes, according to an example of this disclosure.

[0010] Figure 4 A driving path model for an exemplary road, which is also maintained based on vehicle sensor data, is shown as an example according to this disclosure.

[0011] Figure 5 Examples of this disclosure show the junctions where a lane branches off from or merges with another lane.

[0012] Figures 6A to 6C A flowchart illustrating a method for maintaining a driving path model during vehicle operation, according to an example of this disclosure, is shown.

[0013] Figure 7 A vehicle according to an example of this disclosure is shown.

[0014] Figure 8 An example of an automotive control unit according to this disclosure is shown.

[0015] It should be understood that the above-described drawings are in no way intended to limit this disclosure. Rather, these drawings are provided to aid in understanding this disclosure. Those skilled in the art will readily appreciate that aspects of this disclosure illustrated in one drawing may be combined with aspects in another drawing, or may be omitted without departing from the scope of this disclosure. Detailed Implementation

[0016] This disclosure provides a method and corresponding apparatus for maintaining a driving path model during vehicle operation. For a better understanding of this disclosure, reference will first be made to... Figures 1A to 1D The driving path model is discussed, followed by an outline of how to maintain it. Then, the discussion will move on to... Figures 2A to 8 The maintenance of the driving path model is discussed in detail.

[0017] In the context of this disclosure, the driving path model is configured to provide environmental perception about the current road segment to one or more driving automation system features of the vehicle. Figure 1A For example, in the scenario depicted, vehicle 700 is traveling along lane 112 of road 110. Figure 1B The image shows the covering on road 110 and Figure 1A The driving path model 140 corresponding to the situation depicted in the text. From Figure 1B As can be seen, the driving model 140 approximates the current segment of lane 112, which extends behind and in front of the vehicle 700 at its current position. By approximating the current segment of lane 112, the driving path model 140 provides environmental perception of road 110 to one or more driving automation system features of the vehicle 700.

[0018] Based on currently publicly available driving path models, such as Figures 1A to 1D The driving path model 140 approximates a segment of the road as the current segment of the road, so as to indicate how the segment of the road approximated by the driving path model changes as the current position of the vehicle 700 changes. Figure 1C Zhongyu Figure 1A As can be seen from the comparison, vehicle 700 is traveling along lane 112 of road 110. Therefore, compared to... Figure 1B In comparison, such as Figure 1DAs shown, the maintained driving path model 140 for the current segment of road 110 changes. In other words, the driving path model in the context of this disclosure is configured to provide environmental perception of the current location of vehicle 700, which typically changes during operation of vehicle 700.

[0019] More specifically, and as shown throughout the accompanying figures, the driving path model comprises at least one set of multiple lane model segments. Each set of multiple lane model segments approximates the corresponding lane segment of the corresponding lane in the current segment of the road. (Repeatedly) Figure 1B and Figure 1D Taking driving path model 140 as an example, driving path model 140 includes a multi-lane model segment, the start and end of which are indicated by black circles, and it approximates... Figure 1B and Figure 1D The current segment corresponds to lane 112 of road 110. The route of each segment is indicated by a dotted or dashed line between the start and end points. It will be understood that the end of one lane model segment coincides with the start of the next lane model segment.

[0020] Each lane model segment is shown throughout the accompanying drawings as corresponding to the centerline of the corresponding lane. While this method can generally be used given that the driving path model serves as the basis for the lateral and / or longitudinal control of the vehicle 700, the lane model segments among a given plurality of lane model segments can be located anywhere within the corresponding lanes where the longitudinal and / or lateral control of the vehicle 700 is actually achieved.

[0021] As shown throughout the accompanying drawings, lane model segments can have varying lengths. That is, in the context of this disclosure, lane model segments can cover any length of the corresponding segment of the corresponding lane based on the length of the segment of the given lane that exhibits an approximately constant curvature change, as will be discussed in more detail below with respect to steps 610 through 622 of method 600. Therefore, since the curvature of each lane segment changes linearly, the lane model segment can also be referred to as a cycloid spline.

[0022] It will be understood that, for example, due to computational and / or storage requirements, the maximum or minimum length of a lane model segment can be limited. The minimum length of a lane model segment can also be limited considering smoothness requirements. That is, if a given set of lane model segments includes very short segments, such as less than 1 m or less than 60 cm in length, then considering the resulting increase in the number of start and end points of the lane model segments, the longitudinal and / or lateral control of the vehicle 700 based on this driving path model can be altered more frequently. This frequent alteration of the vehicle 700's control can be perceived as bumps by the occupants of the vehicle 700, thereby reducing the perceived comfort of the longitudinal and / or lateral control of the vehicle 700 based on this driving path model. Smoothness requirements also ensure the safety of intended functions (SOTIF). As defined in standard ISO 21448:2022, SOTIF should be understood as the absence of unreasonable risks due to functional deficiencies, i.e., for the intended function of the driving path model, particularly due to performance deficiencies, i.e., providing a basis for making lateral and / or longitudinal control decisions.

[0023] Each lane model segment can be defined by a start point, an end point, an initial heading, an initial curvature, and a constant curvature change. The start and end points can be defined, for example, according to the coordinate system of vehicle 700, the origin of which is at the center of the rear axle. The start and end points in the vehicle 700 coordinate system can be derived from the location information of each lane included in one or more map datasets. It will be understood that, instead of or in addition to the vehicle 700 coordinate system, GPS coordinates or any other type of information indicating location can be used. Furthermore, it will be understood that each lane model segment can be defined by another and / or other parameters, taking into account the concept of a constant curvature change.

[0024] Although the driving path model is shown throughout the accompanying drawings including one or more lane model segments, it will be understood that the driving path model can include any number of lane model segments guaranteed by the travel of a vehicle 700 along a given road. For example, if the vehicle 700 is traveling along a road with two lanes in the direction of travel of the vehicle 700, the driving path model can include and be maintained to include two or more lane model segments, i.e., one or more lane model segments for each lane in the direction of travel of the vehicle 700. In another example, if the vehicle 700 is traveling along a road that typically has a single lane in the direction of travel of the vehicle 700 but partially includes a turning lane, then if the current segment based on the current position of the vehicle 700 includes a turning lane, the driving path model can similarly typically include one or more lane model segments corresponding to the single lane in the direction of travel of the vehicle 700, but may also include multiple lane model segments approximating turning lanes. Furthermore, the driving path model 700 may also include lanes for oncoming traffic (such as... Figures 1A to 1D Multiple lane model segments (lane 111) in lane 112 can be used, for example, to assist one or more driving automation system features of vehicle 700 in determining overtaking maneuvers of slower vehicles in lane 112 or driving maneuvers anticipating oncoming traffic. Using a section of Interstate 10, such as the Katie Expressway (i.e., the widest interstate highway in the United States), as an extreme example, the driving path model can include and be maintained to include 26 multiple lane model segments, i.e., all lanes in the direction of travel of vehicle 700, including through lanes, and all oncoming lanes and associated access roads. In another example, the driving path model can include one multiple lane model segment for the lane along which vehicle 700 is traveling and one multiple lane model segment for each adjacent lane, for a total of two multiple lane model segments on a two-way road or for a total of three multiple lane model segments on a road with multiple lanes in each direction. Therefore, depending on the requirements for the execution of lateral and / or longitudinal control of vehicle 700, the driving path model can include any number of multiple lane model segments based on the number of lanes present in a given current segment of the road along which the vehicle is traveling.

[0025] As described above, the driving path model can consider turning lanes, and more typically lanes branching off from lanes included in the driving path model, as well as lanes merging with lanes included in the driving path model. To this end, the junction points corresponding to the start or end points of new lane model segments can be determined based on lane model segments among multiple lane model segments approximating a given lane and the concept of constant curvature variation, respectively. These new lane model segments correspond to new lane model segments branching off from or merging with a given lane, as will be discussed in more detail below. If the current position of vehicle 700 is behind the junction point, the junction point of the new lane model segment can correspond to the start point of the first lane model segment of the preceding multiple new lane model segments. If the current position of vehicle 700 is in front of the junction point, the junction point can be the start and end points of two lane model segments among the following multiple new lane model segments, respectively.

[0026] As will be clear from the discussion of the number of multiple lane model segments, in the context of this specification, a road can be any type of vehicle infrastructure, ranging from a single-lane road to a controlled access highway with multiple lanes in each direction.

[0027] Each multiple lane model segment includes multiple lane model segments in front of the vehicle's current position and multiple lane model segments behind the vehicle's current position. For example, in... Figure 1B , Figure 1D , Figure 2D , Figure 3D and Figure 4 As can be seen, the corresponding driving path models in front and behind are shown below: the corresponding current position of vehicle 700 in the various attached diagrams is shown by patterned circles, each group of multiple lane model segments behind is shown by dotted line segments, and each corresponding group in front is shown by dashed line segments. To help understand this concept, Figure 1B , Figure 1D , Figure 2D , Figure 3D and Figure 4 This includes a legend explaining the current location and the multiple diagrams shown above, as well as those in front and behind.

[0028] The forward multiple lane model segments in each group can extend forward a distance in front of the vehicle 700 at a given current position. Similarly, the rear multiple lane model segments can extend upward a distance behind the vehicle 700 at a given current position. Since the length of each lane model segment depends on the length of a lane segment with a constant curvature variation, as will be discussed in more detail below, the total number of lane model segments included in the multiple rear and forward lane model segments, and therefore included in each group of multiple lane model segments along the combined rear and forward distances, can vary depending on the current position of the vehicle 700. In some examples of this disclosure, the total number of lane model segments included in each group of multiple lane model segments can be fixed, for example, taking into account storage and / or runtime requirements. In this case, if the corresponding multiple lane model segments would originally include more lane model segments than a fixed number, the rear multiple and / or forward multiple can extend along a distance less than the rear and / or forward distance. Therefore, in the examples disclosed herein, the total number of lane model segments is fixed, and the number of lane model segments can be distributed across the corresponding multiple front and rear lane model segments in a manner that ensures a sufficient percentage of the multiple front and rear lane model segments are covered by the lane model segments.

[0029] What will be understood is that the distance behind and the distance in front can be different from each other. For example, as Figure 1B , Figure 1D , Figure 2D , Figure 3D and Figure 4 This implies that the forward distance can be longer than the rear distance. For example, this could be a case where multiple rear lane model segments can be primarily used to provide perception of the road near the rear of vehicle 700 in order to estimate driving maneuvers of vehicles traveling behind vehicle 700, while multiple forward segments can be primarily used to determine the upcoming lateral and / or longitudinal control of vehicle 700. That is, multiple forward segments, and therefore forward distance, can be considered as indicating that vehicle 700 knows at least that vehicle 700 is traveling along the control horizon. Therefore, the forward distance can be longer than the rear distance. However, it should be understood that, depending on a given implementation of the driving path model, the forward and rear distances can also be substantially equal. For example, in some implementations of this disclosure, the rear and forward distances can be dynamically determined, for example, taking into account the speed of vehicle 700 and the driving environment. For example, in urban environments and / or at low speeds, the rear and forward distances can extend by similar distances from the current position of vehicle 700, because the required environmental perception of vehicles traveling behind vehicle 700 and the required environmental perception of the control horizon in dense urban traffic can be similar.

[0030] In some examples of this disclosure, the rear distance can be reduced to zero, thereby omitting multiple rear lane model segments from one or more sets of lane model segments in the driving path model. If, considering processing power constraints, it is necessary to allocate all available processing power to the multiple forward driving path models and the lateral and / or longitudinal control of the vehicle 700 based thereon, the rear distance can, for example, be reduced to zero. In such examples of this disclosure, backpropagation can be performed based on odometer data of the vehicle 700 to approximate each set of multiple rear lane model segments.

[0031] In the context of this disclosure, a vehicle, such as vehicle 700 in FIG. 6, refers to any type of motor vehicle configured to transport people and / or goods. The power unit of vehicle 700 can be any type of power unit, such as an electric motor or an internal combustion engine. Vehicle 700 can be, for example, a bus or any other type of vehicle, including one or more vehicle sensors 610 and a vehicle control unit 700, which enables vehicle 700 to perform the maintenance of the driving path model discussed below and one or more driving automation system features, including those that directly or indirectly use the driving path model of this disclosure to make lateral and / or longitudinal control decisions. Therefore, the one or more vehicle sensors 610 can be any type of vehicle sensor configured to capture vehicle sensor data indicative of the driving environment of vehicle 700, thereby providing environmental perception to vehicle 700. The one or more vehicle sensors 610 can be, for example, a camera, an infrared camera, a radar sensor, or a lidar sensor with different capture ranges and resolutions. One or more vehicle sensors 610 may also be vehicle sensors configured to capture odometer data (such as acceleration, speed, or any other odometer parameter) of vehicle 700 in any one or more of the three dimensions of either a local coordinate system (such as a vehicle coordinate system centered on the rear axis) or a global coordinate system (such as GPS coordinates).

[0032] Furthermore, in the context of this disclosure, the description of a driving automation system feature will be understood in the sense of SAE International standard J3016 as a design-specific function of a driving automation system at a given level of driving automation, i.e., in the context of this disclosure, as defined in the driving automation classification of standard J3016, such as any one of driving automation levels 1 to 5. More precisely, at least one of one or more driving automation system features of vehicle 610 is configured to determine, at least indirectly, the longitudinal and / or lateral control of vehicle 700 based on a maintained driving path model. Therefore, considering one or more driving automation system features, vehicle 700 can also be referred to as a vehicle 700 that is at least partially automated.

[0033] To maintain the aforementioned driving path model, the curvature profile along which at least one lane vehicle 700 is traveling is first derived from one or more map datasets. Furthermore, the curvature profile can be updated based on vehicle sensor data, for example, if one or more map datasets are incorrect, or if current traffic conditions necessitate changes to the driving path model, such as in the event of an accident or lane closure. Then, segments with constant curvature variations within the curvature profile are identified, which are used to generate lane model segments. If the total number of lane model segments is fixed, these lane model segments are then added to a given number of lane model segments ahead, up to a forward distance, or at least up to a number of lane model segments that can be covered ahead, in order to maintain the forward multiple. These steps are continuously performed as the position of vehicle 700 changes. Simultaneously, any lane model segments that no longer belong to the forward multiple based on the current position of vehicle 700 are reassigned to the rear multiple, and if the total number of lane model segments is fixed, they are eventually deleted from the rear multiple once they are beyond a backward distance or at least beyond a number of covered rear multiple.

[0034] Therefore, the maintenance of the driving path model is based on the curvature variation of the road vehicle 700 as it travels, and lane model segments are determined based on lane segments with constant curvature variation. This approach allows for variable lengths of the generated lane model segments, which in turn enables improvements in the approximation of road geometry while ensuring that the processing of the driving path model is performed in time frames suitable for safe lateral and / or longitudinal guidance of the vehicle, since lane model segments can cover long distances if the curvature variation remains constant over long distances.

[0035] Now refer to Figure 6A and Figure 6C The flowchart describes in detail the general concepts of maintaining a driving path model. Figure 6A and Figure 6C A method 100 for maintaining a driving path model during the operation of vehicle 700 is shown, and reference is made to... Figures 2A to 4 , Figures 2A to 4 An example of how to maintain a driving path model using method 600 is provided.

[0036] What will be understood is... Figures 6A to 6C The dashed boxes in the diagram illustrate optional method steps. It will also be understood that when the current position of vehicle 700 changes, Figures 6A to 6C The steps shown can be performed at least partially simultaneously, as will become apparent from the following discussion of method 600.

[0037] In step 610, method 600 derives a curvature profile of the distance ahead of the current position of vehicle 700 from one or more map datasets. The curvature profile indicates the curvature of at least one lane of the current segment of the road ahead of vehicle 700. .

[0038] Figure 2B It shows in Figure 2A The curvature profile 230 of road 210, indicated by a diamond shape, is between the starting point S and the ending point E. Therefore, the curvature profile 230 is slightly longer than the preceding distance. Figure 2A In this case, the forward distance will only extend from vehicle 700 (e.g., the rear axle of vehicle 700) to endpoint R. As can be seen, the curvature profile 700 indicates a curvature that is approximately zero during the straight sections of road 210. During the two turns on Road 210, the curvature The value of changes continuously, and then returns to a curvature of approximately zero at a constant rate of change. The situation had previously stabilized. Figure 2B The same concept is shown for road 310, i.e., a roundabout, which is shown to have curvature. The curvature profile 330 has a positive constant curvature for the extended portion of the roundabout of road 310, and includes curvature during entry and exit from the roundabout. Sections below zero. For example, in the case of curvature profile 230, the curvature change in curvature profile 330 exhibits a constant curvature change, because steering and roundabouts can generally be designed based on a constant curvature change so that vehicles such as vehicle 700 can travel safely and comfortably along such sections.

[0039] To determine the curvature profile, step 610 may include step 611, wherein method 600 can determine the curvature of the connection between each adjacent set of map data points of at least one lane. That is, lanes are typically included in the map data based on the start and end points of a segment. In step 611, method 600 may, for example, determine the connection between adjacent start and end points, or if the segment between the start and end points includes entrances and exits of turning or roundabouts, it has an approximately constant curvature change, or if the segment between the start and end points includes straight road segments or constant segments of turning or roundabouts, it has a constant curvature. It will be understood that method 600 may, in step 611, determine the curvature of a given road in any other way based on the concept of constant curvature change in the road design, using start and end points provided by one or more map datasets. .

[0040] One or more map datasets used to derive curvature profiles may include low-level map datasets, i.e., map data with low resolution and high detail, such as map data configured for use in at least partially autonomous driving, and / or crowd maps, i.e., map data generated based on vehicle sensor data or other inputs from other vehicles (such as all vehicles from a given original equipment manufacturer). It will be understood that any type of map data or a combination of various types of map data can be used to derive accurate curvature profiles for any lane used to maintain multiple lane model segments.

[0041] In step 620, method 600 identifies segments with approximately constant curvature variation along a forward distance based on curvature profiles. In other words, method 600 determines the curvature in step 620. The change over distance l, i.e., the derivative This concept is based on Figure 2A Road 210 and Figure 2B The curvature profile of 230 Figure 2C Curvature variation profile 240 and based on Figure 3A Road 310 and Figure 3B The curvature profile of 330 Figure 3C The curvature change profile 340 is shown in the figure. Since the curvature change used in road design is usually approximately constant, the curvature change profile typically changes from an approximately zero curvature change to an approximately constant value during sections of turning and roundabout intersections, as shown in curvature change profiles 230 and 330.

[0042] To identify segments where the curvature change is approximately constant, step 620 may include steps 621 and 622. In step 621, method 100 may compare the curvature profile with left-turn curvature change thresholds and right-turn curvature change thresholds. The left-turn curvature change thresholds and right-turn curvature change thresholds are... Figure 2C and Figure 3C The threshold t is used to define the values ​​of the two thresholds respectively. + and threshold t - As shown.

[0043] If the curvature change exceeds one of two thresholds, then method 100 can, in step 622, determine the point exceeding one of the two thresholds as the end point of the previous segment of the approximate constant curvature change and the corresponding starting point of the subsequent segment of the approximate constant curvature change. Figure 2C Taking the curvature change profile 240 as an example, the curvature change profile 230 exceeds the threshold t. + and threshold t - Eight times. Correspondingly, when excluding the start and end points corresponding to the start point S and end point E of road 210, the curvature profile 240 is maintained. Figure 2D The driving path model 250 includes eight combinations of start and end points. Similarly, Figure 3C The curvature change profile of 340 exceeds the threshold t + and threshold t - Eight times, and when excluding the start and end points corresponding to the start point S and end point E of road 310, the corresponding driving path model 350 includes eight combinations of start and end points.

[0044] It will be understood that step 620 can use methods other than the thresholding methods of steps 621 and 622 to identify segments of approximately constant curvature based on curvature profiles different from those discussed. For example, any point on the curvature profile that changes from zero to a non-zero value can be selected as the endpoint and corresponding start point of a given segment of approximately constant curvature. In another example, the derivative of the curvature profile, i.e. Any position containing a non-zero value can be chosen as the end point and corresponding start point of a given segment with approximately constant curvature.

[0045] As described above, one or more sets of lane model segments in the driving path model can each include a fixed number of lane model segments. Therefore, in an implementation of this disclosure, where the number of lane model segments in each set of multiple lane model segments is fixed, method 600 may include step 630, wherein method 600 distributes a fixed number of lane model segments across the forward and rear lane model segments in each set of multiple lane model segments. The fixed number of lane model segments can be distributed across multiple forward and rearward segments in a manner that ensures that a large portion of the rear and forward distances can be covered by the lane model segments. This ensures sufficient environmental awareness for one or more driving automation system features of vehicle 700.

[0046] In step 640, method 600 maintains each group of multiple lane model segments ahead to include lane model segments based on identified segments with approximately constant curvature changes. That is, based on the start and end points of the identified segments with approximately constant curvature changes, method 600 maintains each group of multiple lane model segments ahead to include lane model segments covering each lane of a given road, each lane of which is approximated by a driving path model of part or all of the forward distance along the current position of vehicle 700. To this end, step 640 may include step 641, in which method 600 may add lane model segments to each group of multiple lane model segments ahead based on the current position of vehicle 700 and the identified segments with approximately constant curvature changes (i.e., based on the successive execution of at least steps 610 and 620). It will be understood that steps 640 and 641 may therefore include generating lane model segments as vehicle 700 moves by deriving any of the aforementioned parameters used to define each lane model segment from the start and end points identified in step 620.

[0047] To generate lane model segments, method 600 may include step 641, in which method 600 approximates each identified segment with an approximate constant curvature change using at least one corresponding nth-order Taylor approximation and a corresponding n+1th-order Lagrange remainder. The nth-order Taylor approximation should be understood as an nth-order Taylor polynomial approximating a function connecting the start and end points of a given segment with a constant curvature change. Furthermore, the n+1th-order Lagrange remainder should be understood as an estimate of the error inherent in the function used to approximate the start and end points of the given segment with a constant curvature change using the nth-order Taylor approximation. The order n can be chosen based on the required accuracy of the approximation of the function connecting the start and end points of the given segment with a constant curvature change and the runtime requirements. For example, the order n can be chosen as 3, 4, or 5, i.e., the nth-order Taylor approximation can be a third, fourth, or fifth-order Taylor polynomial with corresponding fourth, fifth, or sixth-order Lagrange remainders.

[0048] To address the approximation error, method 600 may subsequently determine, in step 641a and step 641b, whether the corresponding (n+1)th order Lagrange remainder exceeds the approximation error threshold, and if so, perform any of steps 641c through 641f to resolve the approximation error. To determine whether the (n+1)th order Lagrange remainder exceeds the approximation error threshold, any parameter identifying the corresponding lane model segment can be used as input to the (n+1)th order Lagrange remainder, such as, for example, arc length, initial curvature, and / or the curvature variation of the corresponding lane model segment.

[0049] In step 641c, method 600 may limit the length of a given lane model segment corresponding to a given identified segment with an approximately constant curvature change to a length along which the corresponding n+1th order Lagrange remainder remains below the approximation error threshold; that is, the length is shorter than the length of the corresponding identified segment with an approximately constant curvature change. Therefore, method 600 may use another nth order Taylor approximation in step 641d to approximate the remainder of the length of the corresponding identified segment with a constant curvature change. In other words, method 600 may reduce the approximation error indicated by the n+1th order Lagrange remainder in steps 641c and 641d by approximating the given identified segment with an approximately constant curvature change using more than one lane model segment based on the corresponding nth order Taylor approximation.

[0050] In step 641e, method 600 may use one or more additional corresponding nth-order Taylor approximations and n+1th-order Lagrange remainders to approximate a given segment with constant curvature variation. The number of additional corresponding nth-order Taylor approximations may depend on the number of n+1th-order Lagrange remainders that exceed the approximation error threshold. That is, while in steps 641c and 641d, multiple lane model segments are generated to approximate a given segment with constant curvature variation if the corresponding n+1th-order Lagrange remainder exceeds the approximation error threshold, only one lane model segment is generated to approximate a given segment with constant curvature variation. However, multiple nth-order Taylor approximations are used to generate the lane model segment. The number of nth-order Taylor approximations used to generate the lane model segment is based on the frequency with which the error threshold is exceeded. Taking a 20-m long constant curvature variation section as an example, if the (n+1)th order Lagrange remainder corresponding to the first nth order Taylor approximation exceeds the approximation error threshold after 15 m, a lane model segment approximating the 20-m long constant curvature variation section can be generated based on the first nth order Taylor approximation for the first 15 m, and the remaining 5 m can be generated based on the second nth order Taylor approximation. In another example, this 20-m long constant curvature variation section can also be approximated by lane model segments based on two 10mn order Taylor approximations. If the number of lane model segments in each group of multiple lane model segments is finite, the method of step 641e can be selected, for example, i.e., lane model segments can be generated based on multiple nth order Taylor approximations if necessary, taking into account the (n+1)th order Lagrange remainder.

[0051] In step 641f, method 600 can reduce the curvature of the lane model segment corresponding to the given identified approximately constant curvature change segment to a curvature such that the corresponding n+1 order Lagrange remainder decreases below the approximate error threshold.

[0052] It will be understood that any one of steps 641c through 641f can be performed individually or in combination to generate a lane model segment that sufficiently approximates a given section with constant curvature variation. For example, method 600 may employ an optimizer to generate a lane model segment based on at least one nth-order Taylor approximation and the corresponding n+1th-order Lagrange remainder, aiming to reduce the n+1th-order Lagrange remainder to below an approximation error threshold based on any one of steps 641c through 641f described above.

[0053] In addition to maintaining each group of forward multiple lane model segments by generating and adding new lane model segments based on curvature profiles and curvature change profiles derived from one or more map datasets, method 600 may further modify each group of forward multiple lane model segments based on vehicle sensor data in step 640, as will be discussed below with respect to steps 643 to 646. Steps 643 to 643b address situations where the curvature indicated by one or more map datasets may not correspond to the actual road design encountered by vehicle 700. Steps 644 to 646 address situations where vehicle 700 encounters obstacles and requires corresponding modification of the given forward multiple lanes.

[0054] In step 642, method 600 may detect curvature change deviations within one or more lane model segments in each group of multiple lane model segments ahead, based on vehicle sensor data. That is, as vehicle 700 travels along a given road, method 600 may detect the curvature of the lane along which vehicle 700 is traveling, as well as the curvature of any additional lanes approximated by a given driving path model, based on data provided by one or more vehicle automatic sensors 610, which does not correspond to the curvature of the road derived from one or more map datasets in step 610. In step 643, method 600 may then continue to compare the curvature change deviation with a curvature change deviation threshold. If the curvature change deviation exceeds the curvature change deviation threshold, method 600 may then modify each group of multiple lane model segments ahead in the driving path model according to steps 643 and / or 643b, based on the detected curvature change deviation. In step 643a, method 600 may change the length of one or more lane model segments based on the detected curvature change deviation. In step 643b, method 600 may add one or more additional lane model segments to each group of preceding lane model segments based on the detected curvature change deviation. In other words, method 600 may detect a road curvature in steps 642 to 643b that differs from the curvature of a road as indicated by one or more map datasets, and may then adjust each group of preceding lane model segments as needed by changing their lengths and / or adding discontinuous lane model segments to approximate the actual curvature of the road detected based on vehicle sensor data. Therefore, it will be understood that implementations of method 600 including steps 642 to 643b may also include steps as part of steps 610 and 620, wherein the curvature profile and curvature change profile are updated based on vehicle sensor data.

[0055] The concepts of steps 642 to 643b are in Figure 4 As shown in the image. Figure 4 Road 410 and its sections, which differ from the actual layout of road 410 derived from one or more map datasets, are shown in gray. Therefore, based on one or more map datasets, road 410 and thus driving path model 460 are assumed to include long straight sections. However, the actual layout of road 410 includes a left turn, two right turns, and another left turn, for example due to construction work on the assumed long straight section. Therefore, as part of steps 642 to 643b, the lane model segment covering the assumed long straight section of road 410, which begins after the initial left turn and ends at the start of the final right turn, is either shortened, and 16 lane model segments are added after the shortened lane model segment and the lane model segment at the start of the final right turn, or simply replaced by 17 lane model segments.

[0056] In step 644, method 600 may, based on vehicle sensor data, detect obstacles, such as stationary vehicles or pallets, within one or more lane model segments of each group of preceding lane model segments on a given road. To avoid the obstacles, method 600 may then, in step 645, modify each group of preceding lane model segments by determining that one or more lane model segments have approximately constant curvature changes and bypass the obstacles. In other words, method 600 may employ the concept of designing lane steering to exhibit constant curvature changes in order to determine lane model segments that guide vehicle 700 around obstacles detected by one or more vehicle sensors 610. It will be understood that, in the event of a discrepancy between the curvature derived from map data and the curvature detected based on vehicle sensor data, modification of each group of preceding lane model segments based on the detected obstacles may be performed in the same manner as discussed above with respect to steps 643a and 643b. Therefore, steps 610 and 620 of method 600 may also include the step of updating the curvature profile and curvature change profile if an obstacle is detected in any of the lanes approximated by the driving path model.

[0057] Step 640 may ultimately include step 646, in which, if at least one group of multiple lane model segments exceeds a fixed number of lane model segments, method 600 may limit the number of lane model segments within each group of forward multiple lane model segments to cover distances shorter than the forward distance. That is, if the fixed number allocated to the forward multiple lane model segments is insufficient to cover the entire forward distance—for example, this might occur on very winding roads—step 646 ensures that the forward distance is a compromise between coverage in each group of forward multiple lane model segments and coverage by the fixed number of lane model segments.

[0058] Method 600 may include step 650, in which method 600 may determine, based on one or more map datasets and / or vehicle sensor data, at least one junction point corresponding to the start or end point of a new plurality of lane model segments, the new plurality of lane model segments corresponding to a lane that branches off from or merges with at least one lane of the current segment. The determination of the junction point may be based on identifying a segment on the lane model segment closest to the lane that branches off from or merges with at least one lane of the current segment, with a constant curvature change at the beginning or end. That is, the start of a lane that branches off from a given lane or the end of a lane that merges with a given lane may be considered, respectively, a turn away from or toward a lane with a constant curvature change as described above. Therefore, method 600 may determine the junction point in step 650 as the point on the lane model segment closest to the lane that branches off from or merges with at least one lane of the current segment. That is, the junction point can be any point between the start and end points of the lane model segment closest to the lane, which branches off from or merges with at least one lane in the current segment, from which a turn with a constant curvature change can be achieved. Depending on its position on the given lane model segment, the given lane model segment enables a turn with a constant curvature change on or from at least one lane in the current segment of the road, and the junction point can also coincide with the start or end point of the given lane model segment.

[0059] To illustrate the determination of the joint point as performed in step 650 based on the concept of constant curvature change, Figure 5 An example road 510 with a single lane is shown, from which lanes branch off and then rejoin. It can be seen that the junctions are all located on the single-lane model segment among the multiple lane model segments 5401 preceding it, and in this example, they do not coincide with its start and end points. Based on these junctions, lane model segments of new multiple lane model segments 5402 can be derived to approximate turning into and from the second lane of road 510.

[0060] Although Figure 5 The bifurcated and then merged lanes shown are depicted as parallel lanes, but it will be understood that the concept of determining the junction of lane model segments of multiple lane model segments approaching the bifurcated or merged lanes, and thus the start and / or end points, can also be applied to bifurcated or merged lanes in a given current segment that are not parallel to the road vehicle 700 traveling.

[0061] It will be understood that method 600 may execute steps 610 through 646 after step 650 to populate the front multiple of the new plurality of lane model segments. The rear multiple may initially be left empty, or may be initialized based on the plurality of lane model segments from which the new lanes branch off.

[0062] In step 660, method 600 maintains each group of rear multiple lane model segments by reallocating each group of forward multiple lane model segments to each corresponding rear multiple based on the current position of vehicle 700. To limit the number of lane model segments included in each group of rear multiple lane model segments, step 660 may include step 661, in which method 600 may remove lane model segments from each group of rear multiple lane model segments located beyond each rear distance. It will be understood that method 600 may also remove lane model segments from the rear multiple based on other methods, such as based on storage space allocated to the driving path model maintained by method 600. Furthermore, similar to step 646, step 660 may include step 661, in which method 600 may limit the number of lane model segments within each group of rear multiple lane model segments if at least one or more of the lane model segments exceed a fixed number of lane model segments. That is, in step 661, if the share allocated to a fixed number of rear multiple lane model segments is insufficient to cover the entire rear distance, then method 600 can ensure that the rear distance is a compromise between the coverage of each group of rear multiple lane model segments and the coverage of a fixed number of lane model segments.

[0063] When vehicle 700 starts, as vehicle 700 moves, the rearmost lanes may be filled based on steps 660 and 661 to approximate the corresponding lanes, or may need to be initialized, for example, if vehicle 700 turns from a parallel parking position into traffic. To this end, method 600 can initialize the rearmost lanes by performing steps 610 to 622 discussed above to generate lane model segments for the rearmost lanes. Therefore, method 600 can also derive the curvature profile of the rear distance behind the current position of vehicle 700 during the initialization of the rearmost lanes, and can identify segments with approximately constant curvature changes to generate lane model segments for the rearmost lanes.

[0064] Finally, in step 670, method 600 provides a driving path model to one or more driving automation system features, which are configured to perform lateral and / or longitudinal control of the vehicle based on the driving path model.

[0065] In summary, method 600 maintains a driving path model by continuously determining the curvature of the road in front of vehicle 700 based on map data, and then determining segments with approximately constant curvature changes based on the curvature of the road in front of vehicle 700. The points where the curvature changes are altered serve as the start and end points of lane model segments of the driving path model for one or more lanes of the road in front of vehicle 700. If obstacles or roads detected based on map data exhibit curvature different from what is expected, method 600 can update, for example, by replacing, adding, and / or shortening lane model segments, i.e., modifying the lane model segments.

[0066] Figure 8 An automotive control unit 800, also known as an electronic control unit (ECU), is shown configured to perform method 600. The automotive control unit 800 may include a processor 810, a graphics processing unit (GPU) 820, an automotive processing system 830, a memory 840, a removable storage device 850, a storage device 860, a cellular interface 870, a Global Navigation Satellite System (GNSS) interface 880, and a communication interface 890.

[0067] Processor 810 can be any type of single-core or multi-core processing unit employing Reduced Instruction Set Computing (RISC) or Complex Instruction Set Computing (CISC). Exemplary RISC processing units include ARM-based cores or RISC V-based cores. Exemplary CISC processing units include x86-based cores or x86-64-based cores. Processor 810 can execute instructions that cause vehicle control unit 800 to perform method 600. Processor 810 can be directly coupled to any component of vehicle control unit 810, or it can be directly coupled to memory 840, GPU 820, and device bus 800B.

[0068] GPU 820 can be any type of processing unit optimized for processing graphics-related instructions or, more generally, for parallel processing of instructions. Therefore, GPU 820 can be configured to generate a display of information to the driver of the vehicle, such as information related to one or more driving automation system features or telemetry data, for example, via a head-up display (HUD) or a display positioned within the driver's field of vision. GPU 820 can be coupled to the HUD and / or the display via connection 820C. GPU 820 can also perform at least a portion of method 600 to enable fast parallel processing of instructions capable of implementing method 600. It should be noted that in some embodiments, processor 610 may determine that GPU 820 does not need to execute instructions related to method 600. GPU 820 can be directly coupled to any component of the vehicle control unit 810, or can be directly coupled to processor 810 and memory 840. In some embodiments, GPU 820 can also be coupled to a device bus.

[0069] The automotive processing system 830 can be any type of system-on-a-chip configured to provide trillions of operations per second (TOPS), for example, to enable the automotive control unit 800 to execute one or more driving automation system features while driving, which can perform longitudinal and / or lateral control of the vehicle 700 based on the driving path model maintained by method 600. The automotive processing system 830 may interface only with processor 810, or it may interface with other devices via a system bus.

[0070] Memory 840 can be any type of fast storage device, enabling processor 810, GPU 820, and automotive processing system 830 to store instructions for fast retrieval during instruction processing, as well as cached and buffered data. Memory 840 can be a unified memory coupled to processor 810, GPU 820, and automotive processing system 830, allowing memory 840 to be allocated to processor 810, GPU 820, and automotive processing system 830 as needed. Alternatively, processor 810, GPU 820, and automotive processing system 830 can be coupled to separate processor memory 840a, GPU memory 840b, and automotive processing system memory 840c.

[0071] The removable storage device 850 may be a storage device that can be removably connected to the vehicle control unit 800. Examples include a digital multifunction disc (DVD), a compact disc (CD), a universal serial bus (USB) storage device (such as an external SSD), or magnetic tape. It should be noted that the removable storage device 850 may store data, such as instructions and / or driving path models of method 600, or may be omitted.

[0072] Storage device 860 can be a storage device capable of storing program instructions and other data. For example, storage device 860 can be a hard disk drive (HDD), a solid-state drive (SSD), or some other type of non-volatile memory. Storage device 860 can, for example, store the instructions and / or driving path model of method 600.

[0073] Removable storage device 850 and storage device 860 can be coupled to processor 810 via system bus 810B. System bus 810B can be any type of bus system that enables processor 810 and optionally GPU 820 and automotive processing system 830 to communicate with other devices of automotive control unit 800. System bus can be, for example, a Peripheral Component Interconnect Fast (PCIe) bus or a Serial AT Attach (SATA) bus.

[0074] The cellular interface 870 can be any type of interface that enables the vehicle control unit 800 to communicate with other vehicles and / or roadside infrastructure via a cellular network (such as a 4G or 5G network), for example in V2V or V2X communication.

[0075] The GNSS interface 880 can be any type of interface that enables the vehicle control unit 800 to receive location data provided by satellite networks such as the Global Positioning System (GPS), the Global Navigation Satellite System (GLONASS), or Galileo.

[0076] The communication interface 890 enables the vehicle control unit 800 to interface with external devices directly or via a network, as shown by wiring that couples the communication interface 890 to the outside of the vehicle control unit 800. The communication interface 890 can, for example, enable the vehicle control unit 800 to be coupled to wired or wireless networks, such as Ethernet, Wi-Fi, Controller Area Network (CAN) bus, or any bus system suitable for communication within a vehicle. For example, the vehicle control unit 800 can be coupled to vehicle sensors 710 or any other data source within the vehicle 700 via the communication interface 890. The communication interface 890 may also include a USB port or a serial port to enable direct communication with external devices.

[0077] The vehicle control unit 800 can be integrated with the vehicle 700, for example, under the driver's cab, under the dashboard, or in the trunk of the vehicle 700.

[0078] It should be understood that the vehicle control unit 800 may include... Figure 8 The number of elements shown may vary depending on the actual implementation requirements of the vehicle control unit 800. Furthermore, the elements discussed above may be distributed across multiple (sub)units.

[0079] The present invention can be further illustrated by the following embodiments.

[0080] In the example, one method is configured to maintain a driving path model during vehicle operation. The driving path model is configured to provide environmental perception of the current segment of a road to one or more driving automation system features of the vehicle, and includes at least one set of multiple lane model segments. Each set of multiple lane model segments approximates a corresponding lane segment of the corresponding lane in the current segment of the road. Each set of multiple lane model segments includes a forward multiple lane model segment preceding the vehicle's current position and a rear multiple lane model segment following the vehicle's current position. The method includes: when the current position of a vehicle changes: deriving a curvature profile of a forward distance ahead of the vehicle's current position from one or more map datasets, the curvature profile indicating the curvature of at least one lane of a current segment of the road ahead of the vehicle; identifying segments with approximately constant curvature variation along the forward distance based on the curvature profile; maintaining each set of forward multiple lane model segments to include lane model segments based on the identified segments with approximately constant curvature variation; maintaining each set of rear multiple lane model segments by reassigning each set of forward multiple lane model segments to each corresponding rear multiple lane model segment based on the vehicle's current position; and providing a driving path model to one or more driving automation system features configured to perform lateral and / or longitudinal control of the vehicle based on the driving path model.

[0081] In the exemplary method, maintaining each group of multiple lane model segments ahead includes adding lane model segments to each group of multiple lane model segments ahead based on the vehicle's current position.

[0082] In the example method, adding lane model segments involves approximating each identified segment with an approximate constant curvature change using at least one corresponding nth-order Taylor approximation and a corresponding n+1th-order Lagrange remainder.

[0083] In the example method, adding a lane model segment may further include: if the corresponding n+1 order Lagrange remainder exceeds the approximation error threshold, then limiting the length of the lane model segment corresponding to the given approximate constant curvature change marker segment to a length where the corresponding n+1 order Lagrange remainder remains below the approximation error threshold, which is less than the length of the corresponding approximate constant curvature change marker segment, and using an additional nth order Taylor approximation to approximate the remainder of the length of the corresponding approximate constant curvature change marker segment.

[0084] In the example method, adding a lane model segment may further include: if the corresponding n+1th order Lagrange remainder exceeds the approximation error threshold, then using one or more additional corresponding nth order Taylor approximations and n+1th order Lagrange remainders to approximate a given segment with constant curvature variation, the number of additional corresponding nth order Taylor approximations depending on the number of n+1th order Lagrange remainders exceeding the approximation error threshold.

[0085] In the example method, adding a lane model segment may also include: if the corresponding n+1 order Lagrange remainder exceeds the approximation error threshold, reducing the curvature of the lane model segment corresponding to the given identified approximately constant curvature change segment to a curvature that causes the corresponding n+1 order Lagrange remainder to fall below the approximation error threshold.

[0086] In an exemplary method, maintaining each group of multiple lane model segments ahead may include: detecting curvature change deviations within one or more lane model segments of each group of multiple lane model segments based on vehicle sensor data, and if the curvature change deviation exceeds a curvature change deviation threshold, modifying each group of multiple lane model segments ahead by: changing the length of one or more lane model segments based on the detected curvature change deviation; and / or adding one or more additional lane model segments to each group of multiple lane model segments ahead based on the detected curvature change deviation.

[0087] In the example method, maintaining each group of multiple lane model segments ahead may also include: detecting obstacles in one or more lane model segments of each group of multiple lane model segments ahead based on vehicle sensor data, and modifying each group of multiple lane model segments ahead by determining that one or more lane model segments have approximately constant curvature changes and bypass obstacles.

[0088] In the example method, identifying a segment with approximately constant curvature change may include: comparing the curvature profile with a left-turn curvature change threshold and a right-turn curvature change threshold, and identifying the point of the curvature profile that exceeds one of the left-turn curvature change thresholds and the right-turn curvature change threshold as the end point of the previous segment with approximately constant curvature change and the corresponding start point of the subsequent segment with approximately constant curvature change.

[0089] In the example method, at least one set of multiple lane model segments may include a fixed number of lane model segments, and the example method may also include a fixed number of lane model segments distributed across at least one set of multiple lane model segments across each set of forward multiple lane model segments and each set of backward multiple lane model segments.

[0090] In the example method, maintaining each group of multiple lane model segments ahead may include limiting the number of lane model segments within each group of multiple lane model segments to cover distances shorter than the distance ahead if at least one group of multiple lane model segments exceeds a fixed number of lane model segments.

[0091] In an exemplary method, maintaining multiple lane model segments behind each group may include limiting the number of lane model segments within each group of multiple lane model segments if at least one or more of the lane model segments exceed a fixed number of lane model segments.

[0092] In the example method, each group of rear multiple lane model segments may include one or more lane model segments extending along a rear distance behind the vehicle's current position, and maintaining each group of rear multiple lane model segments may include removing lane model segments from each group of rear multiple lane model segments located beyond each rear distance.

[0093] In the example method, the driving path model can include the number of multiple lane model segments corresponding to the number of lanes in the current segment of the road.

[0094] In the example method, deriving the curvature profile may include determining the curvature of the connection between each adjacent set of map data points for at least one lane.

[0095] Example methods may also include determining, based on one or more map datasets and / or vehicle sensor data, at least one corresponding junction point to the start or end point of a new plurality of lane model segments, the new plurality of lane model segments corresponding to a lane that branches off from or merges with at least one lane in the current segment, wherein determining the junction point may be based on a segment with a constant curvature change that begins or ends on a lane model segment of at least one of at least a group of lane model segments that is closest to at least one lane branching off or merging lane in the current segment.

[0096] An exemplary vehicle control unit includes at least one processing unit and a memory coupled to the at least one processing unit and configured to store machine-readable instructions. The machine-readable instructions cause the at least one processing unit to maintain a driving path model during vehicle operation. The driving path model is configured to provide environmental perception of a current segment of a road to one or more driving automation system features of the vehicle, and includes at least one set of multiple lane model segments. Each set of multiple lane model segments approximates a corresponding lane segment of a corresponding lane in the current segment of the road. Each set of multiple lane model segments includes a forward multiple lane model segment preceding the vehicle's current position and a rear multiple lane model segment following the vehicle's current position. To maintain the driving path model, the machine-readable instructions also enable at least one processing unit, when the vehicle's current position changes: derives from one or more map datasets a curvature profile of the forward distance ahead of the vehicle's current position, the curvature profile indicating the curvature of at least one lane of the current segment of the road ahead of the vehicle; identifies segments with approximately constant curvature variations along the forward distance based on the curvature profile; maintains each set of forward multiple lane model segments to include lane model segments based on the identified segments with approximately constant curvature variations; maintains each set of rear multiple lane model segments by reassigning each set of forward multiple lane model segments to each corresponding rear multiple lane model segment based on the vehicle's current position; and provides the driving path model to one or more driving automation system features configured to perform lateral and / or longitudinal control of the vehicle based on the driving path model.

[0097] In an exemplary vehicle control unit, machine-readable instructions may also cause at least one processing unit to perform any of the exemplary methods described above.

[0098] The example vehicle includes the exemplary vehicle control unit described above.

[0099] The foregoing description has been provided to illustrate methods and systems for maintaining driving path models. It should be understood that this description is in no way intended to limit the scope of this disclosure to the precise embodiments discussed throughout the description. Rather, those skilled in the art will recognize that examples of this disclosure can be combined, modified, or condensed without departing from the scope of this disclosure as defined by the appended claims. List of reference numerals 110 Road 860 Storage Device 111 Oncoming lane 870 Cellular interface 112 This vehicle lane / driving vehicle lane 880 GNSS interface 140 Driving Path Model 890 Communication Interface 210 Road 211 Oncoming lane 212 Driver's Lane / Own Lane 220 Curvature Profile 230 Curvature variation profile 240 Driving Path Model 310 Road 320° curvature profile 330° curvature variation profile 340 Driving Path Model 410 Road 440 Driving Path Model 510 Road 5401 First Multiple Lane Model Segment 5402 Second multiple lane model segment 600 methods Methods and steps (610-670) 700 vehicles 710 Automotive Sensors 800B bus 810 CPU 820 GPU 820c connection 830 Car Processing System 840 memory 850 Removable Storage Device

Claims

1. A method (600) for maintaining a driving path model (140, 240, 340, 440) during operation of a vehicle (700), said driving path model (140, 240, 340, 440) being configured to provide environmental perception of a current segment of a road (110, 210, 310, 410, 510) to one or more driving automation system features of the vehicle (700), and comprising at least one set of multiple lane model segments (5401, 5402), each set of multiple lane model segments ( 5401, 5402) approximate the corresponding lane segments (111, 112, 211, 212) of the current segment of the road (110, 210, 310, 410, 510), wherein each group of multiple lane model segments (5401, 5402) includes multiple lane model segments in front of the current position of the vehicle (700) and multiple lane model segments behind the current position of the vehicle (700), the method including when the current position of the vehicle (700) changes: (610) Derive (220, 320) from one or more map datasets a curvature profile (220, 320) for the forward distance in front of the current position of the vehicle (700), the curvature profile (220, 320) indicating the curvature of at least one lane of the current segment of the road in front of the vehicle. Based on the curvature profile (220, 320), the segment (620) is identified as having an approximately constant curvature variation along the forward distance; Maintain (640) multiple lane model segments ahead of each group to include lane model segments based on the identified approximately constant curvature variation of the segments; By reallocating each group of multiple lane model segments ahead to each group of corresponding multiple lane model segments behind based on the current position of the vehicle (700), the multiple lane model segments behind each group are maintained (660). as well as The driving path model (140, 240, 340, 440) is provided to one or more driving automation system features, the one or more driving automation system features being configured to perform lateral and / or longitudinal control of the vehicle based on the driving path model.

2. The method (600) according to claim 1, wherein the maintenance (640) of each group of multiple lane model segments ahead comprises: Based on the current position of the vehicle, lane model segments are added (641) to each group of multiple lane model segments ahead.

3. The method (600) according to claim 2, wherein adding (641) lane model segments comprises: Each identified segment of approximately constant curvature variation is approximated (641a) using at least one corresponding nth-order Taylor approximation and the corresponding n+1th-order Lagrange remainder.

4. The method (600) according to claim 3, wherein adding (641) the lane model segment further comprises, if the corresponding n+1 order Lagrange remainder of (641b) exceeds the approximation error threshold: The length limit (641c) of the lane model segment corresponding to the given identified segment of approximately constant curvature change is such that the corresponding n+1 order Lagrange remainder is kept below the approximation error threshold, the length being less than the length of the corresponding identified segment of approximately constant curvature change; and The remainder of the length of the segment corresponding to the identified approximately constant curvature variation is approximated using another nth-order Taylor approximation (641d).

5. The method (600) according to any one of claims 3 to 4, wherein the addition (641) lane model segment further comprises: If the corresponding n+1 Lagrange remainder described in (641b) exceeds the approximation error threshold, then one or more additional corresponding n-order Taylor approximations and n+1 Lagrange remainders are used to approximate (641e) a given segment of constant curvature variation, the number of the additional corresponding n-order Taylor approximations depending on the number of n+1 Lagrange remainders that exceed the approximation error threshold.

6. The method (600) according to any one of claims 3 to 5, wherein the addition (641) lane model segment further comprises: If the corresponding n+1 order Lagrange remainder described in (641b) exceeds the approximation error threshold: The curvature of the lane model segment corresponding to the given identified approximately constant curvature change is reduced (641f) to a curvature that reduces the corresponding n+1 order Lagrange remainder to below the approximate error threshold.

7. The method (600) according to any one of the preceding claims, wherein the maintenance (640) of each group of forward multiple lane model segments comprises: Based on vehicle sensor data, the curvature change deviation is detected (642) within one or more lane model segments in each group of multiple lane model segments ahead, and If the curvature change deviation described in (643) exceeds the curvature change deviation threshold, then each group of multiple lane model segments ahead is modified in the following manner: The length of one or more lane model segments is changed (643a) based on the detected curvature change deviation; and / or Based on the detected curvature change deviation, one or more additional lane model segments are added (643b) to each group of multiple lane model segments ahead.

8. The method (600) according to any one of the preceding claims, wherein the maintenance (640) of each group of multiple lane model segments ahead further comprises: Based on vehicle sensor data, obstacles (644) are detected in one or more lane model segments among multiple lane model segments ahead in each group, and Modify (645) each group of multiple lane model segments ahead by determining one or more lane model segments that have approximately constant curvature changes and bypassing the obstacles.

9. The method (600) according to any one of the preceding claims, wherein the segment in which the identifier (620) approximately changes constant curvature comprises: The curvature profile is compared with the left-turn curvature change threshold (t). + ) and the right-turn curvature change threshold (t) - ) for comparison (621), and Determine (622) the curvature profile that exceeds the left-turn curvature change threshold (t). + ) and the right-turn curvature change threshold (t) - One of the points is used as the end point of the preceding segment of the approximately constant curvature change and the corresponding starting point of the subsequent segment of the approximately constant curvature change.

10. The method (600) according to any one of the preceding claims, wherein: The at least one set of multiple lane model segments includes a fixed number of lane model segments. The method further includes: distributing (630) the fixed number of lane model segments in the at least one group of multiple lane model segments across each group of forward multiple lane model segments and each group of rear multiple lane model segments.

11. The method (600) according to any one of the preceding claims, wherein: Each group of rear lane model segments includes one or more lane model segments extending along a rear distance behind the vehicle's current position, and The maintenance (660) of each group of rear multiple lane model segments includes: removing (661) lane model segments from each group of rear multiple lane model segments located at each rear distance.

12. The method (600) according to any one of the preceding claims further includes: Based on the one or more map datasets and / or vehicle sensor data, determine (650) a junction point corresponding to at least one of the start or end points of a new plurality of lane model segments, the new plurality of lane model segments corresponding to a lane that branches off from or merges with the at least one lane of the current segment, wherein determining the junction point is based on identifying a segment with a constant curvature change that begins or ends on the lane model segment closest to the lane that branches off from or merges with the at least one lane of the current segment, respectively.

13. A vehicle control unit (800), comprising: At least one processing unit (810, 820, 830). as well as The memory (840, 850, 860), coupled to the at least one processing unit (810, 820, 830) and configured to store machine-readable instructions, The machine-readable instructions therein cause the at least one processing unit (810, 820, 830) to maintain a driving path model (140, 240, 340, 440) during operation of the vehicle (700). The driving path model (140, 240, 340, 440) is configured to provide environmental perception of the current segment of the road (110, 210, 310, 410, 510) to one or more driving automation system features of the vehicle (700), and includes at least one set of multiple lane model segments (5401, 5402), each set of multiple lane model segments (5401, 5402) approximating the corresponding lane segment of the corresponding lane (111, 112, 211, 212) of the current segment of the road (110, 210, 310, 410, 510), wherein each set of multiple lane model segments includes a forward multiple lane model segment in front of the current position of the vehicle (700) and a rear multiple lane model segment behind the current position of the vehicle (700), and In order to maintain the driving path model (140, 240, 340, 440), the machine-readable instructions also cause the at least one processing unit (810, 820, 830) to: when the current position of the vehicle (700) changes: The curvature profiles (220, 320) for the forward distance in front of the current position of the vehicle (700) are derived from one or more map datasets, the curvature profiles indicating the curvature of at least one lane of the current segment of the road (110, 210, 310, 410, 510) in front of the vehicle (700). Based on the curvature profiles (220, 320), segments with approximately constant curvature variation along the forward distance are identified; Maintain multiple lane model segments ahead for each group to include lane model segments based on the identified approximately constant curvature variation; Each group of rear lane model segments is maintained by reallocating each group of forward lane model segments to the corresponding rear lane model segments based on the current position of the vehicle (700); and The driving path model (140, 240, 340, 440) is provided to one or more driving automation system features, the one or more driving automation system features being configured to perform lateral and / or longitudinal control of the vehicle (700) based on the driving path model (140, 240, 340, 440).

14. The vehicle control unit (800) of claim 13, wherein the machine-readable instructions further cause the at least one processing unit (810, 820, 830) to perform the method (600) of any one of claims 2 to 12.

15. A vehicle (700) comprising a vehicle control unit (800) according to any one of claims 13 and 14.

Images