Vehicle collision detection method and device

Abstract

The invention relates to the technical field of intelligent driving, and discloses a vehicle collision detection method and device, and the method comprises the steps: carrying out the first collision detection in a target grid map along a planned path through employing a target vehicle body model, obtaining a first collision detection result along the planned path, and carrying out the coarse-grained collision detection through employing the grid map, the data processing amount is reduced; and then when the first collision detection result indicates that no collision occurs, performing second collision detection on the obstacle information around the target parking space in the vector map by adopting the original vehicle body model to obtain a second collision detection result, thereby obtaining a second collision detection result aiming at the key area around the target parking space. And second-stage collision detection with finer granularity is carried out by adopting the vector map, so that the calculation amount of the collision detection and the accuracy of the detection result of the key area around the target parking space are considered, and the detection efficiency of the collision detection is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of intelligent driving technology, and in particular to a vehicle collision detection method and device. Background Art

[0002] In the automatic parking assist system (APA), path planning is a key step to ensure that the vehicle can park safely and efficiently in the parking space. Since APA usually operates in unstructured scenarios, such as basements or roadsides, path planning needs to process complex environmental information and dynamic obstacles.

[0003] When planning a path, collision-free is a strong constraint. To achieve this goal, how to balance the efficiency and accuracy of collision detection algorithms has become an urgent problem to be solved. Summary of the invention

[0004] In view of this, the present disclosure provides a vehicle collision detection method and device to solve the problem of how to strike a balance between efficiency and accuracy of collision detection in the prior art.

[0005] In a first aspect, the present disclosure provides a vehicle collision detection method, the method comprising: using a target vehicle body model, performing a first collision detection in a target grid map along a planned path, and obtaining a first collision detection result; wherein the target vehicle body model is generated for a target layer in a target grid map and based on an original vehicle body model; the target grid map is obtained by projecting obstacle elements in a vector map onto a grid map; the vehicle contour and rearview mirror of the original vehicle body model are respectively represented by rectangular frames; when the first collision detection result indicates that no collision has occurred, using the original vehicle body model, performing a second collision detection on obstacle information around the target parking space in the vector map, and obtaining a second collision detection result.

[0006] In a second aspect, the present disclosure provides a vehicle collision detection device, which includes: a first detection module, which is used to use a target vehicle body model to perform a first collision detection in a target grid map along a planned path to obtain a first collision detection result; wherein the target vehicle body model is generated for a target layer in the target grid map and based on an original vehicle body model; the target grid map is obtained by projecting obstacle elements in a vector map onto a grid map; the vehicle contour and rearview mirror of the original vehicle body model are respectively represented by rectangular frames; and a second detection module, which is used to use the original vehicle body model to perform a second collision detection on obstacle information around the target parking space in the vector map to obtain a second collision detection result when the first collision detection result indicates that no collision has occurred.

[0007] In a third aspect, the present disclosure provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the vehicle collision detection method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0008] In a fourth aspect, the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the vehicle collision detection method of the first aspect or any corresponding embodiment thereof.

[0009] In a fifth aspect, the present disclosure provides a computer program product, including computer instructions, which are used to enable a computer to execute the vehicle collision detection method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0010] The vehicle collision detection method in this embodiment adopts the target vehicle body model to perform the first collision detection in the target grid map along the planned path to obtain the first collision detection result along the planned path, thereby using the grid map to perform coarse-grained collision detection to reduce the amount of data processing; thereafter, when the first collision detection result indicates that no collision has occurred, the original vehicle body model is used to perform a second collision detection on the obstacle information around the target parking space in the vector map to obtain the second collision detection result, thereby using the vector map to perform a finer-grained second-stage collision detection on the key area around the target parking space, thereby taking into account both the amount of collision detection calculation and the accuracy of the detection results for the key area around the target parking space, thereby improving the detection efficiency of collision detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the specific embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0012] Figure 1a is a schematic flow chart of a vehicle collision detection method according to an embodiment of the present disclosure;

[0013] Figure 1b is a fused grid map according to an embodiment of the present disclosure;

[0014] Figure 2a is a schematic flow chart of a method for performing a first collision detection on a vehicle according to an embodiment of the present disclosure;

[0015] Figure 2bis an exemplary schematic diagram of using a rectangular frame sequence to perform collision detection in a target layer according to an embodiment of the present disclosure;

[0016] Figure 2c is an exemplary schematic diagram of the outline of a target vehicle body model obtained by expanding an original vehicle body model represented by a rectangular frame according to an embodiment of the present disclosure;

[0017] Figure 2d A schematic diagram of an original vehicle body model including a rearview mirror according to an embodiment of the present disclosure with the rear edge of the vehicle removed;

[0018] Figure 2e A schematic diagram of an original vehicle body model including a rearview mirror according to an embodiment of the present disclosure with the front edge of the vehicle removed;

[0019] Figure 3a is a schematic flow chart of a method for performing a second collision detection on a vehicle according to an embodiment of the present disclosure;

[0020] Figure 3b is an exemplary schematic diagram of generating virtual parking space information based on obstacle information around a target parking space and target parking space information;

[0021] Figure 4 is a schematic flow chart of a specific application scenario of the above vehicle collision detection method according to an embodiment of the present disclosure;

[0022] Figure 5 is a structural block diagram of a vehicle collision detection device according to an embodiment of the present disclosure;

[0023] Figure 6 It is a schematic diagram of the hardware structure of the computer device of the embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0025] In related technologies, the goal of automatic parking assist system (APA) path planning is to plan a collision-free, efficient and comfortable path for the downstream control module. Among them, collision-free is a strong constraint goal. APA planning generally occurs in unstructured scenarios such as basements or roadsides. In addition to generating a vector map, the upstream will also generate a raster map based on the free space information for downstream use.

[0026] APA's upstream data processing and fusion module can generate vector maps and raster maps. Vector map is a method of abstractly representing the environment. It uses vectors (such as line segments, polygons, etc.) to describe features such as obstacles, road boundaries, parking spaces, etc. in the environment. When constructing a raster map, a raster map can be generated based on free space information (such as sensor data such as lidar, ultrasonic sensors, and cameras). Each grid is marked as occupied or free depending on whether it contains obstacles. During the driving process of the vehicle, new sensor data is continuously received, so that the vector map and raster map are updated in real time to reflect changes in the environment.

[0027] In order to take into account both the integrity of map elements and map accuracy, APA needs to rely on two map formats when doing collision detection: one is a vector map (providing higher accuracy); the other is a raster map (providing more complete obstacle information). The vector map contains information such as pillars, cones, signs, ground locks (divided into open and closed states), pedestrians, and limiters, which are stored in the form of bounding boxes. The raster map is constructed based on the perceived segmentation results. In addition to the obstacles in the vector map, it also contains other obstacles that are difficult to identify or irregular in shape, such as walls, curbs, and fire extinguishers, which are stored in the form of occupied grids.

[0028] When there are too many elements in the vector map, the time consumption of collision detection based on the vector map will increase by N times, while the time consumption of collision detection based on the grid map will not increase with the increase of the number of obstacles. In view of this, this paper divides the collision detection into two stages: 1) Project the obstacle elements in the vector map to the grid map to obtain a fused grid map, and then perform the first rough collision detection based on the fused grid map (see the subsequent chapters for detailed scheme); 2) If no collision occurs in the first stage of detection around the target parking space, a more detailed second stage collision detection is performed around the target parking space based on the vector map (see the subsequent chapters for detailed scheme).

[0029] The disclosed embodiment provides a vehicle collision detection method, comprising: using a target vehicle body model, performing a first collision detection in a target grid map along a planned path, and obtaining a first collision detection result; wherein the target grid map is obtained by projecting obstacle elements in a vector map onto a grid map; when the first collision detection result indicates that no collision has occurred, using the original vehicle body model, performing a second collision detection on obstacle information around the target parking space in the vector map, and obtaining a second collision detection result.

[0030] According to an embodiment of the present disclosure, a vehicle collision detection method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] In this embodiment, a vehicle collision detection method is provided, which can be used for an execution subject such as a terminal or a server, such as a vehicle-mounted brain, a cloud server, etc. In some examples, the execution subject can be a decision-making system of the vehicle-mounted brain, including multiple subsystems such as route planning, path planning, motion planning and control. In other examples, the execution subject can be an automatic parking assistance system in a smart vehicle.

[0032] Figure 1a is a flow chart of a vehicle collision detection method according to an embodiment of the present disclosure, such as Figure 1a As shown, the process includes the following steps:

[0033] Step S101 , using the target vehicle body model, performing a first collision detection in the target grid map along the planned path to obtain a first collision detection result.

[0034] In this embodiment, the target vehicle model is generated based on the target layer in the target grid map and the original vehicle model, for example, the original vehicle model can be obtained by scaling and / or removing the model boundary. The specific method of scaling and / or removing the model boundary can adopt the method of scaling and / or removing the model boundary in the prior art, or can be determined according to the actual application scenario and / or the experience of those skilled in the art, and this application does not limit this.

[0035] The original vehicle body model may use a rectangle or a polygon to represent the outline of the vehicle. In some examples, the original vehicle body model may include a vehicle outline represented by a rectangular frame. In some examples, the original vehicle body model may also include a rearview mirror represented by a rectangular frame.

[0036] The planned path may be generated by the above-mentioned execution subject or its upstream processing node based on the results of environmental perception and parking space detection, and is a safe and collision-free driving path from the current vehicle position to the parking space, which usually includes a series of coordinate points or line segments.

[0037] The target raster map is a fused raster map obtained by projecting the obstacle elements in the vector map onto the raster map. Obstacle elements refer to elements that hinder the operation of the vehicle and are located on the driving path and / or the target parking space. In some specific examples, they can be pillars, cones, signs, ground locks (in open and closed states), pedestrians, and limiters, etc., which can usually be stored in the form of a bounding box.

[0038] For example, see Figure 1b , Figure 1b It is a fused grid map according to an embodiment of the present disclosure, in which yellow pillars and green limiters are projected into the grid map.

[0039] When there are too many elements in the vector map, the time required for collision detection based on the vector map will increase by N times, while the time required for collision detection based on the grid map will not increase with the increase in the number of obstacles. Therefore, the above-mentioned execution entity can perform the first relatively rough collision detection based on the fused grid map, that is, perform the first collision detection.

[0040] Step S102, when the first collision detection result indicates that no collision occurs, a second collision detection is performed on the obstacle information around the target parking space in the vector map using the original vehicle body model to obtain a second collision detection result.

[0041] In this embodiment, the obstacle information around the target parking space refers to the obstacle information around the target parking space that prevents the vehicle from parking in the parking space. In some specific examples, it may include information such as pillars, cones, signs, ground locks (divided into open and closed states), pedestrians, and limiters, which are stored in the vector map in the form of a bounding box.

[0042] When the above-mentioned execution entity obtains the first collision detection result, it can perform a rough screening based on the first collision detection result to reduce subsequent processing data, and for situations where no collision occurs, perform a more detailed second-stage collision detection based on the vector map around the target parking space.

[0043] The vehicle collision detection method in this embodiment adopts the target vehicle body model to perform the first collision detection in the target grid map along the planned path to obtain the first collision detection result along the planned path, thereby using the grid map to perform coarse-grained collision detection to reduce the amount of data processing; thereafter, when the first collision detection result indicates that no collision has occurred, the original vehicle body model is used to perform a second collision detection on the obstacle information around the target parking space in the vector map to obtain the second collision detection result, thereby using the vector map to perform a finer-grained second-stage collision detection on the key area around the target parking space, thereby taking into account both the amount of collision detection calculation and the accuracy of the detection results for the key area around the target parking space, thereby improving the detection efficiency of collision detection.

[0044] In this embodiment, a method for performing a first collision detection on a vehicle is provided, which can be used for the above-mentioned execution subject. Figure 2a is a flow chart of a method for performing a first collision detection on a vehicle according to an embodiment of the present disclosure. The method for performing a first collision detection on a vehicle is the above Figure 1a Some optional implementations of step S101 in . Specifically, Figure 2a As shown, the process of the method for performing the first collision detection of the vehicle includes the following steps:

[0045] Step S201 , using the target vehicle body model, respectively performing collision detection in multiple layers of the target grid map along the planned path, and obtaining collision detection results corresponding to each layer.

[0046] In this embodiment, the target vehicle body model is generated based on the target layer in the target grid map and the original vehicle body model. For example, the target vehicle body model can be obtained by scaling the original vehicle body model and / or removing the model boundary. The original vehicle body model may include a vehicle outline represented by a rectangular frame. In some examples, the original vehicle body model may also include a rearview mirror represented by a rectangular frame.

[0047] The target grid map can include multiple types of layers: drivable space layer (also known as free space layer, freespace layer), target parking space layer, and obstacle layer. For different types of layers, targeted target vehicle models can be generated, and collision detection can be performed on the drivable space layer, target parking space layer, and obstacle layer along the planned path to obtain the collision detection results corresponding to each type of layer.

[0048] It is understandable that when generating targeted target vehicle body models for different types of layers, the different characteristics of the drivable space class layer, the target parking space class layer and the obstacle class layer can be considered, and the original vehicle body model can be obtained by scaling it to different degrees and / or removing the model boundary.

[0049] For example, the sizes of the target vehicle models generated for the three types of layers, namely, the drivable space layer, the target parking space layer, and the obstacle layer, are expanded in sequence. Alternatively or additionally, the degree to which the rearview mirror is considered in the target vehicle models generated for the three types of layers, namely, the drivable space layer, the target parking space layer, and the obstacle layer, is enhanced in sequence.

[0050] When the target vehicle model is used to perform collision detection in multiple layers of the target grid map along the planned path, the target vehicle model can be placed at each detection point or line segment of the planned path in each layer, and the actual space occupied by it is considered; then, the grids within the spatial range are traversed to check the layer information to which they belong, and if the target vehicle model intersects with the obstacle point in the traversed grid, it is determined to be a collision.

[0051] Step S202: when there is a target detection result indicating a collision in the collision detection results corresponding to each type of layer, a first collision detection result including collision information is output based on the target detection result.

[0052] In this embodiment, after obtaining the collision detection results of each type of layer, if there is a detection result indicating a collision, it is used as the target detection result. Then, based on the target detection result, a first collision detection result can be output, and the collision detection result includes collision information, and the collision information indicates the layer where the target vehicle model collides and the position where the target vehicle model collides.

[0053] The method for performing the first collision detection on the vehicle in this embodiment obtains the first collision detection result by adopting different target vehicle models to perform collision detection on different types of layers, thereby realizing differentiated collision detection of obstacles, and taking into account the efficiency, accuracy and adaptability to various scenes and various types of obstacles of collision detection in both spatial and category dimensions.

[0054] In the above Figure 2a In some optional implementations of the illustrated embodiment, the above step S201 includes:

[0055] Step S2011, using the target vehicle body model, in the target layer, a sequence of rectangular frames representing the vehicle body contour is generated along the planned path.

[0056] In this embodiment, the target layer is each layer in the multiple types of layers of the target grid map. The rectangular frame sequence corresponds to the target vehicle body model, that is, the target vehicle body model sequence moving along the planned path.

[0057] Step S2012, using a rectangular frame sequence to perform collision detection in the target layer, and obtaining a collision detection result corresponding to the target layer.

[0058] In this embodiment, when performing collision detection, a series of rectangular boxes representing the outline of the vehicle body are generated along the planned path in each layer, and then the four sides of each rectangular box are judged in turn whether they collide with the obstacle points in the grid map. If any rectangular box of any layer collides, it is determined that a collision has occurred, and the collision result (which side of the vehicle body has collided in which layer) is immediately returned.

[0059] In a specific example, Figure 2b As shown, Figure 2b It is an exemplary schematic diagram of using a rectangular frame sequence to perform collision detection in a target layer according to an embodiment of the present disclosure.

[0060] In some optional implementations of the above embodiments, the drivable space class layer may include at least one of the following: an empty drivable space layer; a vacant parking space layer; the target parking space class layer may include: a target parking space layer; the obstacle class layer includes at least one of the following: an unknown obstacle layer; a layer of obstacles that can invade the vehicle body; a layer of obstacles that can be pressed by the rearview mirror and can be attached to the vehicle body; a layer of obstacles inside and around the target parking space; a layer of obstacles at the entrance of the target parking space; a layer of obstacles outside the target parking space; a layer of dangerous collision obstacles inside and around the target parking space; a layer of dangerous collision obstacles outside the target parking space.

[0061] In this embodiment, the layer generation is performed based on the fused raster map. The purpose of the layer generation is to achieve differentiated obstacle collision detection. The differentiation is reflected in two dimensions: space and category. For example: if the distance between the vehicle and the pillar outside the target parking space is less than 40cm, it is judged as a collision, but if the distance between the vehicle and the pillars on both sides of the parking space is less than 30cm, it is judged as a collision; if the distance between the vehicle and the wall is less than 40cm, it is judged as a collision, but if the distance between the vehicle and the curb is less than 10cm, it is judged as a collision.

[0062] According to common parking scenarios, the drivable space layer, target parking space layer and obstacle layer can be divided into the following layers respectively.

[0063] The drivable space layer may include at least one of the following: an empty drivable space layer; and an empty parking space layer.

[0064] The target parking space class layer may include: target parking space layer.

[0065] The obstacle layers include at least one of the following: unknown obstacle layer; vehicle-invasive obstacle layer; mirror-pressable and vehicle-body-stickable obstacle layer; target parking space obstacle layer; target parking space entrance obstacle layer; target parking space obstacle layer; target parking space outside obstacle layer; target parking space inside and around dangerous collision obstacle layer; target parking space outside dangerous collision obstacle layer.

[0066] In a specific example, based on common parking scenarios, it can be divided into the following centralized layers:

[0067]

[0068]

[0069] in,

[0070] COST_FREE refers to the open drivable space layer;

[0071] COST_FREE_STALL refers to the free parking space layer;

[0072] COST_GOAL_STALL refers to the target parking space layer;

[0073] COST_UNKNOWN refers to the unknown obstacle layer;

[0074] COST_CAN_INVADE_PARTIAL_VEHICLE means that the vehicle obstacle layer can be invaded;

[0075] COST_ROAD_SIDE means that the rearview mirror can be pressed and the vehicle body can be attached with obstacle layers;

[0076] COST_COLLISION_INSTALL refers to the obstacle layer around the target parking space;

[0077] COST_COLLISION_STALL_DOOR refers to the obstacle layer at the target parking space entrance;

[0078] COST_COLLISION refers to the obstacle layer outside the target parking space;

[0079] COST_DANGER_IN_STALL refers to the dangerous collision obstacle layer around the target parking space;

[0080] COST_DANGER_OUT_STALL refers to the dangerous collision obstacle layer outside the target parking space.

[0081] In some optional implementations of the above embodiment, the target vehicle body model is generated by at least one of the following:

[0082] For the first target layer, the front of the original vehicle model is expanded forward by 0 distance, the rear of the vehicle is expanded backward by 0 distance, the left and right sides of the vehicle are expanded by 0 distance, and the rectangular frame representing the rearview mirror is deleted to obtain the target vehicle model; the first target layer is at least one of the empty drivable space layer, the vacant parking space layer, and the target parking space layer.

[0083] For the second target layer, the front of the original vehicle model is expanded forward by a first preset distance, and the rear of the vehicle is expanded backward by a first preset distance to obtain a target vehicle model; the second target layer is at least one of an unknown obstacle layer, a rearview mirror compressible and / or vehicle body stickable obstacle layer; the first preset distance is less than 0.

[0084] For the third target layer, the front of the original vehicle model is expanded forward by a first preset distance, the rear of the vehicle is expanded backward by a first preset distance, and the rectangular frame representing the rearview mirror is deleted to obtain the target vehicle model; the third target layer is a layer of obstacles that can invade the vehicle body.

[0085] For the fourth target layer, the front of the original vehicle model is expanded forward by the second preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the first preset distance to obtain the target vehicle model; the fourth target layer is the surrounding obstacle layer in the target parking space and the surrounding dangerous collision obstacle layer in the target parking space; the second preset distance is greater than 0.

[0086] For the fifth target layer, the front of the original vehicle model is expanded forward by the third preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the first preset distance to obtain the target vehicle model; the fifth target layer is the obstacle layer at the target parking space entrance and the dangerous collision obstacle layer outside the target parking space; the third preset distance is greater than the second preset distance.

[0087] For the sixth target layer, the front of the original vehicle model is expanded forward by the third preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the second preset distance to obtain a target vehicle model; the sixth target layer is an obstacle layer outside the target parking space; the second preset distance is greater than the first preset distance.

[0088] In this implementation, after the layers are generated, the vehicle model is constructed according to different layers, and the vehicle outline and rearview mirror are represented by rectangular boxes. Considering the efficiency of the algorithm, the traditional method of expanding the grid map is changed to expand the vehicle itself.

[0089] The expansion distance of the car body in different layers is different, and whether the rearview mirror is considered is also different. In a specific example, please refer to the following table:

[0090]

[0091]

[0092] in,

[0093] COST_FREE refers to the open drivable space layer;

[0094] COST_FREE_STALL refers to the free parking space layer;

[0095] COST_GOAL_STALL refers to the target parking space layer;

[0096] COST_UNKNOWN refers to the unknown obstacle layer;

[0097] COST_CAN_INVADE_PARTIAL_VEHICLE means that the vehicle obstacle layer can be invaded;

[0098] COST_ROAD_SIDE means that the rearview mirror can be pressed and the vehicle body can be attached with obstacle layers;

[0099] COST_COLLISION_INSTALL refers to the obstacle layer around the target parking space;

[0100] COST_COLLISION_STALL_DOOR refers to the obstacle layer at the target parking space entrance;

[0101] COST_COLLISION refers to the obstacle layer outside the target parking space;

[0102] COST_DANGER_IN_STALL refers to the dangerous collision obstacle layer around the target parking space;

[0103] COST_DANGER_OUT_STALL refers to the dangerous collision obstacle layer outside the target parking space.

[0104] The distance the front of the vehicle expands forward is represented by dis_front, the distance the rear of the vehicle expands backward is represented by dis_rear, and the distance the left and right sides of the vehicle expand are represented by dis_Ir.

[0105] For a specific example, see Figure 2c , Figure 2c It is an exemplary schematic diagram of the outline of the target vehicle body model obtained by expanding the original vehicle body model represented by a rectangular frame according to an embodiment of the present disclosure.

[0106] Alternatively or additionally, based on the above-mentioned method of generating the target vehicle body model, the above-mentioned target vehicle body model can, when the distance to the obstacle is less than a preset distance, remove the contour edge of the original vehicle body model that is opposite to the planned gear position according to the planned gear position; the planned gear position is determined based on the planned path.

[0107] In some specific examples, in some scenarios where the vehicle is very close to an obstacle, the recognition error of the obstacle is relatively large (sometimes even intruding into the vehicle body contour). For these special scenarios, when modeling the vehicle, an edge of the vehicle body contour is removed according to the planned gear direction.

[0108] Figure 2d FIG. 1 is a schematic diagram showing an original vehicle body model including a rearview mirror according to an embodiment of the present disclosure, with the rear edge of the vehicle removed. Figure 2d As shown, if the vehicle is in forward gear, the rear end (with an optical blind spot monitoring system OBS installed) of the vehicle body contour can be removed.

[0109] Figure 2e FIG. 2 shows a schematic diagram of an original vehicle body model including a rearview mirror according to an embodiment of the present disclosure, with the front edge of the vehicle removed. Figure 2e As shown, if the vehicle is in reverse gear, the front side of the vehicle body contour (equipped with an optical blind spot monitoring system OBS) can be removed.

[0110] In this embodiment, a method for performing a second collision detection on a vehicle is provided, which can be used in the above-mentioned execution subject. Figure 3a is a flow chart of a method for performing a second collision detection on a vehicle according to an embodiment of the present disclosure. The method for performing a second collision detection on a vehicle is the above Figure 1a Some optional implementations of step S102 in . Specifically, Figure 3a As shown, the process of the method for performing a second collision detection on the vehicle includes the following steps:

[0111] Step S301, generating virtual parking space information based on obstacle information around the target parking space and target parking space information in the vector map.

[0112] In this embodiment, collision detection is performed based on a vector map. When the number of obstacles increases, the detection efficiency will be significantly reduced. Therefore, in the above embodiment, the obstacle elements in the original vector map are mapped to the grid map. However, in order to perform more precise collision detection in narrow parking space scenarios, collision detection still needs to be performed based on a vector map.

[0113] Here, you can only focus on the obstacle information around the target parking space when doing vector map collision detection. The obstacle information around the target parking space refers to the obstacle information located around the target parking space that is used to hinder the parking of vehicles and the parking space. In some specific examples, it can include information such as pillars, cones, signs, ground locks (divided into open and closed states), pedestrians, and limiters, which are stored in the vector map in the form of a bounding box. In order to further reduce the amount of calculation, the above-mentioned execution entity can generate a virtual parking space based on the obstacle information around the target parking space and the target parking space information.

[0114] Step S302: using the original vehicle body model, a second collision detection is performed on the virtual parking space information.

[0115] In this embodiment, each time a collision detection based on a vector map is performed, only a collision detection between the vehicle body contour (ie, the original vehicle body model) and the virtual parking space needs to be performed.

[0116] In a specific example, virtual parking space information (such as Figure 3b The virtual parking space information includes the obstacle information around the parking space (including the limiter and the pillars on both sides of the parking space) and the boundary line information of the target parking space. Each time a collision detection based on the vector map is performed, only the collision detection between the body contour of the original vehicle model and the virtual parking space is required.

[0117] In the above Figure 3a In some optional implementations of the illustrated embodiment, the above step S302 includes:

[0118] Step S3021, projecting the virtual parking space information and the original vehicle model to the first dimension and the second dimension respectively.

[0119] Step S3022: if a collision is determined to occur according to the projection in the first dimension and / or the projection in the second dimension, a second collision detection result including collision information is generated, where the first dimension and the second dimension are respectively the x-axis and the y-axis of the binary coordinate system.

[0120] Alternatively or additionally, the above step S302 may further include:

[0121] Step S3023: if it is determined that no collision occurs according to the projection in the first dimension and / or the projection in the second dimension, the normal direction of each group of contour edges of the obstacle element and the normal direction of each group of contour edges of the original vehicle body model are respectively used as target separation axes; wherein each group of contour edges includes contour edges without parallel contour edges or a group of parallel contour edges;

[0122] Step S3024: project the virtual parking space information and the original vehicle body model onto each target separation axis, and perform at least one of the following:

[0123] Step S3025, if it is determined that a collision occurs according to the projection on each target separation axis, a second collision detection result including collision information is generated;

[0124] Step S3026: If it is determined based on the projection on each target separation axis that no collision occurs, a second collision detection result including information that no collision occurs is generated.

[0125] In the method for performing the second collision detection in this embodiment, the obstacle information and the vehicle body contour information are first projected onto the x-axis and the y-axis to detect whether a collision occurs. If a collision occurs, the collision result is returned, thereby greatly reducing the amount of calculation for the second collision detection; if no collision occurs, the number of separation axes is increased, and the normal direction of each contour edge of the obstacle is used as a separation axis, and the normal direction of each vehicle body contour is used as a separation axis (considering that the vehicle body contour is a rectangle, two mutually parallel axes can be reduced here, reducing half of the calculation), and then the obstacle edge and the vehicle body contour edge are projected onto each separation axis respectively to detect whether a collision occurs. If a collision is found on any separation axis, the collision result is returned, otherwise it is determined that no collision occurs. Therefore, when no collision occurs in the two-dimensional projection, the second collision detection result is determined by projection on multiple separation axes, thereby improving the accuracy and safety of the collision detection result.

[0126] As one or more specific application examples of the embodiments of the present disclosure, please refer to Figure 4 , Figure 4 A schematic flow chart of a specific application scenario of the vehicle collision detection method is disclosed.

[0127] In this embodiment, a vehicle collision detection device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0128] This embodiment provides a vehicle collision detection device, such as Figure 5 As shown, including:

[0129] The first detection module 501 is used to use the target vehicle model to perform a first collision detection in the target grid map along the planned path to obtain a first collision detection result; wherein the target vehicle model is generated based on the target layer in the target grid map and based on the original vehicle model; the target grid map is obtained by projecting the obstacle elements in the vector map onto the grid map; the vehicle outline and rearview mirror of the original vehicle model are respectively represented by rectangular frames;

[0130] The second detection module 502 is used to perform a second collision detection on obstacle information around the target parking space in the vector map using the original vehicle model when the first collision detection result indicates that no collision has occurred, so as to obtain a second collision detection result.

[0131] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0132] The vehicle collision detection device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0133] See also Figure 6 , Figure 6 1 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present disclosure, and the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system).

[0134] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0135] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0136] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer 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.

[0137] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0138] The computer device further includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means.

[0139] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0140] The embodiments of the present disclosure also provide a computer-readable storage medium. The above-mentioned method according to the embodiments of the present disclosure can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium and downloaded through a network, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0141] A part of the present disclosure may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present disclosure through the operation of the computer. Those skilled in the art should understand that the existence of computer program instructions in computer-readable media includes, but is not limited to, source files, executable files, installation package files, etc., and accordingly, the way in which computer program instructions are executed by a computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to the computer.

[0142] Although the embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A vehicle collision detection method, characterized in that: The method comprises: Using the target vehicle body model, a first collision detection is performed in the target grid map along the planned path to obtain a first collision detection result; wherein the target vehicle body model is generated for the target layer in the target grid map and based on the original vehicle body model; the target grid map is obtained by projecting the obstacle elements in the vector map onto the grid map; the vehicle outline and rearview mirror of the original vehicle body model are respectively represented by rectangular frames; When the first collision detection result indicates that no collision occurs, the original vehicle body model is used to perform a second collision detection on obstacle information around the target parking space in the vector map to obtain a second collision detection result.

2. The method according to claim 1, characterized in that The target grid map includes multiple layers: a drivable space layer, a target parking space layer, and an obstacle layer; The target vehicle body model is used to perform a first collision detection in the target grid map along the planned path to obtain a first collision detection result, including: Using the target vehicle model, collision detection is performed in multiple layers of the target grid map along the planned path to obtain collision detection results corresponding to each layer. When there is a target detection result indicating a collision in the collision detection results corresponding to each layer, a first collision detection result including collision information is output based on the target detection result, wherein the collision information indicates the layer where the target vehicle body model collides and the position where the target vehicle body model collides.

3. The method according to claim 2, characterized in that The target vehicle model is used to perform collision detection in multiple layers of the target grid map along the planned path, and the collision detection results corresponding to each layer are obtained, including: Using a target vehicle body model, a sequence of rectangular frames representing the vehicle body contour is generated in a target layer along a planned path; the target layer is each layer in the multiple layers of the target grid map; The rectangular frame sequence is used to perform collision detection in the target layer to obtain a collision detection result corresponding to the target layer.

4. The method according to claim 3, characterized in that The drivable space layer includes at least one of the following: an open drivable space layer; an empty parking space layer; The target parking space class layer includes: a target parking space layer; The obstacle class layers include at least one of the following: an unknown obstacle layer; a layer of obstacles that can invade the vehicle body; a layer of obstacles that can be pressed by rearview mirrors and stuck to the vehicle body; a layer of obstacles around the target parking space; a layer of obstacles at the entrance of the target parking space; a layer of obstacles outside the target parking space; a layer of dangerous collision obstacles around the target parking space; and a layer of dangerous collision obstacles outside the target parking space.

5. The method according to claim 4, characterized in that The target vehicle body model is generated by at least one of the following: For the first target layer, the front of the original vehicle model is expanded forward by 0 distance, the rear of the vehicle is expanded backward by 0 distance, the left and right sides of the vehicle are expanded by 0 distance, and the rectangular frame representing the rearview mirror is deleted to obtain the target vehicle model; the first target layer is at least one of the open drivable space layer, the vacant parking space layer, and the target parking space layer; For the second target layer, the front of the original vehicle model is expanded forward by a first preset distance, and the rear of the vehicle is expanded backward by the first preset distance to obtain the target vehicle model; the second target layer is at least one of the unknown obstacle layer, the rearview mirror pressable and / or vehicle body stickable obstacle layer; the first preset distance is less than 0; For the third target layer, the front of the original vehicle model is expanded forward by the first preset distance, the rear of the vehicle is expanded backward by the first preset distance, and the rectangular frame representing the rearview mirror is deleted to obtain the target vehicle model; the third target layer is a layer of obstacles that can invade the vehicle body; For the fourth target layer, the front of the original vehicle model is expanded forward by the second preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the first preset distance to obtain the target vehicle model; the fourth target layer is a layer of obstacles around the target parking space and a layer of dangerous collision obstacles around the target parking space; the second preset distance is greater than 0; For the fifth target layer, the front of the original vehicle model is expanded forward by the third preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the first preset distance to obtain the target vehicle model; the fifth target layer is a target parking space entrance obstacle layer and a target parking space outside dangerous collision obstacle layer; the third preset distance is greater than the second preset distance; For the sixth target layer, the front of the original vehicle model is expanded forward by the third preset distance, the rear of the vehicle is expanded backward by the second preset distance, and the left and right sides of the vehicle are expanded by the second preset distance to obtain the target vehicle model; the sixth target layer is an obstacle layer outside the target parking space; the second preset distance is greater than the first preset distance.

6. The method according to claim 5, characterized in that When the target vehicle body model is less than a preset distance from the obstacle, the contour edge of the original vehicle body model that is opposite to the planned gear position is removed according to the planned gear position; the planned gear position is determined based on the planned path.

7. The method according to any one of claims 1 to 6, characterized in that: The original vehicle model is used to perform a second collision detection on the obstacle information around the target parking space in the vector map, including: Generate virtual parking space information based on obstacle information around the target parking space and target parking space information in the vector map; The original vehicle body model is used to perform a second collision detection on the virtual parking space information.

8. The method according to claim 7, characterized in that The method of using the original vehicle body model to perform a second collision detection on the virtual parking space information includes: The virtual parking space information and the original vehicle body model are projected to a first dimension and a second dimension respectively. If a collision is determined to occur based on the projection in the first dimension and / or the projection in the second dimension, a second collision detection result including collision information is generated. The first dimension and the second dimension are respectively the x-axis and the y-axis of a binary coordinate system.

9. The method according to claim 8, characterized in that The method of using the original vehicle body model to perform a second collision detection on the virtual parking space information further includes: If it is determined that no collision occurs according to the projection in the first dimension and / or the projection in the second dimension, the normal direction of each group of contour edges of the obstacle element and the normal direction of each group of contour edges of the original vehicle body model are respectively used as target separation axes; wherein each group of contour edges includes contour edges without parallel contour edges or a group of parallel contour edges; Project the virtual parking space information and the original vehicle body model onto each target separation axis respectively, and perform at least one of the following: If a collision is determined to occur according to the projection on each of the target separation axes, a second collision detection result including collision information is generated; If it is determined that no collision occurs according to the projection on each of the target separation axes, a second collision detection result including information that no collision occurs is generated.

10. A vehicle collision detection device, characterized in that: The device comprises: A first detection module is used to use a target vehicle body model to perform a first collision detection in a target grid map along a planned path to obtain a first collision detection result; wherein the target vehicle body model is generated based on a target layer in a target grid map and based on an original vehicle body model; the target grid map is obtained by projecting obstacle elements in a vector map onto a grid map; and the vehicle outline and rearview mirror of the original vehicle body model are respectively represented by rectangular frames; The second detection module is used to use the original vehicle body model to perform a second collision detection on obstacle information around the target parking space in the vector map when the first collision detection result indicates that no collision has occurred, so as to obtain a second collision detection result.

11. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the vehicle collision detection method according to any one of claims 1 to 7 by executing the computer instructions.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the vehicle collision detection method according to any one of claims 1 to 7.

13. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to cause a computer to execute the vehicle collision detection method according to any one of claims 1 to 7.

Images