An obstacle detour decision method and device of an autonomous vehicle and the autonomous vehicle

Abstract

The embodiment of the specification discloses a kind of obstacle bypass decision method, device and automatic driving vehicle of automatic driving vehicle;Wherein method includes: based on the sampling density of determined first sampling point, the first sampling layer is sampled;Total cost of all sampling points of the first sampling layer is traversed and calculated, and first alternative point is determined in the first sampling layer according to preset rule, to update the first sampling point as the first alternative point;Based on the first sampling point updated, the sampling density of next sampling layer is determined and sampling is carried out, until reaching preset sampling termination condition;According to the first sampling point determined by multiple sampling layers, first trajectory line is determined, and according to the first trajectory line, the planning result of bypass target obstacle is obtained.The present application scheme is based on the degree of judgment obstacle density, corresponding adjustment sampling density, combined with optimization sampling point cost category design, effectively takes into account bypass decision ability and calculation complexity.

Description

Technical Field

[0001] This specification relates to the field of autonomous vehicles, and more particularly to an obstacle avoidance decision-making method and device for autonomous vehicles, as well as autonomous vehicles, and corresponding electronic devices and storage media. Background Technology

[0002] During autonomous vehicle operation, various static or dynamic obstacles frequently obstruct the driving lane, hindering normal travel. In such situations, the vehicle needs to automatically calculate detour plans to determine whether and how to detour, significantly impacting the vehicle's trafficability and the rationality of its autonomous driving behavior. Current obstacle detour decision-making methods include sampling-based assessments; however, these methods suffer from slow computation speeds, fixed sampling densities, or failure to consider obstacle distribution density. Therefore, overcoming the shortcomings of existing obstacle detour decision-making methods and proposing an obstacle detour decision-making scheme for autonomous vehicles that balances detour accuracy and computational complexity is a pressing technical problem. Summary of the Invention

[0003] The purpose of the embodiments in this specification is to address the above-mentioned problems by providing an obstacle avoidance decision-making method, device, autonomous vehicle, and corresponding electronic devices and storage media for autonomous vehicles.

[0004] To solve the above-mentioned technical problems, the embodiments in this specification are implemented as follows:

[0005] Firstly, a method for obstacle avoidance decision-making in autonomous vehicles is proposed, including:

[0006] Based on the determined sampling density of the first sampling point, the first sampling layer is sampled;

[0007] The total cost of all sampling points in the first sampling layer is calculated by traversing the layer, and a first candidate point is determined in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point.

[0008] Based on the updated first sampling point, the sampling density of the next sampling layer is determined and sampling is performed until the preset sampling termination condition is reached.

[0009] Based on multiple first sampling points determined by multiple sampling layers, a first trajectory line is determined, and based on the first trajectory line, a detour decision result for the target obstacle is obtained.

[0010] Further, based on the density of target obstacles within a first distance threshold from the first sampling point, the sampling density of the first sampling point is determined, including:

[0011] If the number of target obstacles within a preset distance from the first sampling point is greater than a preset number, the sampling density is increased;

[0012] If the number of target obstacles within a preset distance from the first sampling point is less than a preset number, or if the first sampling point is located within the outer boundary of the target obstacle, the sampling density is reduced.

[0013] Furthermore, the total cost includes the sum of a first cost and a second cost; wherein the first cost includes one or more costs characterizing the relationship between the current sampling point and the relevant parameters of the existing planning decision results, and the second cost includes one or more prediction costs characterizing the relationship between the current sampling point and the endpoint.

[0014] Further, the first cost includes one or more of the following: distance cost, reference line cost, boundary cost, static obstacle cost, dynamic obstacle cost, and similarity cost; wherein, the distance cost is used to characterize the cost of the distance between the current sampling point and the sampling point of the previous sampling layer; the reference line cost is used to characterize the cost of the lateral displacement between the current sampling point and the reference line; the boundary cost is used to characterize the cost of the distance between the autonomous vehicle and the lane boundary line; the static obstacle cost is used to characterize the cost of the distance between the current sampling point and static obstacles within a preset range; and the similarity cost is used to characterize the cost of the distance between the current sampling point and the sampling point at the same location as the previous planning decision result.

[0015] Furthermore, the second cost includes one or the sum of two costs: the endpoint distance cost and the obstacle compensation cost; wherein, the endpoint distance cost is used to characterize the predicted cost of the distance from the current sampling point to the endpoint; and the obstacle compensation cost is used to characterize the predicted cost of compensation when the obstacle cost cannot be obtained within a preset range behind the obstacle.

[0016] Furthermore, a first candidate point is determined in the first sampling layer according to a preset rule, including selecting the sampling point with the minimum total cost in the first sampling layer as the first candidate point.

[0017] Furthermore, before sampling the first sampling layer based on the determined sampling density, the current position of the autonomous vehicle is obtained as the initial first sampling point.

[0018] Furthermore, the step of stopping sampling until a preset sampling termination condition is reached includes: terminating sampling when the distance between the second projection point of the updated first sampling point on the reference line and the first projection point of the initial first sampling point on the reference line along the s direction in the Frenet coordinate system is greater than or equal to a second distance threshold.

[0019] Further, obtaining the detour decision result for the target obstacle based on the first trajectory line includes:

[0020] If the outer boundary of the target obstacle is located to the left of the envelope of the autonomous vehicle formed by the first trajectory line, then the decision to detour around the target obstacle to the right is output.

[0021] If the outer boundary of the target obstacle is located to the right of the envelope of the autonomous vehicle formed by the first trajectory line, then the decision to detour around the target obstacle to the left is output.

[0022] If the outer boundary of the target obstacle intersects with the envelope of the autonomous vehicle formed by the first trajectory line, then the decision result of not detouring around the target obstacle is output.

[0023] Further, based on multiple first sampling points determined by multiple sampling layers, a first trajectory line is determined, including: connecting multiple first sampling points to obtain the first trajectory line.

[0024] Secondly, an obstacle avoidance decision-making device for autonomous vehicles is proposed, comprising:

[0025] The first module is capable of sampling the first sampling layer based on the sampling density of the determined first sampling points;

[0026] The second module is capable of traversing and calculating the total cost of all sampling points in the first sampling layer, and determining a first candidate point in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point;

[0027] The third module is able to determine the sampling density of the next sampling layer and perform sampling based on the updated first sampling point until the preset sampling termination condition is reached.

[0028] The fourth module is able to determine a first trajectory line based on multiple first sampling points determined by multiple sampling layers, and obtain a detour decision result for the target obstacle based on the first trajectory line.

[0029] Thirdly, an autonomous vehicle is proposed that executes the method described in the first aspect to obtain obstacle detour planning decisions for the autonomous vehicle.

[0030] Fourthly, an electronic device is proposed, comprising: a processor; and

[0031] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method described in the first aspect.

[0032] Fifthly, a computer-readable storage medium is provided that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the method described in the first aspect.

[0033] This manual can achieve at least the following technical effects:

[0034] The present invention is based on a sampling evaluation algorithm. When obstacles are dense, the sampling density is increased to improve the detour decision-making ability of the target vehicle. When there are no obstacles or obstacles are sparse, the sampling density is reduced to reduce the computational complexity. At the same time, combined with the design of optimized sampling point cost categories, the detour decision-making ability and computational complexity are effectively balanced. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a schematic diagram of the Frenet coordinate system provided for embodiments of this specification.

[0037] Figure 2 This is one of the flowcharts illustrating an obstacle avoidance decision-making method for an autonomous vehicle provided in the embodiments of this specification.

[0038] Figure 3 This is a schematic diagram illustrating sampling in Frenet coordinates for an autonomous vehicle, as provided in an embodiment of this specification.

[0039] Figure 4 This is the second schematic diagram of an obstacle avoidance decision-making method for an autonomous vehicle provided in the embodiments of this specification.

[0040] Figure 5 This is the third schematic diagram of an obstacle avoidance decision-making method for an autonomous vehicle provided in the embodiments of this specification.

[0041] Figure 6 This is a schematic diagram of an obstacle avoidance decision-making device for an autonomous vehicle provided in an embodiment of this specification.

[0042] Figure 7 This is a schematic diagram of the structure of an electronic device provided as an embodiment of this specification. Detailed Implementation

[0043] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0044] The following detailed description of an obstacle avoidance decision-making scheme for an autonomous vehicle, as described in this specification, is illustrated through a specific example.

[0045] This invention focuses on overcoming the shortcomings of existing obstacle detour decision-making methods and proposes a technical solution for obstacle detour decision-making in autonomous vehicles that balances detour accuracy and computational complexity. During autonomous vehicle operation, various static or dynamic obstacles frequently obstruct the driving lane, hindering normal vehicle movement. In such cases, the vehicle needs to automatically calculate detour plans to determine whether and how to detour, significantly impacting the vehicle's trafficability and the rationality of its autonomous driving behavior. Currently, obstacle detour decision-making methods include sampling-based evaluation schemes and machine learning-based schemes. Sampling-based evaluation schemes include depth-first search, breadth-first search, and A / B search. * Search algorithms, etc. Among them, depth-first search and breadth-first search do not effectively guide search priority, resulting in a large amount of invalid search space and slow speed; A * The search algorithm uses a fixed sampling density, which makes it unable to provide a refined solution for detour decisions with dense obstacles, thus affecting the passability of the detour planning decision results.

[0046] Because autonomous vehicles typically use a Cartesian coordinate system to describe their position, this system is not optimal for the vehicle. Even if the vehicle's position is given, it doesn't describe the specific location within the driving lane, making it impossible to determine the vehicle's distance traveled and whether it has deviated from the lane center. Furthermore, since most driving lanes are not straight but curved, and autonomous vehicles only need to know their lane position and the lateral and longitudinal distances to surrounding obstacles to control themselves, a Cartesian coordinate system is usually chosen... Figure 1The detour planning solution is performed in the Frenet coordinate system shown. The Frenet coordinate system describes the vehicle's position relative to the driving lane. In the Frenet coordinate system, 's' represents the distance along the driving lane and is called the ordinate, and 'd' represents the displacement relative to the longitudinal line and is called the abscissa. This ensures that at each point in the driving lane, the abscissa and ordinate are perpendicular. The ordinate represents the distance traveled in the driving lane, and the abscissa represents the distance the vehicle deviates from the center line. In the Frenet coordinate system, the center line of the driving lane is used as the reference line, and a coordinate system is established using the tangent vector 't' and the normal vector 'n' of the reference line. With the vehicle itself as the origin, the coordinate axes are perpendicular to each other, divided into the 's' direction along the reference line and the 'd' direction, which is the current normal to the reference line. The Frenet coordinate system is suitable for route planning in autonomous driving curves, and it can also reduce the dimensionality of map data, thereby reducing the computational load. It simplifies the problem, making it easier to obtain parameters such as the center line of the driving lane reference line and the distances in the 's' and 'd' directions during vehicle movement. Therefore, the obstacle detour decision of the autonomous vehicle in this embodiment of the invention is described based on a sampling evaluation algorithm in the Frenet coordinate system.

[0047] Example 1

[0048] To effectively balance the detour decision-making capability and computational complexity of the sampling evaluation algorithm, this invention provides a variable-density sampling evaluation algorithm, while optimizing the cost type of sampling points in the algorithm. Therefore, as... Figure 2 The diagram shown is a flowchart illustrating an obstacle avoidance decision-making method for an autonomous vehicle according to an embodiment of the present invention. The method includes:

[0049] S1: Based on the determined sampling density of the first sampling point, sample the first sampling layer.

[0050] In some embodiments of the present invention, before sampling the first sampling layer based on a determined sampling density, the current position of the autonomous vehicle is obtained as an initial first sampling point.

[0051] In some embodiments of the present invention, determining the sampling density of the first sampling point based on the density of target obstacles within a first distance threshold from the first sampling point includes:

[0052] S111: If the number of target obstacles within a preset distance from the first sampling point is greater than a preset number, the sampling density is increased;

[0053] S112: If the number of target obstacles within a preset distance from the first sampling point is less than a preset number, or if the first sampling point is located within the outer boundary of the target obstacle, the sampling density is reduced.

[0054] Specifically, when the number of target obstacles within a preset distance from the first sampling point is greater than a preset number, it is determined that the area is in a dense obstacle space. In this case, the sampling density should be increased. For example, in some embodiments, a sampling point can be selected every 0.1m to improve the calculation accuracy of detour planning.

[0055] Specifically, when the number of target obstacles within a preset distance from the first sampling point is less than a preset number, or when the first sampling point is located within the outer boundary of the target obstacle, it is determined that the area is in a dense obstacle space. In this case, the sampling density should be reduced. For example, in some embodiments, a sampling point can be selected every 0.3m to improve the calculation efficiency of detour planning.

[0056] S2: Iterate through and calculate the total cost of all sampling points in the first sampling layer, and determine the first candidate point in the first sampling layer according to the preset rules, so as to update the first sampling point as the first candidate point.

[0057] In some embodiments of the present invention, the total cost includes the sum of a first cost and a second cost; wherein the first cost includes one or more costs characterizing the relationship between the current sampling point and the relevant parameters of the existing planning decision results, and the second cost includes one or more prediction costs characterizing the relationship between the current sampling point and the endpoint.

[0058] In some embodiments of the present invention, the first cost includes one or more of the following: distance cost, reference line cost, boundary cost, static obstacle cost, dynamic obstacle cost, and similarity cost.

[0059] In some embodiments of the present invention, the distance cost is used to characterize the cost of the distance between the current sampling point and the sampling point of the previous sampling layer.

[0060] Specifically, the distance cost is established based on the Euclidean distance between the current sampling point and the sampling point in the previous layer; the greater the distance, the greater the cost. In some embodiments, the distance cost g can be defined. cost_dist for:

[0061] g cost_dist =k dist_factor ×distance(cur_node, pre_node)

[0062] Where k dist_factor This represents the distance cost factor; distance(,) represents the Euclidean distance.

[0063] In some embodiments of the present invention, the reference line cost is used to characterize the cost of the lateral displacement between the current sampling point and the reference line.

[0064] Specifically, a greater deviation from the reference line means a greater distance from the road centerline. To address this, a reference line cost is incorporated to prevent vehicles from straying too far from the reference line when selecting sampling points. In some embodiments, a reference line cost g can be defined. cost_reference_line for:

[0065] g cost_reference_line =k reference_line_factor ×|cur_node_l-reference_line_l|

[0066] Where k reference_line_factor represents the reference line cost factor; || represents the distance in the s direction.

[0067] In some embodiments of the present invention, the boundary cost is used to characterize the cost of the distance between the autonomous vehicle and the lane boundary line.

[0068] Specifically, considering the additional risk that vehicles may exceed the lane boundaries while maneuvering, a boundary cost is considered to be added near the lane boundary line in this case. Assume the vehicle bounding box is a rectangle centered at the current sampling point, with the vehicle's length, width, and orientation information as its dimensions. In some embodiments, the distance from the left side of the vehicle bounding box to the left boundary is set to l_left; if the vehicle exceeds the left boundary of the lane, l_left is set to 0. Similarly, the distance from the right side of the vehicle bounding box to the right boundary is set to l_right; if the vehicle exceeds the right boundary of the lane, l_right is set to 0. The smaller of l_left and l_right is taken as the final boundary distance value: boundary_dist = min(l_left, l_right). In some embodiments, the boundary cost can be defined as:

[0069]

[0070] Where k boundary_factor This represents the boundary cost factor.

[0071] In some embodiments of the present invention, the static obstacle cost is used to characterize the cost of the distance between the current sampling point and a static obstacle within a preset range.

[0072] Specifically, the closer the sampling point is to the obstacle, the higher the cost of that sampling point should be. First, the distance between the current sampling point and the static obstacle is calculated. The minimum distance between the outer boundary of the obstacle and the vehicle bounding box extending from the current sampling point is used as `dist_static`. If the outer boundary of the obstacle intersects with the vehicle bounding box, then `dist_static` is 0. In some embodiments, the static obstacle cost can be defined as:

[0073]

[0074] Where k static_obsfactor This represents the static obstacle cost factor. It iterates through all static obstacles within a certain distance and takes the maximum value as the final static obstacle cost for that sampling point.

[0075] In some embodiments of the present invention, the cost of dynamic obstacles is less than that of static obstacles because the future position of dynamic obstacles is more difficult to predict.

[0076] In some embodiments of the present invention, the similarity cost is used to characterize the cost of the distance between the current sampling point and the sampling point at the same location as the previous planning decision result.

[0077] Specifically, to ensure stable obstacle avoidance decisions during vehicle operation and avoid abrupt changes in the avoidance planning results (i.e., to prevent significant discrepancies between the current sampling point and the planned point at the same location in the previous frame), the distance in the s-direction can be used to obtain the corresponding sampling point from the existing trajectory line planned in the previous frame. In some embodiments, the similarity cost can be defined as:

[0078]

[0079] Where k similar_factor This represents the similarity cost factor.

[0080] In some embodiments of the present invention, the second cost includes one or the sum of two costs: a destination distance cost and an obstacle compensation cost.

[0081] In some embodiments of the present invention, the endpoint distance cost is used to characterize the prediction cost of the distance from the current sampling point to the endpoint.

[0082] Specifically, the distance cost h can be defined. cost_dist for:

[0083] h cost_dist =k dist_h_factor ×distance(end_point,cur_node)

[0084] Where k dist_h_factor The endpoint distance cost factor is represented by `end_point`; `end_point` represents the endpoint.

[0085] In some embodiments of the present invention, the obstacle compensation cost is used to characterize the predicted cost of compensation when the obstacle cost cannot be obtained within a preset range behind the obstacle. Specifically, for cases where there is no obstacle at the sampling point, if the sampling point is located within a certain range behind a static obstacle, an additional term h should be added. cost_cost_staticobs To compensate for the lack of g in a certain space after the static obstacle. cost_cost_staticobs The limitations of this approach include, for example, when the obstacle is a thin wall, the space behind that wall is practically impassable. However, if it is determined that there are no static obstacles at the current sampling point, it will be difficult to bypass that thin wall. Similarly, the cost of dynamic obstacle compensation should also be considered, but will not be elaborated here.

[0086] In some embodiments of the present invention, a first candidate point is determined in the first sampling layer according to a preset rule, including selecting the sampling point with the minimum total cost in the first sampling layer as the first candidate point.

[0087] S3: Based on the updated first sampling point, determine the sampling density of the next sampling layer and perform sampling until the preset sampling termination condition is reached.

[0088] In some embodiments of the present invention, the step of stopping sampling until a preset sampling termination condition is reached includes: terminating sampling when the distance between the second projection point of the updated first sampling point on the reference line and the first projection point of the initial first sampling point on the reference line along the s direction in the Frenet coordinate system is greater than or equal to a second distance threshold.

[0089] Specifically, in some embodiments, a fixed interval can be selected as the second distance threshold along the s-direction of the Frenet coordinate system, for example, the second distance threshold can be selected as 1 meter. Therefore, as long as the distance between the second projection point of the updated first sampling point on the reference line and the first projection point of the initial first sampling point on the reference line is less than 1 meter, lateral sampling can be performed on the next sampling layer.

[0090] S4: Based on multiple first sampling points determined by multiple sampling layers, determine a first trajectory line, and based on the first trajectory line, obtain the detour decision result for the target obstacle.

[0091] In some embodiments of the present invention, obtaining the detour decision result of the target obstacle based on the first trajectory line includes:

[0092] S411: If the outer boundary of the target obstacle is located to the left of the envelope of the autonomous vehicle formed by the first trajectory line, then output the decision result of going right around the target obstacle.

[0093] S412: If the outer boundary of the target obstacle is located to the right of the envelope of the autonomous vehicle formed by the first trajectory line, then output a decision to detour around the target obstacle to the left.

[0094] S413: If the outer boundary of the target obstacle intersects with the envelope surface of the autonomous vehicle formed by the first trajectory line, then output a decision result of not detouring around the target obstacle.

[0095] In some embodiments of the present invention, determining a first trajectory line based on a plurality of first sampling points determined by a plurality of sampling layers includes: connecting the plurality of first sampling points to obtain the first trajectory line.

[0096] Specifically, in some embodiments, the first trajectory line can also be used for similarity cost calculation in the next planning, which will not be elaborated here.

[0097] The present invention is based on a sampling evaluation algorithm. When obstacles are dense, the sampling density is increased to improve the detour decision-making ability of the target vehicle. When there are no obstacles or obstacles are sparse, the sampling density is reduced to reduce the computational complexity. At the same time, combined with the design of optimized sampling point cost categories, the detour decision-making ability and computational complexity are effectively balanced.

[0098] Example 2

[0099] Figure 6 This is a schematic diagram of an obstacle avoidance decision-making device 600 for an autonomous vehicle provided as an embodiment of this specification. Please refer to... Figure 6 An obstacle avoidance decision-making device 600 for an autonomous vehicle in one embodiment includes:

[0100] The first module 601 is capable of sampling the first sampling layer based on the determined sampling density of the first sampling points;

[0101] The second module 602 is capable of traversing and calculating the total cost of all sampling points in the first sampling layer, and determining a first candidate point in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point;

[0102] The third module 603 is able to determine the sampling density of the next sampling layer and perform sampling based on the updated first sampling point until a preset sampling termination condition is reached.

[0103] The fourth module 604 is capable of determining a first trajectory line based on multiple first sampling points determined by multiple sampling layers, and obtaining a detour decision result for the target obstacle based on the first trajectory line.

[0104] It should be understood that the obstacle avoidance decision-making device 600 for an autonomous vehicle according to an embodiment of this specification can also perform... Figures 1 to 5 A method for the obstacle avoidance decision-making device (or equipment) of an autonomous vehicle, and to realize the obstacle avoidance decision-making device (or equipment) of an autonomous vehicle in Figures 1 to 5 The functionality of the example shown will not be elaborated upon here.

[0105] Example 3

[0106] This embodiment provides an autonomous vehicle. The autonomous vehicle is capable of navigating around obstacles and obtains obstacle detour planning by executing the method described in Embodiment 1. Therefore, the method for obtaining obstacle detour planning by the autonomous vehicle possesses the technical features of the method described in Embodiment 1 and has corresponding beneficial effects; specific technical features will not be elaborated further. The present invention is based on a sampling evaluation algorithm. When obstacles are dense, the sampling density is increased to improve the detour decision-making ability of the target vehicle; when there are no obstacles or obstacles are sparse, the sampling density is reduced to lower computational complexity; simultaneously, combined with the optimized sampling point cost category design, it effectively balances detour decision-making ability and computational complexity.

[0107] Example 4

[0108] Figure 7 This is a schematic diagram of the structure of an electronic device according to one embodiment of this specification. Please refer to it. Figure 7 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for other business operations.

[0109] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0110] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0111] The processor reads the corresponding computer program from non-volatile memory into main memory and then executes it, forming a shared resource access control mechanism at the logical level. The processor executes the program stored in memory and specifically performs the following operations:

[0112] Based on the determined sampling density of the first sampling point, the first sampling layer is sampled;

[0113] The total cost of all sampling points in the first sampling layer is calculated by traversing the layer, and a first candidate point is determined in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point.

[0114] Based on the updated first sampling point, the sampling density of the next sampling layer is determined and sampling is performed until the preset sampling termination condition is reached.

[0115] Based on multiple first sampling points determined by multiple sampling layers, a first trajectory line is determined, and based on the first trajectory line, a detour decision result for the target obstacle is obtained.

[0116] The above is as described in this instruction manual. Figures 1 to 5The obstacle avoidance decision-making method for autonomous vehicles disclosed in the illustrated embodiments can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this specification. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this specification can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0117] Of course, in addition to the software implementation, the electronic devices in the embodiments of this specification do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0118] Example 5

[0119] This specification also provides an embodiment of a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by a portable electronic device including multiple applications, enable the portable electronic device to perform... Figures 1 to 5 The obstacle avoidance decision-making method for an autonomous vehicle shown in the embodiment is specifically used to perform the following methods:

[0120] Based on the determined sampling density of the first sampling point, the first sampling layer is sampled;

[0121] The total cost of all sampling points in the first sampling layer is calculated by traversing the layer, and a first candidate point is determined in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point.

[0122] Based on the updated first sampling point, the sampling density of the next sampling layer is determined and sampling is performed until the preset sampling termination condition is reached.

[0123] Based on multiple first sampling points determined by multiple sampling layers, a first trajectory line is determined, and based on the first trajectory line, a detour decision result for the target obstacle is obtained.

[0124] In summary, the above description is merely a preferred embodiment of this specification and is not intended to limit the scope of protection of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

[0125] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0126] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0127] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0128] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

Claims

1. An obstacle avoidance decision-making method for an autonomous vehicle, characterized in that, include: Based on the determined sampling density of the first sampling point, the first sampling layer is sampled; The total cost of all sampling points in the first sampling layer is calculated by traversing the layer, and a first candidate point is determined in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point. Based on the updated first sampling point, the sampling density of the next sampling layer is determined and sampling is performed until the preset sampling termination condition is reached. Based on multiple first sampling points determined by multiple sampling layers, a first trajectory line is determined, and based on the first trajectory line, a detour decision result for the target obstacle is obtained; The sampling density of the first sampling point is determined based on the density of target obstacles within a first distance threshold from the first sampling point, including: If the number of target obstacles within a preset distance from the first sampling point is greater than a preset number, the sampling density is increased; If the number of target obstacles within a preset distance from the first sampling point is less than a preset number, or if the first sampling point is located within the outer boundary of the target obstacle, the sampling density is reduced. The step of obtaining the detour decision result for the target obstacle based on the first trajectory line includes: If the outer boundary of the target obstacle is located to the left of the envelope of the autonomous vehicle formed by the first trajectory line, then the decision to detour around the target obstacle to the right is output. If the outer boundary of the target obstacle is located to the right of the envelope of the autonomous vehicle formed by the first trajectory line, then the decision to detour around the target obstacle to the left is output. If the outer boundary of the target obstacle intersects with the envelope of the autonomous vehicle formed by the first trajectory line, then the decision result of not detouring around the target obstacle is output.

2. The method according to claim 1, characterized in that, The total cost includes the sum of a first cost and a second cost; wherein the first cost includes one or more costs characterizing the relationship between the current sampling point and the relevant parameters of the existing planning decision results, and the second cost includes one or more prediction costs characterizing the relationship between the current sampling point and the endpoint.

3. The method according to claim 2, characterized in that, The first cost includes one or more of the following: distance cost, reference line cost, boundary cost, static obstacle cost, dynamic obstacle cost, and similarity cost. The distance cost is used to characterize the cost of the distance between the current sampling point and the sampling point of the previous sampling layer; the reference line cost is used to characterize the cost of the lateral displacement between the current sampling point and the reference line; the boundary cost is used to characterize the cost of the distance between the autonomous vehicle and the lane boundary line; the static obstacle cost is used to characterize the cost of the distance between the current sampling point and static obstacles within a preset range; and the similarity cost is used to characterize the cost of the distance between the current sampling point and the sampling point at the same location as the previous planning decision result.

4. The method according to claim 2, characterized in that, The second cost includes one or the sum of two costs: the endpoint distance cost and the obstacle compensation cost; wherein, the endpoint distance cost is used to characterize the predicted cost of the distance from the current sampling point to the endpoint; and the obstacle compensation cost is used to characterize the predicted cost of compensation when the obstacle cost cannot be obtained within a preset range behind the obstacle.

5. The method according to any one of claims 2 to 4, characterized in that, Determine a first candidate point in the first sampling layer according to preset rules, including selecting the sampling point with the minimum total cost in the first sampling layer as the first candidate point.

6. The method according to claim 1, characterized in that, Before sampling the first sampling layer based on a determined sampling density, the current position of the autonomous vehicle is also obtained as the initial first sampling point.

7. The method according to claim 6, characterized in that, The sampling termination condition until the preset conditions are met includes: when the distance between the second projection point of the updated first sampling point on the reference line and the first projection point of the initial first sampling point on the reference line along the s direction in the Frenet coordinate system is greater than or equal to a second distance threshold, sampling is terminated.

8. The method according to any one of claims 1 or 7, characterized in that, Determining a first trajectory line based on multiple first sampling points determined by multiple sampling layers includes: connecting multiple first sampling points to obtain the first trajectory line.

9. An obstacle avoidance decision-making device for an autonomous vehicle, characterized in that, The apparatus for performing the method according to any one of claims 1 to 8 comprises: The first module is capable of sampling the first sampling layer based on the sampling density of the determined first sampling points; The second module is capable of traversing and calculating the total cost of all sampling points in the first sampling layer, and determining a first candidate point in the first sampling layer according to a preset rule, so as to update the first sampling point as the first candidate point; The third module is able to determine the sampling density of the next sampling layer and perform sampling based on the updated first sampling point until the preset sampling termination condition is reached. The fourth module is able to determine a first trajectory line based on multiple first sampling points determined by multiple sampling layers, and obtain a detour decision result for the target obstacle based on the first trajectory line.

10. An autonomous vehicle, characterized in that, Obtain obstacle detour planning decisions for the autonomous vehicle by performing the method according to any one of claims 1 to 8.

11. An electronic device, characterized in that, include: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method according to any one of claims 1 to 8.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs that, when executed by an electronic device including a plurality of applications, cause the electronic device to perform the method of any one of claims 1 to 8.

Images