Continuous reference line decision method and device, vehicle, and storage medium

Abstract

The present invention discloses a continuous reference line decision method and device, the method comprising determining semantic information of a reference line, the determined semantic information comprising a first semantic information group for intention decision and a second semantic information group for direction decision; determining a reference line switching intention based on the first semantic information group; selecting an optimal reference line based on the reference line switching intention and the second semantic information group of each reference line; and determining a target reference line based on the selected optimal reference line. The solution of the present invention uses the semantic information determined for the reference line to make a reference line decision, which can ensure the rationality and reliability of the decision result; and applies part of the semantic information to perform decision analysis at each stage of the decision, which not only makes the decision process clear, intuitive and explainable, but also facilitates the flexible adjustment of the evaluation parameters of the decision analysis, i.e., the semantic information, according to the needs, thereby improving the debuggability and scalability of the reference line cost evaluation.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a continuous reference line decision method, a continuous reference line decision device, a vehicle, and a storage medium. Background Art

[0002] With the development of artificial intelligence, autonomous or semi-autonomous vehicle driving technologies are gaining widespread application. In the field of autonomous driving, reference lines are an important reference and basis for path planning for autonomous vehicles. In this field, reference lines typically refer to the centerline of a lane or road.

[0003] In real-world scenarios, due to the complexity of structured roads, autonomous vehicles often need to adjust their current lane in real time based on the driving task and surrounding environment. Therefore, lane selection and decision-making have become crucial aspects of path planning. Currently, lane selection and decision-making are typically based on reference lines. Therefore, lane selection and decision-making during driving are essentially decisions and choices about the target reference line. Therefore, the reliability of target reference line selection significantly impacts the safety, comfort, and efficiency of path planning. However, the complexity and variability of the driving environment, the uncertainty of upper-level perception, and the differences in passenger subjective experience all pose significant challenges to target reference line selection. Therefore, designing a safe, robust, and robust target reference line selection method has become an urgent need that needs to be addressed in the industry. Summary of the Invention

[0004] The embodiment of the present invention provides a continuous reference line decision solution to solve the problem in the prior art that it is difficult to safely and reasonably select a target reference line when performing path planning.

[0005] In a first aspect, an embodiment of the present invention provides a continuous reference line decision method, the method comprising:

[0006] Determining semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision-making and a second semantic information group for direction decision-making;

[0007] determining a reference line switching intention according to the first semantic information group;

[0008] Selecting an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line;

[0009] The target reference line is determined based on the selected optimal reference line.

[0010] In a second aspect, an embodiment of the present invention provides a continuous reference line decision device, the device comprising:

[0011] a semantic generation module, configured to determine semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision-making and a second semantic information group for direction decision-making;

[0012] an intention decision module, configured to determine a reference line switching intention based on the first semantic information group;

[0013] a direction decision module, configured to select an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line; and

[0014] The target decision module is used to determine the target reference line according to the selected optimal reference line.

[0015] In a third aspect, an embodiment of the present invention provides another continuous reference line decision device, comprising:

[0016] a memory for storing executable instructions; and

[0017] A processor is configured to execute executable instructions stored in a memory, wherein the executable instructions implement the steps of the method of any embodiment of the present invention when executed by the processor.

[0018] In a fourth aspect, an embodiment of the present invention provides a vehicle, comprising:

[0019] A decision maker, configured to make a target reference line decision according to any one of the continuous reference line decision methods according to the embodiments of the present invention; and

[0020] A controller is used to control the vehicle according to the target reference line determined by the decision maker.

[0021] In a fifth aspect, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method of any embodiment of the present invention.

[0022] The beneficial effects of the embodiments of the present invention are: the method provided by the embodiments of the present invention realizes the evaluation of the status of the reference line from different dimensions by determining multiple groups of semantic information of the reference line, and the method provided by the embodiments of the present invention divides the reference line decision into three stages: intention decision, direction decision and switching target decision, and applies part of the semantic information to perform decision analysis in each stage of the decision, which not only makes the decision process clear, intuitive and explainable, but also can facilitate the flexible adjustment of the evaluation parameters of the decision analysis, i.e., the semantic information, according to needs, thereby improving the debuggability and scalability of the reference line cost evaluation, and thus effectively ensuring the consistency of the decision results; in addition, the method of the embodiments of the present invention uses the semantic information determined for the reference line to make a reference line decision, and can also intuitively reflect the influence of the reference line status of different dimensions on the decision results in the decision process, so that the decision results are explainable and the rationality and reliability of the decision results are guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flow chart of a continuous reference line decision method according to one embodiment of the present invention;

[0025] Figure 2 A schematic diagram showing the specific scenario effect of determining the first semantic option of the reference line based on the high-precision map and perception information;

[0026] Figure 3 A specific processing flow chart for determining the second semantic option of a reference line based on a high-precision map and perception information is schematically shown;

[0027] Figure 4 The following diagram schematically shows a specific scenario effect of determining the second semantic option of the reference line based on the high-precision map and perception information, wherein: Figure 4 A shows the relationship map between the vehicle and the perceived obstacles under structured road conditions. Figure 4 B shows the pair Figure 4 Reference line ST diagram for construction of the environmental map shown in A;

[0028] Figure 5 A schematic diagram showing the specific scenario effect of the third semantic option for determining the reference line based on the high-precision map and perception information;

[0029] Figure 6Schematically shows a flow chart of a continuous reference line decision method according to another embodiment of the present invention;

[0030] Figure 7 This is a principle block diagram of a continuous reference line decision device according to one embodiment of the present invention;

[0031] Figure 8 This is a principle block diagram of a continuous reference line decision device according to another embodiment of the present invention;

[0032] Figure 9 This is a principle block diagram of an autonomous driving vehicle according to one embodiment of the present invention;

[0033] Figure 10 FIG. 4 is a structural diagram of an embodiment of a continuous reference line determination device of the present invention. DETAILED DESCRIPTION

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0035] It should be noted that, unless there is any conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0036] The present invention may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0037] In the present invention, "module", "device", "system" and the like refer to related entities applied to a computer, such as hardware, a combination of hardware and software, software or software in execution, etc. Specifically, for example, an element can be, but is not limited to, a process running on a processor, a processor, an object, an executable element, an execution thread, a program and / or a computer. In addition, an application or script program running on a server, or a server can all be an element. One or more elements can be in an execution process and / or thread, and an element can be localized on a computer and / or distributed between two or more computers, and can be run by various computer-readable media. An element can also communicate through local and / or remote processes based on a signal having one or more data packets, for example, a signal from a data packet interacting with another element in a local system, a distributed system, and / or a signal from a network on the Internet that interacts with other systems via signals.

[0038] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include" and "comprise" include not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or device. In the absence of further limitations, the elements defined by the phrase "include..." do not exclude the presence of other identical elements in the process, method, article or device that includes the elements.

[0039] The continuous reference line decision method in the embodiments of the present invention can be applied to a continuous reference line decision device, allowing a user or autonomous driving device to use the continuous reference line decision device to obtain a target reference line, thereby controlling the autonomous driving device to plan a path or execute corresponding driving actions based on the target reference line. These continuous reference line decision devices include, but are not limited to, a planner or decision maker on an autonomous driving vehicle, a smart tablet, a personal computer, a computer, a cloud server, and the like. In particular, the continuous reference line decision method in the embodiments of the present invention can also be directly applied to autonomous driving devices such as autonomous vehicles, which is not limited by the present invention.

[0040] Figure 1The continuous reference line decision method according to one embodiment of the present invention is schematically shown. The execution subject of the method can be a planner, controller or decision maker on an autonomous driving vehicle, or a processor of a continuous reference line decision device such as a smart tablet, a personal PC, a computer, a cloud server, or a processor of an autonomous driving device such as an unmanned cleaning vehicle, an unmanned sweeping vehicle, a sweeping robot, an autonomous driving vehicle, etc. The embodiment of the present invention is not limited to this. The embodiment of the present invention is first described in detail by taking the execution subject as a decision maker on an autonomous driving vehicle as an example. Figure 1 As shown, the method of the embodiment of the present invention includes:

[0041] Step S10: determining semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision, a second semantic information group for direction decision, and a third semantic information group for feasibility decision;

[0042] Step S11: determining a reference line switching intention according to the first semantic information group;

[0043] Step S12: selecting an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line;

[0044] Step S13: determining a target reference line according to the third semantic information group of the selected optimal reference line.

[0045] Since the selection of the target reference line requires consideration of many factors, such as whether it belongs to the global path, whether it is reachable, whether it has a speed advantage, etc., the embodiment of the present invention preferably considers converting the factors affecting the selection of the target reference line into semantic information, so that the decision of the target reference line can be made based on the determined semantic information, ensuring the consistency and rationality of the decision result. Among them, in the embodiment of the present invention, the converted semantic information is continuous information with clear physical meaning, which is preferably reflected by a continuous and quantifiable cost value attribute. Continuity means that the cost value used to represent its physical meaning is a continuous quantity rather than a discrete quantity.

[0046] In step S10, the semantic information of the reference line can be flexibly determined based on the influencing factors of the reference line, as long as it is possible to characterize the state of the reference line within a period of time from multiple dimensions so that a reasonable reference line can be selected based on the state. Exemplarily, the semantic information determined in the embodiment of the present invention preferably includes three types, namely, semantic information for intention decision-making, semantic information for direction decision-making, and semantic information for feasibility decision-making, which are respectively referred to as the first group of semantic information, the second group of semantic information, and the third group of semantic information in the embodiment of the present invention. Among them, when determining the semantic information of the reference line, it can be determined specifically based on the high-precision map and perception information obtained in real time. Since the influencing factors of the reference line can basically be determined and obtained through the high-precision map and perception information, determining the semantic information of the reference line based on the acquired high-precision map and perception information can also ensure the real-time, authenticity, and validity of the semantic information.

[0047] Specifically, semantic information for a reference line can be constructed based on predetermined factors influencing the selection of the reference line. This semantic information can be constructed based on the physical meaning of the semantic information. Subsequently, associated data information is extracted from the high-precision map and perception information based on the constructed semantic information, and the semantic information for the reference line is determined based on the extracted data information. The following text will explain how to determine the semantic information for a reference line based on the high-precision map and perception information, combining the specific factors determined and the physical meaning of the defined semantic information.

[0048] After constructing and determining the semantic information of the reference line, the embodiment of the present invention implements the decision-making of continuous reference lines through three stages, steps S11 to S13. Specifically, in step S11, the determined partial semantic information will be applied to make an intention decision; in step S12, when there is an intention to switch the reference line, the determined partial semantic information will be applied to make a direction decision to screen out the optimal reference line; and in step S13, the determined partial semantic information will be applied to make a feasibility decision on the optimal reference line determined based on the direction decision. Through these three stages of evaluation and decision-making from different angles, the consistency of the decision results can be guaranteed and the decision results can be made more reasonable.

[0049] It should be noted that, in a specific implementation, depending on the specific content of the semantic information of the reference line determined in step S10, the semantic information selected can also be adaptively adjusted when making decisions in the three stages of steps S11 to S13. For example, different partial semantic information can be selected in each of the three stages to make the corresponding decision. Alternatively, partially overlapping or partially different partial semantic information can be selected in two stages to make the corresponding decision. In certain circumstances, the same partial semantic information can also be used in two stages to make the corresponding decision. The embodiments of the present invention do not restrict the content and overlap of the partial semantic information used in each specific decision. As long as the embodiments of the present invention can be implemented by constructing the influencing factors of the reference line as the semantic information of the reference line and applying this semantic information in three stages (intention, direction, and feasibility decision) to achieve continuous reference line decision-making, thereby determining a more reasonable, comfortable, and robust target reference line, it is sufficient. In this embodiment of the present invention, continuous reference line decision-making means that the reference line semantic information used for reference line decision-making and the corresponding reference line cost value calculated therefrom are continuous quantities rather than discrete quantities.

[0050] As a preferred embodiment, the partial semantic information applied in the intention decision stage, i.e., the first semantic information group, of the embodiment of the present invention partially overlaps with the partial semantic information applied in the direction decision stage, i.e., the second semantic information group, while the partial semantic information applied in the feasibility decision stage, i.e., the third semantic information group, does not overlap with either the first semantic information group or the second semantic information group, that is, completely different semantic information is applied. In a specific implementation, depending on the content of the determined semantic information, the first semantic information group, the second semantic information group, and the third semantic information group of the application may each include at least one semantic option.

[0051] As a more preferred embodiment, the embodiment of the present invention is based on the consideration of the influencing factors of the reference line, and the specific content of the semantic information determined is implemented as including: a first semantic option for measuring the distance between the reference line and the global path, a second semantic option for describing the congestion degree of the reference line, a third semantic option for measuring the difficulty of reaching the reference line from the current position of the vehicle, a fourth semantic option for describing the impact of the vehicle reaching the reference line on the driving smoothness, and a fifth semantic option for describing the collision risk when the vehicle reaches the reference line.

[0052] Preferably, the physical meaning of each semantic option can be set to be represented by a cost value, so that each semantic option included in the semantic information actually represents a cost value with a specific meaning. Exemplarily, based on the specific meaning represented by each semantic option, each semantic option can be defined as follows: the first semantic option is defined as being positively correlated with the urgency of lane change, the second semantic option is defined as being positively correlated with the congestion level of the lane, the third semantic option is defined as being positively correlated with the difficulty of the vehicle reaching the reference line from the current position, the fourth semantic option is defined as being negatively correlated with the driving smoothness of the vehicle reaching the reference line, and the fifth semantic option is defined as being positively correlated with the collision risk of the vehicle reaching the reference line.

[0053] After the influencing factors to be considered and their physical meanings are determined in advance, the semantic information of each reference line can be dynamically and real-time determined based on the acquired high-precision map and perception information, wherein each reference line refers to all reference lines to be selected as target reference lines. In an embodiment of the present invention, the semantic information of all selectable reference lines is determined separately, and a target reference line that can be switched is determined based on the semantic information of each selectable reference line. The target reference line refers to the reference line of the target lane that can be switched to.

[0054] For example, taking the case where the first semantic option is defined as the ratio of the ideal lane change length to the remaining average lane change distance, the first semantic option of each reference line can be calculated using the following formula:

[0055] cost dev =S / (s / num) (Formula 1)

[0056] Among them, cost dev Represents the first semantic option, S represents the ideal lane change length, s represents the remaining length of the reference line, num represents the number of lane changes required to reach the global path from the reference line, and s / num represents the remaining average lane change distance. In a specific implementation, the ideal lane change length can be set and adjusted as needed. For example, it can be set or adjusted according to the influence weight of the global path on the target reference line. When the global path is used as an important decision factor or influencing factor of the target reference line, the ideal lane change length can be set to a larger value. Otherwise, the ideal lane change length can be adjusted to a smaller value. Preferably, the ideal lane change length of an embodiment of the present invention can be set to 100 meters. In other preferred embodiments, the ideal lane change length can also be calculated by designing a function with parameters such as the vehicle speed as the independent variables of the design function to obtain the value of the ideal lane change length.

[0057] Therefore, Figure 2As an example of the scenario shown in FIG. 1 , the first semantic option of determining the reference line based on the high-precision map and the perception information can be implemented as follows: Figure 2 As shown, it can be determined through high-precision maps and perception information that there are currently three lanes, namely the first lane 1, the second lane 2 and the third lane 3. The reference lines of each lane are the first reference line, the second reference line, and the third reference line. At the same time, the information that the vehicle needs to turn right at the intersection can also be determined based on the perception information. Therefore, based on this information extracted from the high-precision map and perception information, combined with the physical meaning of the first semantic option, it can be calculated that in order to turn right at the intersection, two lane changes (num=2) are required to the right from the first reference line to achieve the global path goal, only one lane change (num=1) is required to the right from the second reference line to achieve the global path goal, and no lane change (num=0) is required from the third reference line to achieve the global path goal; at the same time, based on the perception information and high-precision map, the remaining length of each reference line from the right turn intersection can also be extracted, assuming it is Figure 2 As shown in FIG1 , the ideal lane change length is 100 meters (s=100). Therefore, based on the information extracted from the high-precision map and the perception information and combined with Formula 1, it can be determined that the first semantic option of the first reference line is 2, the first semantic option of the second reference line is 1, and the first semantic option of the third reference line is 0.

[0058] Preferably, the embodiment of the present invention describes the degree of congestion of each reference line by considering the impact of dynamic and static obstacles on the passable space of the own vehicle. Specifically, the second semantic option can be defined as being positively correlated with the impact of the obstacle on the passable space of the own vehicle, that is, the greater the impact of the obstacle on the passable space of the own vehicle, the larger the second semantic option. Thus, the second semantic option can be designed to be positively correlated with the size of the space occupied by the obstacle on the current reference line. Exemplarily, taking the second semantic option as the ratio of the space occupied by the obstacle to the passable space of the own vehicle as an example, wherein the space occupied by the obstacle is the size of the space occupied by the obstacle on the current reference line, which can be represented by the projection area and congestion coefficient of the obstacle on the ST diagram (i.e., the displacement time diagram) corresponding to the reference line, and the passable space of the own vehicle is the space that the own vehicle can occupy within the planning time when traveling at the expected speed, which can be represented by the projection area of ​​the spatial area in the ST diagram. Thus, the second semantic option of each reference line is specifically calculated by the following formula:

[0059] cost con =(area*α) / Th_area (Formula 2)

[0060] Among them, cost conIndicates the second semantic option, Th_area represents the projected area of ​​the traversable space of the ego vehicle on the ST diagram, area represents the projected area of ​​the obstacle on the ST diagram corresponding to the reference line, and α represents the congestion coefficient, which is calculated by the following formula:

[0061] α=max(1,10 w-d )

[0062] Where d is the maximum distance from the obstacle to the reference line boundary, and w is the ego vehicle's width. Since the maximum distance d from the obstacle to the reference line boundary is greater than the ego vehicle's width w when the vehicle can travel normally along the reference line, as the lateral distance of the obstacle from the reference line increases, the maximum distance d from the obstacle to the reference line boundary decreases accordingly, narrowing the road width available for the ego vehicle. When the maximum distance d from the obstacle to the reference line boundary decreases to equal the ego vehicle's width w, the distance between the obstacle and the boundary becomes insufficient for the ego vehicle to pass, maximizing the ego vehicle's sense of urgency to change lanes. This is when the obstacle's impact on the ego vehicle's traversable space reaches its maximum. Conversely, as the maximum distance d from the obstacle to the reference line boundary increases, the impact of the obstacle on the ego vehicle's traversable space decreases accordingly. Therefore, the congestion coefficient defined by the above formula accurately reflects the impact of the obstacle on the ego vehicle's traversable space. When combined with the obstacle's projected area on the ST diagram corresponding to the reference line, it more accurately reflects the space occupied by the obstacle and, consequently, the degree of congestion along the reference line.

[0063] Therefore, in a specific implementation, Figure 3 As shown, the second semantic option of determining the reference line based on the high-precision map and perception information can be implemented as including:

[0064] Step S30: Calculate the congestion coefficient of each obstacle on the reference line based on the high-precision map and perception information.

[0065] Step S31: All obstacles on the reference line that may conflict with the reference line are determined based on the high-precision map and perception information, and the determined obstacles are projected onto the ST map corresponding to the reference line according to preset rules. With reference to the definition of the second semantic option above, in this embodiment of the present invention, projecting the determined obstacles onto the ST map corresponding to the reference line according to preset rules means projecting the obstacles onto the ST map corresponding to the reference line based on the final projected area obtained by vertically projecting the obstacles onto the ST map corresponding to the reference line and multiplying the congestion coefficient by the original area.

[0066] Step S32: Determine the spatial area that the vehicle can occupy within the planning time when traveling at the desired speed based on the perception information and high-precision information, and project the spatial area onto the ST diagram of the reference line.

[0067] Step S33: Calculate the area of ​​the obstacle-occupied space on the reference line and the passable space for the vehicle based on the ST diagram of the reference line and the congestion coefficient of each obstacle.

[0068] Step S34: Determine the second semantic option based on the area of ​​the space occupied by the obstacle and the traversable space of the vehicle.

[0069] The following will be Figure 4 For example, Figure 3 The processing flow shown is described in detail.

[0070] like Figure 4 As shown in A, in this scene, through the high-precision map and perception information, it can be learned that there are three lanes in the current scene, namely the first lane 1, the second lane 2 and the third lane 3, and there is a dynamic obstacle B in front of the ego vehicle A traveling along the reference line of the second lane 2, and there is a static obstacle C in front of the left of the ego vehicle A occupying part of the space of the reference line of the second lane 2. Based on the data of the high-precision map and perception information, the reference line ST graph can be modeled through step S31. For example, taking the reference line of the second lane 2 as an example, with the starting point of the ego vehicle at the reference line as the origin, the driving time T of the ego vehicle along the reference line as the horizontal axis, i.e., the T axis, and the vertical axis, i.e., the S axis, the ST graph corresponding to the reference line is constructed in the frenet coordinate system, and the situation where the obstacle conflicts with the reference line at time t is projected at (t, s) in the ST graph to obtain the ST graph corresponding to the reference line as shown in FIG. Figure 4 As shown in B, since the dynamic obstacle B continuously occupies the passable space in front of the ego vehicle A, its projection on the ST diagram of the reference line of the second lane 2 is the upper parallelogram D. Although the static obstacle C does not completely block the passage of the ego vehicle A, the ego vehicle A will still lose a certain speed when passing through, so it also needs to be projected on the ST diagram of the reference line. Its projection on the ST diagram is the lower rectangle E. Afterwards, the driving speed and obstacle avoidance path of the ego vehicle A within the planning time can be planned based on the high-precision map and perception information. Therefore, the passable space of the ego vehicle within the planning time can be determined based on this and projected on the ST diagram of the reference line. For example, the maximum displacement at each moment can be determined based on the speed limit value of the ego vehicle, which is the upper limit of the S axis. The upper limit of the S axis and the planning time together constitute the passable space. Projecting it on the ST diagram of the reference line can be obtained as shown below. Figure 4The dotted line in B contains area F, where the upper boundary of area F represents the distance traveled by the ego vehicle at time t, the lower boundary of area F represents the time corresponding to the upper limit of the S axis, and the slope of area F is the speed of the ego vehicle. The area F defined by the upper boundary, lower boundary, and slope is the spatial area that the ego vehicle can occupy within the planned time if it travels at the desired speed. This can achieve the projection of the ego vehicle's traversable spatial area in step S31.

[0071] In practical applications, as a preferred embodiment, in order to reflect the different degrees of encroachment of each obstacle on the reference line, a blocking coefficient may be further added to the projected area of ​​the obstacle on the reference line ST diagram as needed. For example, the blocking coefficient is set to be obtained by the following calculation formula:

[0072] α=max(1,10 w-d )(Formula 3)

[0073] In step S32, Figure 4 Using the high-precision map and perception information acquired in the scenario of vehicle A, the maximum distance d from the obstacle to the boundary of lane 2 and the width w of the ego vehicle are determined. The congestion coefficient for each obstacle can then be calculated using the above formula. The physical meaning of the congestion coefficient defined in accordance with the present invention is readily apparent. In this embodiment, the maximum distance d from the obstacle to the reference line boundary is greater than the ego vehicle width w during normal driving. As the lateral distance of the obstacle from the reference line increases, the maximum distance d decreases, and the congestion coefficient increases. When the distance d between the obstacle and the lane boundary is insufficient for the ego vehicle to pass, the congestion coefficient is 1.

[0074] According to the physical meaning of the aforementioned second semantic option, in step S33, the space occupied by the obstacle can be determined by calculating the product of the projected area of ​​the obstacle in the reference line ST diagram and the congestion coefficient. And by calculating the area of ​​the projection area in the reference line ST diagram that the vehicle can occupy when traveling at the expected speed within the planning time, the passable space of the vehicle can be determined. Then, in step S34, the second semantic option can be obtained by calculating the ratio of the space occupied by the obstacle to the passable space of the vehicle. It should be noted that when there are multiple obstacles, the sum of the occupied spaces of the various obstacles can be used as the obstacle occupied space of the reference line, such as Figure 4 In the scenario shown, the occupied spaces of the dynamic obstacle B and the static obstacle C can be calculated separately, and the sum of their occupied spaces is used as the obstacle occupied space of the reference line of the second lane 2, and the second semantic option is determined based on this.

[0075] Therefore, in the embodiment of the present invention, the above method is used to realize the traffic efficiency of the reference line using the ratio of the obstacle projection space to the traversable space of the vehicle, and the space is determined based on the ST diagram modeling of the reference line. Therefore, what is actually obtained is the spatiotemporal occupancy rate of the obstacle, that is, the embodiment of the present invention realizes the traffic efficiency of the reference line using the spatiotemporal occupancy rate of the obstacle. The second semantic option determined accordingly fully considers the impact of dynamic and static obstacles on the traversable space, and can therefore accurately and reasonably describe the congestion degree of the reference line, ensuring the rationality of the consideration of the influencing factors of the reference line selection.

[0076] Preferably, the third semantic option is defined as the number of lane changes required for the vehicle to reach the reference line from its current position. Specifically, the third semantic option is calculated using the following formula:

[0077]

[0078] Among them, cost arr For the third semantic option, m i is the minimum number of lane changes required for the vehicle to reach the reference line i from its current position.

[0079] In a specific implementation, the high-precision map and perception information can be used to determine whether the vehicle can reach reference line i from its current position. If it is determined that the vehicle can reach reference line i, the third semantic option is determined to be the minimum number of lane changes required for the vehicle to reach the reference line from its current position. Otherwise, the third semantic option is determined to be infinite. This allows the determination of the third semantic option. For example, the feasibility and specific number of lane changes are primarily affected by static obstacles and virtual and real lines. Therefore, using existing graph search or sampling algorithms based on the acquired high-precision map and perception information can roughly determine the feasibility and minimum number of lane changes required for the vehicle to reach each reference line. Figure 5 The embodiment of determining the third semantic option in a specific scenario is schematically shown, as shown in FIG. Figure 5 As shown, based on the high-precision map and perception information, it can be determined that the ego vehicle A is traveling in the second lane 2 and there is a static obstacle C in the first lane 1. Therefore, using the existing graph search or sampling algorithm, the ego vehicle cannot search for a collision-free path with the reference line of the first lane 1, indicating that the ego vehicle cannot reach the reference line of the first lane from its current position. Therefore, the third semantic option of the reference line of the first lane is determined to be infinity. Correspondingly, the ego vehicle searches for the reference line of the second lane 2 and needs to borrow a lane, that is, it needs to change lanes twice to reach it. Therefore, the third semantic option of the reference line of the second lane is determined to be 2. Similarly, since the search shows that the ego vehicle needs to change lanes to the right once to reach the reference line of the third lane 3, the third semantic option of the reference line of the third lane is determined to be 1.

[0080] Thus, embodiments of the present invention can determine the third semantic option for the reference line using the aforementioned method. Because the determination of the third semantic option takes into account whether the vehicle can reach the reference line and the difficulty of reaching the reference line, reasonable reference line decisions can be made based on the third semantic option. For example, when the third semantic option indicates increased difficulty, lane keeping may be preferred, thereby avoiding the drawbacks and disadvantages of frequently changing reference lines due to fluctuations in the external environment.

[0081] Preferably, the fourth semantic option is defined as being characterized by the deviation between the current driving state of the ego vehicle and the stable lane keeping state, where the stable lane keeping state refers to the ideal state of maintaining stable driving in the lane. Specifically, the current driving state of the ego vehicle and the stable lane keeping state can be characterized by the lateral distance of the ego vehicle from the reference line, the deviation of the ego vehicle's heading from the reference line, and the time interval since the last reference line change. Therefore, the fourth semantic option of each reference line can be calculated using the following formula:

[0082] cost comf =l / L+angle / Angle+t / T (Formula 5)

[0083] Among them, cost comf Indicates the fourth semantic option, l represents the lateral distance between the vehicle and the reference line, angle represents the deviation between the vehicle's heading and the reference line angle, t represents the time since the last reference line was changed, L represents the ideal distance deviation, Angle represents the ideal angle deviation, and T represents the ideal time interval. In specific applications, the ideal distance deviation, angle deviation and time interval can be flexibly set and adjusted according to requirements (such as passenger comfort). For example, the ideal distance deviation can be set to 0.5 meters, the ideal angle deviation can be set to 5 degrees, and the ideal time interval can be set to 5 seconds. In other preferred embodiments, the ideal distance deviation, angle deviation and time interval can also be calculated by designing a function, using parameters such as the vehicle speed as the independent variables of the design function to obtain the preset values ​​of these ideal parameters. It should be noted that in the embodiment of the present invention, the values ​​of the ideal distance deviation, angle deviation and time interval can reflect the state of the vehicle when starting to change lanes. Therefore, they can be adjusted according to the requirements of ride comfort to avoid the discomfort caused by continuous lane changes or large-angle lane changes through adjustment.

[0084] As a preferred embodiment, the lateral distance of the vehicle from the reference line, the deviation of the vehicle's heading from the reference line angle, and the time since the last reference line change can be obtained based on the perception information of the perception input. These can all be obtained using the front-end perception module (including but not limited to cameras, laser radars, ultrasonic radars, etc.), so they will not be repeated here. After obtaining the corresponding information, the fourth semantic option of the reference line can be calculated and determined based on the physical meaning of the fourth semantic option. Therefore, when determining the semantic information of the reference line, the embodiment of the present invention can determine the fourth semantic option by considering the distance item, the angle item, and the time item. By setting the distance item and the angle item, the tracking state of the vehicle when changing the reference line can be improved. By setting the time item, continuous lane changes and multiple obstacle avoidance behaviors can be avoided as much as possible. Therefore, continuous reference line decision-making based on this can achieve the goal of avoiding large maneuvers as much as possible when the tracking error is large, thereby effectively ensuring the stability of control and ride comfort.

[0085] Preferably, the fifth semantic option is defined as being characterized by the collision risk between the ego vehicle and obstacles in front of and behind the ego vehicle on the reference line. The collision risk can be described based on the gap theory. The specific content of the gap theory can be referenced in the prior art and will not be elaborated here. For example, the fifth semantic option of each reference line is calculated using the following formula:

[0086] cost safe =∑(dis / Dis+ttc / TTC) (Formula 6)

[0087] Among them, cost safe Indicates the fifth semantic option, dis indicates the absolute distance between the vehicle and the obstacle, ttc indicates the collision time between the vehicle and the obstacle, Dis indicates the ideal longitudinal distance, and TTC indicates the ideal collision time. In specific applications, the ideal longitudinal distance and collision time reflect the radicalness of the decision. Therefore, the ideal longitudinal distance and collision time can be set and adjusted according to the demand for the radicalness of the decision. For example, when the radicalness of the decision is high and speed is preferred, the ideal longitudinal distance and ideal collision time can be set to smaller values. When the radicalness of the decision is low and safety is preferred, the ideal longitudinal distance and ideal collision time can be set to larger values. In a preferred embodiment of the present invention, the ideal longitudinal distance is set to 10 meters and the ideal collision time is set to 6 seconds. In other preferred embodiments, the ideal longitudinal distance and collision time can also be calculated by designing a function with parameters such as the vehicle speed as the independent variables of the design function to obtain the values ​​of the ideal longitudinal distance and collision time.

[0088] As a preferred embodiment, the absolute distance between the vehicle and the obstacle and the collision time between the vehicle and the obstacle can be obtained based on the perception information of the perception input. These can be obtained using the front-end perception module and / or planning module, so they will not be repeated here. After obtaining the corresponding information, the fifth semantic option of the reference line can be calculated and determined based on the physical meaning of the fifth semantic option. Therefore, when determining the semantic information of the reference line, the embodiment of the present invention can determine the fifth semantic option by setting the distance item and the collision time item, and realize the fifth semantic option by selecting the nearest obstacles in front of and behind the vehicle on the reference line, and calculating and summing the distance item and the collision time item. Based on this, when there is no corresponding obstacle on the reference line, there is no need to calculate its cost, that is, in the obstacle-free state, there is no need to calculate the fifth semantic option on the reference line. In determining the fifth semantic option, the embodiment of the present invention enables the ego vehicle to fully consider the size of the space when switching the reference line to execute a lane change by setting a distance item, and enables the ego vehicle to fully consider the rationality of time when switching the reference line to execute a lane change by setting a collision time item. Therefore, the determined fifth semantic option, by considering the costs of distance and collision time, enables the ego vehicle to seek a larger space and time gap to execute lane changes, thereby improving the rationality and comfort of reference line decisions.

[0089] Thus, embodiments of the present invention can predetermine semantic information, including deviation degree, driving efficiency, execution difficulty, stability, and safety, based on factors such as whether a route belongs to a global path, whether it is reachable, and whether it has a speed advantage. Furthermore, embodiments of the present invention define cost value attributes for this semantic information to represent its physical meaning. This defines semantic information for semantic options, including a first semantic option for representing deviation degree, a second semantic option for representing driving efficiency, a third semantic option for representing execution difficulty, a fourth semantic option for representing stability, and a fifth semantic option for representing safety. This allows embodiments of the present invention to construct and determine semantic options for reference lines based on acquired high-precision maps and perception information. This allows for quantitative evaluation of semantic information when associated with specific influencing factors. Consequently, reference line decisions and selections based on this information are more interpretable, reasonable, and consistent, thereby ensuring driving comfort and safety. Furthermore, based on the physical meaning of each semantic information discussed above, it can be seen that the calculated cost values ​​for each reference line are continuous rather than discrete quantities, thus ensuring continuity in the resulting reference line decisions and ensuring their rationality.

[0090] Considering that in actual applications, the change of the target reference line is usually caused by forced lane change or free lane change, for example, if the current reference line deviates from the global path or it is difficult to reach the expected speed, it is necessary to reselect the target reference line and change lanes. Therefore, as a preferred embodiment, the switching intention decision can be made based on the degree of deviation and traffic efficiency. Exemplarily, in an embodiment of the present invention, the first semantic option for measuring the distance between the reference line and the global path and the second semantic option for describing the congestion degree of the reference line are determined as semantic information for intention decision-making, that is, the first semantic information group is set to include the first semantic option and the second semantic option, so as to make a decision on the reference line switching intention based on the first semantic option and the second semantic option.

[0091] Taking the first semantic information group as an example, which is set to include the first semantic option and the second semantic option, the determination of the reference line switching intention based on the first semantic information group in the above step S11 can be implemented as determining the reference line switching intention based on the comparison result of the first semantic option and the preset threshold; or it can be implemented as determining the reference line switching intention based on the comparison result of the second semantic option and the preset threshold; of course, it can also be implemented as determining the reference line switching intention based on the comparison result of the first semantic option and the preset threshold, and the comparison result of the second semantic option and the preset threshold at the same time.

[0092] Taking the physical meanings of the first semantic option and the second semantic option determined in the previous text as an example, in a preferred embodiment, the switching intention of the reference line can be determined based on the comparison result of the first semantic option and the preset threshold value and / or the comparison result of the second semantic option and the preset threshold value. Specifically, since the first semantic option reflects the degree of deviation, and the degree of deviation reflects the urgency of forced lane change, therefore, according to the physical meaning of the first semantic option, when the remaining lane change distance is less than the ideal lane change distance, there will be a need to return to the global path as soon as possible. Therefore, a preset threshold value can be set for the first semantic option. For example, the preset threshold value of the first semantic option can be set to 1, and the determined first semantic option can be compared with the preset threshold value 1. When the first semantic option is greater than 1, the switching intention of the reference line is determined as the intention to switch the reference line. Otherwise, the switching intention of the reference line is determined as the absence of the switching intention. Since the second semantic option reflects driving efficiency and the degree of reference line encroachment, according to the physical meaning of the second semantic option, if the current reference line is encroached upon by a large amount of space, an intention to switch the reference line will also be generated. Therefore, a preset threshold can be set for the second semantic option. For example, the preset threshold for the second semantic option can be set to 0.2. When the second semantic option is determined, the second semantic option is compared with the preset threshold of 0.2. If the second semantic option is determined to be less than 0.2, the intention to switch the reference line is determined to exist; otherwise, the intention to switch the reference line is determined to exist. It should be noted that since the threshold of the first semantic option is a deviation threshold and the threshold of the second semantic option is a driving efficiency threshold, they represent and reflect the degree of aggressiveness of forced lane changes and free lane changes, respectively, and can be adjusted and set according to the desired driving style. The embodiments of the present invention do not limit their specific setting values. For example, the threshold of the second semantic option represents the proportion of the reference line encroached on the S axis and can be adjusted according to the driving style, increasing for aggressiveness and decreasing for conservativeness.

[0093] Preferably, the selection of the optimal reference line according to the reference line switching intention and the second semantic information group of each reference line in the above step S12 can be realized as follows: when there is an intention to switch the reference line, the total cost value of each reference line is calculated according to the second semantic information group of each reference line, and the optimal reference line is selected according to the total cost value. The optimal reference line refers to the current optimal reference line determined according to the calculated total cost value of each reference line. Since it is necessary to comprehensively consider the distance from the global path, the driving efficiency and the cost of reaching the reference line when evaluating the reference line, the embodiment of the present invention preferably sets the second semantic information group to include a first semantic option for measuring the degree of deviation of each reference line, a second semantic option for measuring the driving efficiency of each reference line and a third semantic option for measuring the execution difficulty of each reference line, so as to comprehensively evaluate the cost of each reference line based on these three, so that the direction intention decision can accurately reflect the rationality and comfort of the reference line, and further ensure the rationality and consistency of the reference line decision result at the direction decision level.

[0094] Taking the physical meanings of the first semantic option, the second semantic option, and the third semantic option determined in the previous text as an example, since the semantic information of the first semantic option, the second semantic option, and the third semantic option are dimensionless and of similar order of magnitude, in a preferred embodiment, the total cost value of each reference line calculated according to the second semantic information group of each reference line can be implemented by directly summing up the semantic options of the second semantic information group, and taking the sum of the semantic options of the second semantic information group as the total cost value of the reference line. And selecting the optimal reference line according to the total cost value can be implemented by selecting the reference line with the lowest cost value as the optimal reference line, thereby achieving a directional decision. In an embodiment of the present invention, by reasonably setting the semantic options in the second semantic information group and calculating the total cost value of the reference line by summing up the semantic options in combination with the physical meaning of each semantic option, not only the degree of influence of each important influencing factor on the reference line selection is comprehensively considered, but also the implementation method is simple, so that the directional decision of the reference line is reasonable and explainable, so that the reference line decision result can ensure consistency.

[0095] In other preferred embodiments, based on the differences in the degree of attention paid to various influencing factors of the reference line and the differences in factors considered when making decisions, when calculating the total cost of the reference line, weights can be set for each semantic option in the second semantic information group as needed, that is, weighted calculations can be performed on the first semantic option, the second semantic option, and the third semantic option to enhance the influence of a certain factor on the decision-making result.

[0096] As a preferred embodiment, after the current optimal reference line is determined through the above steps, the timing of the switch can be further judged, that is, the feasibility and rationality of changing the reference line at the current moment can be further judged, so as to determine the target reference line based on the feasibility and rationality. Since in actual scenarios, when preparing to change lanes, the main focus is on whether there is enough clearance in the target lane and whether the current state of the vehicle is stable, the fourth and fifth semantic options can be used to judge whether the safety and stability of the reference line meet the requirements, thereby achieving the determination of the target reference line based on feasibility and rationality. Based on this, in a preferred embodiment of the present invention, the third semantic information group can be set to include the fourth semantic option and the fifth semantic option.

[0097] Taking the third semantic information group as an example, which is set to include the fourth semantic option and the fifth semantic option, the determination of the target reference line according to the third semantic information group of the selected optimal reference line in the above step S13 can be implemented as determining the target reference line according to the comparison result of the fourth semantic option of the selected optimal reference line and the preset threshold; or it can be implemented as determining the target reference line according to the comparison result of the fifth semantic option of the selected optimal reference line and the preset threshold; of course, it can also be implemented as determining the target reference line according to the comparison result of the fourth semantic option of the selected optimal reference line and the preset threshold, and the comparison result of the fifth semantic option of the selected optimal reference line and the preset threshold at the same time.

[0098] Taking the physical meanings of the fourth semantic option and the fifth semantic option determined in the previous text as an example, in a preferred embodiment, the target reference line can be determined based on the comparison results of the fourth semantic option and the preset threshold and the comparison results of the fifth semantic option and the preset threshold. Specifically, it can be implemented as follows: when it is judged that the fourth semantic option is lower than the preset threshold and the fifth semantic option is lower than the preset threshold, the selected optimal reference line is directly changed to a new target reference line; otherwise, it indicates that it is not suitable for lane changing at present (neither feasibility nor rationality is good enough), and the target reference line is temporarily not changed, that is, the target reference line continues to be determined as the reference line currently being traveled. Exemplarily, the preset thresholds of the fourth semantic option and the fifth semantic option can be set according to engineering experience data.

[0099] In a preferred embodiment, when it is determined based on the third semantic information group that it is not currently suitable to change lanes, a lane change assistance operation may be further performed to prepare for or provide a prompt for the lane change. For example, the lane change assistance operation performed may be to first light up the turn signal in the target direction based on the selected optimal reference line, to continue to perform the judgment of step S13 to seek a better lane change opportunity, and when it is determined in step S13 that it is suitable to change lanes, to immediately perform the operation of changing the target reference line.

[0100] In other embodiments, the rationality and feasibility judgment process of the above step S13 may not be performed, but the target reference line may be directly determined based on the selected optimal reference line.

[0101] Figure 6 The following schematically shows the process of the continuous reference line decision method according to another embodiment. Figure 6 As shown, its implementation includes:

[0102] Step S60: determining semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision and a second semantic information group for direction decision;

[0103] Step S61: determining a reference line switching intention according to the first semantic information group;

[0104] Step S62: selecting an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line;

[0105] Step S63: determining a target reference line according to the selected optimal reference line.

[0106] The specific implementation of steps S60 to S62 of the embodiment of the present invention can refer to the above description. Figure 1 The difference from the illustrated embodiment is that the embodiment of the present invention only determines the first semantic information group and the second semantic information group in step S60. After step S62, that is, after the current optimal reference line is selected, the target reference line is determined directly based on the selected optimal reference line.

[0107] In a specific implementation, for example, in step S63, a judgment can be made based on the selected optimal reference line, such as whether the vehicle needs to change lanes to reach the target reference line from the current lane. If the judgment is made based on the third semantic option and it is determined that the target reference line can be reached without changing lanes, the current optimal reference line is directly used as the target reference line.

[0108] For example, in step S63, after selecting the current optimal reference line, the target reference line can be directly changed, i.e., the target reference line can be directly determined as the selected current optimal reference line. Thus, after determining the current optimal reference line in step S62, lane changes can be performed based on the determined optimal reference line, i.e., the determined optimal reference line is used as the target reference line, and the vehicle is directly switched to the lane containing the target reference line. The target reference line refers to the determined reference line of the lane to which the vehicle is to be switched.

[0109] It should be noted that when determining whether lane change is required based on the selected optimal reference line and determining the target reference line and executing the lane change operation based on the judgment result, step S63 and the above-mentioned Figure 1 Step S13 of the illustrated embodiment is combined to form a new and more optimal embodiment. The new and more optimal embodiment may be, for example: after selecting the current optimal reference line, determine whether the vehicle needs to change lanes to reach the optimal reference line from the current lane. If it is determined that the optimal reference line can be reached without changing lanes, then the current optimal reference line is directly used as the target reference line, that is, the target reference line is changed to the optimal reference line; and when it is determined that the optimal reference line can be reached only after changing lanes, the rationality and feasibility judgment of step S13 is executed, and it is determined whether to change the target reference line to the current optimal reference line based on the rationality and feasibility judgment results of step S13.

[0110] In other embodiments, when making the rationality and feasibility judgment of step S13, the fourth semantic option and the fifth semantic option may not be compared with the preset thresholds respectively to make a feasibility decision. Instead, certain parameters selected from the physical meaning of the fourth semantic option or the fifth semantic option may be used as constraints for separate judgment.

[0111] Therefore, the embodiment of the present invention can realize reference line decision-making based on the semantic information of the reference line, and the cost value obtained based on the semantic information is smooth and continuous, which can ensure the consistency of the decision result. Moreover, since the embodiment of the present invention divides the reference line decision into multiple stages, each stage uses part of the semantic information for simple comparison, and the decision-making process is clear, intuitive and explainable. In addition, since the parameters selected by the embodiment of the present invention have clear physical meanings and value patterns when determining the semantic information, the parameters can be adjusted quickly and flexibly according to needs to achieve flexible semantic information determination based on the influencing factors of the reference line, so as to ensure the rationality and compliance of the decision result.

[0112] Figure 7 The continuous reference line decision device according to one embodiment of the present invention is schematically shown. Figure 7 As shown, the device includes

[0113] a semantic generation module 70 for determining semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision and a second semantic information group for direction decision;

[0114] An intention decision module 71 is configured to determine a reference line switching intention based on the first semantic information group;

[0115] a direction decision module 72, configured to select an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line; and

[0116] The target decision module 73 is used to determine the target reference line according to the selected optimal reference line.

[0117] As a preferred embodiment, there are at least some identical semantic options in the first semantic information group and the second semantic information group.

[0118] In a preferred embodiment, the first semantic information group includes a first semantic option for measuring the distance between the reference line and the global path and a second semantic option for measuring the congestion level of the reference line; the second semantic information group includes a first semantic option for measuring the distance between the reference line and the global path, a second semantic option for measuring the congestion level of the reference line, and a third semantic option for measuring the difficulty of reaching the reference line from the current position of the vehicle.

[0119] Among them, preferably, the first semantic option and the second semantic option are both defined as being positively correlated with the urgency of lane change. Exemplarily, the first semantic option is defined as the ratio of the ideal lane change length to the average remaining lane change distance. The second semantic option is defined as being positively correlated with the impact of the obstacle on the traversable space of the ego vehicle, that is, positively correlated with the size of the space occupied by the obstacle on the current reference line. Specifically, the second semantic option is defined as the ratio of the space occupied by the obstacle to the traversable space of the ego vehicle, wherein the space occupied by the obstacle is the size of the space occupied by the obstacle on the current reference line, and the traversable space of the ego vehicle is the space that the ego vehicle can occupy within the planning time when traveling at the desired speed.

[0120] In another preferred embodiment, the determined semantic information further includes a third semantic information group for feasibility decision, and the target decision module is specifically configured to determine the target reference line according to the third semantic information group of the selected optimal reference line.

[0121] In a preferred embodiment, the third semantic information group includes a fourth semantic option for measuring the impact of the ego vehicle's arrival at the reference line on driving smoothness, and a fifth semantic option for measuring the collision risk associated with the ego vehicle's arrival at the reference line. For example, the fourth semantic option is defined as being characterized by the ego vehicle's tracking state, lane change frequency, and obstacle avoidance frequency when changing reference lines; the fifth semantic option is defined as being characterized by the state of obstacles ahead and behind the ego vehicle on the reference line.

[0122] In another preferred embodiment, the target decision module is specifically configured to determine the target reference line according to the second semantic information group and the third semantic information group of the selected optimal reference line.

[0123] In other preferred embodiments, the target decision module may also be configured to directly determine the target reference line according to the selected optimal reference line.

[0124] It should be noted that the specific implementation process of each module of the continuous reference line decision device and the definition of each semantic information in the embodiment of the present invention can refer to the description of the method part above, so they will not be repeated here.

[0125] Figure 8 The continuous reference line decision device according to another embodiment of the present invention is schematically shown. Figure 8 As shown, its implementation includes:

[0126] Memory 80, for storing executable instructions; and

[0127] The processor 81 is configured to execute executable instructions stored in the memory, wherein the executable instructions, when executed by the processor, implement the steps of the continuous reference line determination method described in any one of the aforementioned embodiments.

[0128] In specific practice, the above-mentioned continuous reference line decision device can be used, for example, on autonomous driving equipment such as autonomous driving vehicles, unmanned cleaners, unmanned sweepers, robots, etc., to select and switch the reference lines of these devices to improve the comfort of unmanned driving. Specifically, the above-mentioned continuous reference line decision device can be implemented as a decision maker, controller, or planner on autonomous driving equipment such as autonomous driving vehicles.

[0129] Figure 9 A vehicle according to an embodiment of the present invention is schematically shown. Figure 9 As shown, the vehicle includes:

[0130] A decision maker 90, configured to determine a target reference line according to the method of any of the above embodiments; and

[0131] The controller 91 is used to control the vehicle according to the target reference line determined by the decision maker 90, such as performing direction control, path planning, turn signal control, etc.

[0132] In other embodiments, the decision maker and controller in the vehicle may be integrated into a single control module. Alternatively, in practical applications, the vehicle may also include a perception and recognition module and other planning and control modules, such as a path planning controller and a bottom-level controller. For example, the vehicle may be an autonomous vehicle or a semi-autonomous vehicle, which is not limited in this embodiment of the present invention.

[0133] In some embodiments, an embodiment of the present invention provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the continuous reference line decision method of any of the above embodiments of the present invention.

[0134] In some embodiments, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the continuous reference line decision method of any one of the above embodiments.

[0135] In some embodiments, an embodiment of the present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the continuous reference line decision method of any of the above embodiments.

[0136] In some embodiments, an embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the continuous reference line decision method of any of the above embodiments.

[0137] Figure 10 FIG. 1 is a schematic diagram of the hardware structure of a continuous reference line decision device according to another embodiment of the present invention. The continuous reference line decision device can be implemented with the structure shown in the figure. Figure 10 As shown, the continuous reference line decision device includes:

[0138] One or more processors 610 and memory 620, Figure 10 A processor 610 is taken as an example.

[0139] The continuous reference line decision device may further include: an input device 630 and an output device 640 .

[0140] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 10 The bus connection is taken as an example.

[0141] Memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the continuous reference line determination method in the embodiments of the present application. Processor 610 executes the non-volatile software programs, instructions, and modules stored in memory 620 to execute various server functional applications and data processing, thereby implementing the continuous reference line determination method in the above-described method embodiment.

[0142] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of the path planning method, etc. In addition, the memory 620 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely located relative to the processor 610, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0143] The input device 630 may receive input digital or character information and generate signals related to user settings and function control of the image processing device. The output device 640 may include a display device such as a display screen.

[0144] The one or more modules are stored in the memory 620 , and when executed by the one or more processors 610 , perform the continuous reference line determination method in any of the above method embodiments.

[0145] The above-mentioned product can execute the method provided in the embodiment of this application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of this application.

[0146] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0147] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and are primarily designed to provide voice and data communications. These terminals include smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones.

[0148] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, have computing and processing capabilities, and generally also have mobile Internet access. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0149] (3) Portable entertainment devices: These devices can display and play multimedia content. These devices include audio and video players (such as iPods), handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0150] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0151] (5) Other electronic devices with data interaction functions.

[0152] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0153] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, or of course, by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the relevant technology, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiment.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. Continuous reference line decision method, characterized in that, The method comprises: Determining semantic information of the reference line, wherein the determined semantic information includes a first semantic information group for intention decision-making and a second semantic information group for direction decision-making, wherein the first semantic information group includes a first semantic option for measuring the distance between the reference line and the global path and a second semantic option for measuring the congestion degree of the reference line; determining a reference line switching intention according to the first semantic information group; Selecting an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line; The target reference line is determined based on the selected optimal reference line.

2. The method according to claim 1, characterized in that There are at least some identical semantic options in the first semantic information group and the second semantic information group.

3. The method according to claim 2, characterized in that Determining the reference line switching intention according to the first semantic information group includes: Determining the switching intention of the reference line based on a comparison result of the first semantic option and a preset threshold; and / or The switching intention of the reference line is determined according to a comparison result of the second semantic option and a preset threshold.

4. The method according to claim 3, characterized in that The first semantic option is defined as being positively correlated with the urgency of lane change, and the second semantic option is defined as being positively correlated with the congestion level of the lane.

5. The method according to claim 4, characterized in that The first semantic option is defined as a ratio of an ideal lane change length to a remaining average lane change distance.

6. The method according to claim 5, characterized in that The first semantic option is calculated by the following formula: cost dev =S / (s / num) Among them, cost dev Indicates the first semantic option, S represents the ideal lane change length, s represents the remaining length of the current reference line, num represents the number of lane changes required to reach the global path from the current reference line, and s / num represents the remaining average lane change distance.

7. The method according to claim 4, characterized in that The second semantic option is defined as being positively correlated with the impact of the obstacle on the traversable space of the ego vehicle.

8. The method according to claim 7, characterized in that The second semantic option is defined as the ratio of the space occupied by the obstacle to the space passable by the vehicle, wherein the space occupied by the obstacle is the size of the space encroached by the obstacle on the reference line, which is represented by the projection area of ​​the obstacle on the ST diagram corresponding to the reference line and the congestion coefficient; the space passable by the vehicle is the space that the vehicle can occupy within the planning time when traveling at the expected speed, which is represented by the projection area of ​​the space on the ST diagram corresponding to the reference line.

9. The method according to claim 8, characterized in that The second semantic option is calculated by the following formula: cost con =(area*α) / Th_area Among them, cost con Indicates the second semantic option, Th_area represents the projection area of ​​the traversable space of the vehicle on the ST graph corresponding to the reference line, area represents the projection area of ​​the obstacle on the ST graph corresponding to the reference line, and α represents the congestion coefficient, which is calculated by the following formula: α=max(1.10 w-d ) Where d is the maximum distance from the obstacle to the reference line boundary, and w is the width of the vehicle.

10. The method according to any one of claims 1 to 9, characterized in that The selecting the optimal reference line according to the reference line switching intention and the second semantic information group of each reference line includes: When it is determined that the reference line switching intention is to switch the reference line, calculating the total cost value of each reference line according to the second semantic information group of each reference line; An optimal reference line is selected according to the total cost value.

11. The method according to claim 10, characterized in that The second semantic information group includes a first semantic option for measuring the distance between the reference line and the global path, a second semantic option for measuring the congestion degree of the reference line, and a third semantic option for measuring the difficulty of reaching the reference line from the current position of the vehicle; Calculating the total cost value of each reference line according to the second semantic information group of each reference line includes: A total cost value of each reference line is calculated according to the first semantic option, the second semantic option, and the third semantic option of each reference line.

12. The method according to claim 11, characterized in that The third semantic option is defined as being positively correlated with the difficulty of the vehicle reaching the reference line from the current position.

13. The method according to claim 12, characterized in that The third semantic option is defined to be characterized by the number of lane changes required for the ego vehicle to reach the reference line from its current position.

14. The method according to claim 12, characterized in that The third semantic option is calculated by the following formula: Among them, cost arr For the third semantic option, m i is the minimum number of lane changes required for the vehicle to reach the reference line i from its current position.

15. The method according to claim 10, characterized in that The determined semantic information also includes a third semantic information group for feasibility decision-making, wherein the third semantic information group includes a fourth semantic option for measuring the impact of the vehicle reaching the reference line on driving smoothness and a fifth semantic option for measuring the collision risk existing when the vehicle reaches the reference line; then, The determining the target reference line according to the selected optimal reference line includes: determining the target reference line according to the third semantic information group of the selected optimal reference line.

16. The method according to claim 15, characterized in that Determine the target reference line according to the third semantic information group of the selected optimal reference line, including The target reference line is determined according to a comparison result of the fourth semantic option and a preset threshold, and / or a comparison result of the fifth semantic option and the preset threshold.

17. The method according to claim 15, characterized in that The fourth semantic option is defined as being negatively correlated with the driving smoothness of the ego vehicle reaching the reference line.

18. The method according to claim 17, characterized in that The fourth semantic option is defined as being characterized by a deviation between the current driving state of the vehicle and a stable lane keeping state.

19. The method according to claim 18, characterized in that The fourth semantic option is calculated by the following formula: cost comf =l / L+angle / Angle+t / T Among them, cost comf Indicates the fourth semantic option, l represents the lateral distance of the ego vehicle from the reference line, angle represents the angular deviation between the ego vehicle heading and the reference line, t represents the time since the last reference line change, L represents the ideal distance deviation, Angle represents the ideal angle deviation, and T represents the ideal time interval.

20. The method according to claim 15, wherein The fifth semantic option is defined as being positively correlated with the collision risk existing when the ego vehicle reaches the reference line.

21. The method according to claim 20, characterized in that The fifth semantic option is defined as representing the collision risk between the ego vehicle and obstacles in front of and behind the ego vehicle on the reference line.

22. The method according to claim 21, characterized in that The fifth semantic option is calculated by the following formula: cost safe =∑(dis / Dis+ttc / TTC) Among them, cost safe Indicates the fifth semantic option, dis represents the absolute distance between the ego vehicle and the obstacle, ttc represents the collision time between the ego vehicle and the obstacle, Dis represents the ideal longitudinal distance, and TTC represents the ideal collision time.

23. A continuous reference line decision device, characterized in that: The device includes a semantic generation module, configured to determine semantic information of a reference line, wherein the determined semantic information includes a first semantic information group for intention decision-making and a second semantic information group for direction decision-making, wherein the first semantic information group includes a first semantic option for measuring the distance of the reference line from the global path and a second semantic option for measuring the degree of congestion of the reference line; an intention decision module, configured to determine a reference line switching intention based on the first semantic information group; a direction decision module, configured to select an optimal reference line according to the reference line switching intention and the second semantic information group of each reference line; and The target decision module is used to determine the target reference line according to the selected optimal reference line.

24. The continuous reference line decision device according to claim 23, characterized in that: The determined semantic information also includes a third semantic information group for feasibility decision; The third semantic information group includes a fourth semantic option for describing the impact of the vehicle reaching the reference line on driving stability and a fifth semantic option for describing the collision risk existing when the vehicle reaches the reference line; The target decision module is specifically configured to determine a target reference line according to the third semantic information group of the selected optimal reference line.

25. A continuous reference line decision device, characterized in that: include: a memory for storing executable instructions; as well as A processor, configured to execute executable instructions stored in a memory, wherein the executable instructions implement the steps of the method according to any one of claims 1 to 22 when executed by the processor.

26. A vehicle, characterized in that The vehicle comprises: A decision maker, configured to determine a target reference line according to the method according to any one of claims 1 to 22; and A controller is used to control the vehicle according to the target reference line determined by the decision maker.

27. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 22 are implemented.

Images