Obstacle detection method and vehicle path planning method and device

Abstract

The present invention provides an obstacle detection method and a vehicle path planning method and apparatus. The obstacle detection method includes: obtaining historical state values ​​of obstacles within a target area where a target vehicle is located over a certain period of time, the historical state values ​​of the obstacles including the obstacle's position and speed; determining whether the obstacle is a static obstacle based on the historical state values; updating the current state value of the obstacle based on the determination result and the historical state values ​​of the obstacle; and detecting the obstacle based on the current state value of the obstacle to obtain an obstacle detection result. By analyzing and determining the obstacle, misjudgments due to slight jitter of the obstacle are avoided, improving static obstacle detection while significantly reducing the number of subsequent invalid vehicle path planning attempts, further increasing production efficiency.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to an obstacle detection method and a vehicle path planning method and device. Background Art

[0002] Environmental perception is a prerequisite for the realization of autonomous driving systems and provides an effective basis for autonomous vehicle path planning. Stable and reliable obstacle perception technology plays a key role in the execution of autonomous vehicles such as obstacle stopping and obstacle avoidance. The stability of obstacle perception results directly affects the safety, operating efficiency and driving experience of autonomous driving systems. In existing technologies, obstacle detection is mostly achieved by identifying the obstacle's posture frame by frame, and then planning the vehicle's path based on the single-frame recognition results. When the obstacle's posture shakes, the vehicle's path planning needs to be re-performed accordingly, and the efficiency of obstacle detection and vehicle path planning is very low. Summary of the Invention

[0003] Therefore, the technical problem to be solved by the present invention is to overcome the defect that obstacle detection in the prior art is mostly single-frame recognition, and subsequent vehicle driving path planning is performed based on the single-frame recognition results. When the obstacle appears to be shaking, the vehicle driving path planning needs to be re-performed accordingly, resulting in very low efficiency of obstacle detection and vehicle driving path planning, thereby providing an obstacle detection method and a vehicle path planning method and device.

[0004] According to a first aspect, an embodiment of the present invention provides an obstacle detection method, the method comprising:

[0005] Obtaining historical state values ​​of obstacles in a target area where the target vehicle is located within a certain period of time, wherein the historical state values ​​of the obstacles include position information and speed information of the obstacles;

[0006] Determining whether the obstacle is a static obstacle based on the position information and speed information of the obstacle;

[0007] Based on the judgment result and the historical state value of the obstacle, updating the current state value of the obstacle;

[0008] The obstacle is detected based on the current state value of the obstacle to obtain an obstacle detection result.

[0009] Optionally, before determining whether the obstacle is a static obstacle based on the position information and speed information of the obstacle, the method further includes:

[0010] Calculating a distance of the obstacle from the target vehicle based on the position information of the obstacle and the current position of the target vehicle;

[0011] Determining whether the distance between the obstacle and the target vehicle is less than a preset distance threshold;

[0012] When the distance between the obstacle and the target vehicle is less than the preset distance threshold, the position information of the obstacle is subjected to coordinate axis conversion to be converted into position information based on an absolute coordinate system.

[0013] Optionally, the determining whether the obstacle is a static obstacle based on the converted position information and speed information of the obstacle includes:

[0014] Extracting the position information and speed information of the obstacle at the starting time and the current time within the certain period of time respectively to obtain a first position, a second position, a first speed, and a second speed;

[0015] Calculating a first absolute position change of the obstacle based on the first position and the second position;

[0016] Calculating an average absolute speed of the obstacle based on the first speed and the second speed;

[0017] When the first absolute position change is smaller than the target position change, and / or the absolute speed average is smaller than the target speed average, it is determined that the obstacle is a static obstacle.

[0018] Optionally, the state value further includes a size and an orientation angle of the obstacle, and updating the current state value of the obstacle based on the judgment result and the state value of the obstacle includes:

[0019] When the obstacle is a static obstacle, the average values ​​of the position information, the speed information, the size, and the orientation angle are calculated based on the historical state values ​​of the obstacle within a certain period of time;

[0020] An average value of the position information, the speed information, the size, and the orientation angle is determined as a static state value at a current moment, and the static state value is used as a current state value of the obstacle for updating.

[0021] Optionally, when the obstacle is not a static obstacle, the position information, speed information, size, and orientation angle corresponding to the current moment are determined as the current state value of the obstacle.

[0022] Optionally, the calculating the distance of the obstacle relative to the target vehicle based on the position information of the obstacle and the current position of the target vehicle includes:

[0023] Extracting the coordinates of each inflection point of the space occupied by the obstacle from the position information of the obstacle;

[0024] Determining the distance of the obstacle relative to the target vehicle based on the distance between the current position and the coordinates of each inflection point of the space occupied by the obstacle; determining the distance of the obstacle relative to the target vehicle based on the distance between the current position and the coordinates of each inflection point of the space occupied by the obstacle, including:

[0025] sorting the distances between the current position and the coordinates of each inflection point of the space occupied by the obstacle;

[0026] The minimum distance in the sorting results is determined as the distance of the obstacle relative to the target vehicle.

[0027] According to a second aspect, an embodiment of the present invention provides a vehicle path planning method, the method comprising:

[0028] Perform obstacle detection using the obstacle detection method described in the first aspect or any optional embodiment of the first aspect to obtain an obstacle detection result;

[0029] Calculating the inflection point coordinates of the obstacle based on the obstacle detection result;

[0030] Determining a spatial contour of the obstacle based on the inflection point coordinates of the obstacle;

[0031] Based on the spatial contour of the obstacle, a driving path of the target vehicle is determined.

[0032] According to a third aspect, an embodiment of the present invention provides an obstacle detection device, comprising:

[0033] An acquisition module is used to obtain the historical state value of the obstacle in the target area where the target vehicle is located within a certain period of time, wherein the historical state value of the obstacle includes the position information and speed information of the obstacle;

[0034] a first processing module, configured to determine whether the obstacle is a static obstacle based on the position information and speed information of the obstacle;

[0035] a second processing module, configured to update a current state value of the obstacle based on the judgment result and the historical state value of the obstacle;

[0036] The third processing module is configured to detect the obstacle based on the current state value of the obstacle to obtain an obstacle detection result.

[0037] According to a fourth aspect, an embodiment of the present invention provides a vehicle path planning device, the device comprising:

[0038] a detection module, configured to perform obstacle detection using the obstacle detection device as described in the third aspect and obtain an obstacle detection result;

[0039] A first calculation module, configured to calculate the coordinates of the inflection point of the obstacle based on the obstacle detection result;

[0040] a second calculation module, configured to determine a spatial contour of the obstacle based on the inflection point coordinates of the obstacle;

[0041] The third calculation module is used to determine the driving path of the target vehicle based on the spatial contour of the obstacle.

[0042] According to a fifth aspect, an embodiment of the present invention provides an electronic device, including:

[0043] A memory and a processor, wherein the memory and the processor are communicatively connected to each other, computer instructions are stored in the memory, and the processor executes the method described in the first / second aspect, or any optional embodiment of the first / second aspect, by executing the computer instructions.

[0044] According to the sixth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute the method described in the first / second aspect, or any optional implementation of the first / second aspect.

[0045] The technical solution of the present invention has the following advantages:

[0046] The obstacle detection method and device provided by the present invention obtains the historical state values ​​of obstacles within a target area where a target vehicle is located over a certain period of time. The historical state values ​​include the obstacle's position and speed information. Based on the obstacle's position and speed information, the device determines whether the obstacle is a static obstacle. Based on the determination result and the obstacle's state value, the device updates the obstacle's current state value. Finally, the device detects the obstacle based on the obstacle's current state value to obtain an obstacle detection result. Obtaining the obstacle's historical state values ​​provides preliminary data support for subsequent comprehensive determination of whether the obstacle is a static obstacle. By performing a static and dynamic determination of the obstacle and specifically updating the obstacle's current state value, the device detects the obstacle based on the updated current state value to obtain a detection result. By analyzing and determining the obstacle, the device avoids misjudgments caused by slight obstacle motion. This improves static obstacle detection while significantly reducing the number of subsequent invalid vehicle path planning attempts, further increasing production efficiency.

[0047] The vehicle path planning method and device provided by the present invention utilize an obstacle detection method provided by another embodiment of the present invention to perform obstacle detection and obtain an obstacle detection result; based on the obstacle detection result, calculate the inflection point coordinates of the obstacle; based on the inflection point coordinates of the obstacle, determine the spatial contour of the obstacle; and, based on the spatial contour of the obstacle, determine the target vehicle's driving path. The target vehicle's driving path is determined based on the obstacle detection result. While ensuring that the planned vehicle path effectively avoids obstacles, the device further analyzes and determines the obstacles to avoid misjudgments caused by slight jitters of the obstacles. This significantly reduces the number of invalid vehicle path planning attempts, improving planning efficiency and the user's driving experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 This is a schematic diagram of conventional vehicle path planning;

[0050] Figure 2 is a flow chart of an obstacle detection method according to an embodiment of the present invention;

[0051] Figure 3 Schematic diagram of the obstacle detection process according to an embodiment of the present invention;

[0052] Figure 4 Schematic diagram of the obstacle recognition process based on a neural network model according to an embodiment of the present invention;

[0053] Figure 5 A schematic diagram of a state value freezing process based on a certain frame according to an embodiment of the present invention;

[0054] Figure 6 A schematic diagram of a multi-frame state value freezing process according to an embodiment of the present invention;

[0055] Figure 7 A flow chart of a vehicle path planning method according to an embodiment of the present invention;

[0056] Figure 8 Schematic diagram of obstacle outline position according to an embodiment of the present invention;

[0057] Figure 9 Schematic diagram of the structure of an obstacle detection device according to an embodiment of the present invention;

[0058] Figure 10A schematic structural diagram of a vehicle path planning device according to an embodiment of the present invention;

[0059] Figure 11 The figure is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0061] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0062] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; internal connections between two components; wireless connections or wired connections. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0063] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0064] Environmental perception is a prerequisite for the realization of autonomous driving systems and provides an effective basis for autonomous vehicle path planning. Stable and reliable obstacle perception technology plays a key role in the execution of autonomous vehicles such as obstacle stopping and obstacle avoidance. The stability of obstacle perception results directly affects the safety, operating efficiency and driving experience of the autonomous driving system. Figure 1As shown, when an autonomous driving system encounters a static obstacle blocking its path, it generally makes an obstacle avoidance decision. If the obstacle remains stable during the autonomous driving system's obstacle avoidance process, the autonomous driving system can successfully pass through it in one go. On the contrary, if the position and size of the obstacle perceived by the autonomous driving system fluctuate during the obstacle avoidance process, blocking the initial obstacle avoidance path, the autonomous driving system will need to update the obstacle avoidance path, which seriously affects planning efficiency and driving experience. Therefore, improving obstacle perception is essential for improving the performance of the entire autonomous driving system. Based on the characteristics of static obstacles, whose position and size remain unchanged, the embodiment of the present invention proposes the idea of ​​"freezing" static obstacles, thereby effectively avoiding the situation where multiple vehicle path planning operations are required due to the jitter of static obstacles.

[0065] The embodiment of the present invention provides an obstacle detection method, such as Figure 2 As shown, the obstacle detection method specifically includes the following steps:

[0066] Step S101: Obtain historical state values ​​of obstacles in a target area where a target vehicle is located within a certain period of time. The historical state values ​​of obstacles include position information and speed information of the obstacles.

[0067] Specifically, in practical applications, Figure 3 As shown, the embodiment of the present invention obtains the historical status values ​​of obstacles in the target area where the target vehicle is located within a certain period of time and caches them in the form of images. Since the target area may contain multiple obstacles, the embodiment of the present invention constructs a list of historical data for each obstacle in advance. By establishing the historical data list, the historical status values ​​of the obstacles are cached. The historical status values ​​of the obstacles include not only the location information and speed information of the obstacles, but also the ID, category, size and other information of the obstacles, thereby laying a data foundation for subsequent vehicle path planning.

[0068] Among them, the category of the obstacle refers to the type of obstacle, such as: motor vehicle, pedestrian, bicycle, cone barrel, etc.; the size refers to the physical size of the obstacle, including length, width, and height, and the unit can adopt the international standard unit system.

[0069] Specifically, in practical applications, in order to better match the image with the relevant numerical data, the time can be limited by limiting the number of image frames.

[0070] Specifically, in actual applications, the historical state value of the embodiment of the present invention is obtained by the upstream algorithm module, and its value can be a relative state value relative to the target vehicle or an absolute state value, which is not affected by the change of the posture of the target vehicle, that is, the position information and speed information in the embodiment of the present invention are absolute position and absolute speed.

[0071] In actual applications, when the obstacle state value is a relative state value, coordinate system conversion is required during the subsequent static obstacle recognition and planning process to convert the relative state value into an absolute coordinate value based on the absolute coordinate system. When the obstacle state value is an absolute state value, no conversion is required.

[0072] The embodiment of the present invention is described in detail using relative state values ​​as an example. For example, the absolute coordinate system can be the East-North-Sky coordinate system ENU (i.e., the station-centered coordinate system), and the ENU coordinate system is used for calibration to better understand the motion laws of other objects with the observer as the center.

[0073] Step S102: Based on the position information and speed information of the obstacle, determine whether the obstacle is a static obstacle.

[0074] Specifically, in practical applications, considering that there are errors in the obstacle status values ​​output by upstream algorithm modules such as obstacle detection and tracking, and that there are certain limitations in manually designed rules for determining whether an obstacle is static, the embodiment of the present invention uses a neural network model to determine whether an obstacle is a static obstacle.

[0075] like Figure 4 As shown in the figure, the neural network model can be constructed using M fully connected layers. The model input is obstacle information from the current frame to the previous K frames, and the model output is a flag indicating whether the obstacle is static. M can generally be set to 2. If the obstacle processing frame rate is 10, K is the target number of frames, generally set to a value between 5 and 15, corresponding to a time of 0.5s to 1.5s.

[0076] Each frame of obstacle information input to the model includes some of the following: position, orientation, speed, category, and size. Position, speed, and orientation are generally required, as they are the primary basis for determining whether an obstacle is static. Category information is used to account for the varying error distributions of upstream algorithms such as obstacle detection and tracking for different types of obstacles, such as vehicles and pedestrians. The variation in size information across multiple frames for the same obstacle indirectly reflects the error characteristics of the obstacle processing results of upstream algorithms such as obstacle detection and tracking, thus serving as a feature to provide additional basis for neural network model recognition. Position, speed, and orientation represent the absolute position, absolute speed, and absolute orientation of the obstacle, respectively. These data are unaffected by changes in the vehicle's position and posture, and the obstacle's position, orientation, and speed data in the ENU coordinate system can be used directly.

[0077] The fully connected layer in the model structure generally consists of a linear transformation W and a nonlinear activation function f. The function of the fully connected layer is to convert input features into output states. The weight parameters W contained in the fully connected layer are learned during the training phase and can be directly used during model inference. The training steps of a neural network model generally include: collecting and labeling data to obtain the model input data and the corresponding output true value; passing the input data into the neural network model to obtain the model prediction value, and calculating the loss function value based on the difference between the prediction value and the true value; calculating the gradient of the backpropagation based on the loss function value and updating the weight parameter W; iterating until the loss function value is less than a certain threshold or the number of iterations reaches a certain threshold. The training of the neural network model refers to the training process of the neural network model in the prior art and will not be repeated here.

[0078] After recognition by the neural network model, the output state can be an integer or Boolean value with a value of 0 or 1, representing the sign of whether it is a static obstacle or not; it can also be a decimal with a value of 0 to 1.0, representing the confidence level of a static obstacle. When the confidence level exceeds a certain threshold, it can be judged as a static obstacle.

[0079] The embodiment of the present invention judges obstacles based on state value information such as obstacle type, position, orientation angle, speed and size through a neural network model, shortening the recognition time and significantly reducing the probability of misjudgment.

[0080] Step S103: Based on the judgment result and the historical state value of the obstacle, the current state value of the obstacle is updated.

[0081] Specifically, in actual applications, there may be slight shaking of obstacles, and most existing technologies solve this problem by re-planning the vehicle path. However, this method is not only inefficient, but also seriously affects the user's driving experience.

[0082] The embodiment of the present invention first determines whether an obstacle is a static obstacle or a dynamic obstacle, and based on the determination result, promptly updates the historical state value of the obstacle, thereby laying the foundation for subsequent vehicle path planning.

[0083] Step S104: Detect the obstacle based on the current state value of the obstacle to obtain an obstacle detection result.

[0084] Specifically, in practical applications, the embodiments of the present invention will perform obstacle detection based on the updated current state value of the obstacle, while ensuring the accuracy of the obstacle detection result and providing data support for subsequent vehicle path planning.

[0085] By executing the above steps, the obstacle detection method provided by the embodiments of the present invention obtains the historical state value of the obstacle, providing preliminary data support for the subsequent comprehensive judgment of whether the obstacle is a static obstacle. By performing static and dynamic judgment on the obstacle and specifically updating the current state value of the obstacle, the obstacle is detected based on the updated current state value of the obstacle to obtain a detection result. By analyzing and judging the obstacle, misjudgment caused by slight vibration of the obstacle is avoided. While improving the static obstacle perception effect, the number of subsequent invalid vehicle path planning is significantly reduced, further improving production efficiency.

[0086] Specifically, in one embodiment, before executing the above step S102 to determine whether the obstacle is a static obstacle based on the position information and speed information of the obstacle, the following steps are further included:

[0087] Step S201: Calculate the distance of the obstacle relative to the target vehicle based on the position information of the obstacle and the current position of the target vehicle.

[0088] Step S202: Determine whether the distance between the obstacle and the target vehicle is less than a preset distance threshold.

[0089] Step S203: When the distance of the obstacle relative to the target vehicle is less than a preset distance threshold, coordinate axis conversion is performed on the position information of the obstacle to convert it into position information based on an absolute coordinate system.

[0090] Specifically, in actual applications, before determining whether the obstacle is a static obstacle, the embodiment of the present invention will calculate the distance between the obstacle and the target vehicle. Only when the distance between the obstacle and the target vehicle is less than the preset distance threshold d will the obstacle be determined to be within the target area, and then the coordinate system conversion of the position information and the subsequent related judgment process for static obstacles will be performed.

[0091] By setting a preset distance threshold, the embodiment of the present invention fully takes into account the situation where an obstacle is too far away from the target vehicle and basically has no impact on the autonomous driving system. When the distance between the obstacle and the target vehicle is less than the preset distance threshold, subsequent processing will not be performed temporarily, thereby saving computing resources and improving processing efficiency.

[0092] Specifically, different preset distance thresholds can be set for different obstacle categories. For example, for motor vehicles, d can be set to 80m; for pedestrians, d can be set to 60m, etc. Each preset distance threshold can be adjusted according to the actual operation scenario.

[0093] Specifically, in one embodiment, step S201 calculates the distance of the obstacle relative to the target vehicle based on the location information of the obstacle and the current location of the target vehicle, and specifically includes the following steps:

[0094] Step S301: extracting the coordinates of each inflection point of the space occupied by the obstacle from the position information of the obstacle.

[0095] Step S302: Determine the distance of the obstacle relative to the target vehicle based on the distance between the current position and the coordinates of each inflection point of the space occupied by the obstacle.

[0096] Specifically, in one embodiment, the above step S302 specifically includes the following steps:

[0097] Step S401: sorting the distances between the current position and the coordinates of each inflection point in the space occupied by the obstacle;

[0098] Step S402: Determine the minimum distance in the sorting results as the distance between the obstacle and the target vehicle.

[0099] Specifically, in practical applications, in order to better determine the distance between an obstacle and a target vehicle, an embodiment of the present invention sorts the distances between the current position and the coordinates of each inflection point in the space occupied by the obstacle, and determines the minimum distance in the sorting results as the distance between the obstacle and the target vehicle, thereby minimizing the changes in the target vehicle's path during driving and significantly improving the user's driving experience.

[0100] Specifically, in one embodiment, based on the converted position information and speed information of the obstacle, step S203 specifically includes the following steps:

[0101] Step S501: extracting the position information and speed information of the obstacle at the starting time and the current time within a certain period of time, respectively, to obtain a first position, a second position, a first speed, and a second speed.

[0102] Step S502: Calculate a first absolute position change of the obstacle based on the first position and the second position.

[0103] Step S503: Calculate the average absolute speed of the obstacle based on the first speed and the second speed.

[0104] Step S504: When the first absolute position change is smaller than the target position change, and / or the absolute speed average is smaller than the target speed average, the obstacle is determined to be a static obstacle.

[0105] Specifically, in practical applications, in order to effectively plan vehicle paths, embodiments of the present invention will determine whether a static obstacle is present based on the position and speed information of the target vehicle within a certain period of time. If at least one of the following constraints is met: the first absolute position change is less than the target position change, and the absolute speed average is less than the target speed average, the obstacle is determined to be a static obstacle.

[0106] Specifically, in one embodiment, the state value further includes the size and orientation angle of the obstacle. The above step S103 updates the current state value of the obstacle based on the judgment result and the historical state value of the obstacle, and specifically includes the following steps:

[0107] Step S601: When the obstacle is a static obstacle, the average values ​​of the position information, speed information, size, and orientation angle are calculated based on the historical state values ​​of the obstacle within a certain period of time.

[0108] Step S602: Determine the average value of the position information, speed information, size, and orientation angle as the static state value at the current moment, and update the static state value as the current state value of the obstacle.

[0109] Specifically, in practical applications, when the obstacle is a static obstacle, in order to avoid the automatic driving system from re-planning the driving path due to slight changes in the numerical values ​​of the relevant state values ​​caused by the jitter of the obstacle, the embodiment of the present invention outputs a fixed position, orientation angle, speed, and size of the static obstacle to achieve a similar freezing effect, wherein the position, orientation angle, and speed represent the absolute position, absolute orientation, and absolute speed of the obstacle respectively. The fixed values ​​of the position, orientation angle, speed, and size of the obstacle can be as follows: Figure 5 As shown, it is directly set to the corresponding data of a certain frame in the history of the obstacle; it can also be set as Figure 6 As shown in the figure, the obstacle is obtained by calculating the average value of the corresponding data of several frames in the history. Both methods can calculate the current state value of the static obstacle, reducing the number of static obstacle state value calculations while significantly improving processing speed and enhancing the user's driving experience.

[0110] Specifically, in one embodiment, the above step S103 further includes the following steps:

[0111] Step S701: When the obstacle is not a static obstacle, the position information, speed information, size, and orientation angle corresponding to the current moment are determined as the current state value of the obstacle.

[0112] Specifically, in actual applications, when the obstacle is not a static obstacle, it is necessary to obtain the relevant state value at the current moment according to the actual situation and update it in real time to avoid misjudgment of subsequent vehicle path planning due to incorrect data.

[0113] The embodiment of the present invention provides a vehicle path planning method, such as Figure 7 As shown, the vehicle path planning method specifically includes the following steps:

[0114] Step S801: Obstacle detection is performed using the above-mentioned obstacle detection method to obtain an obstacle detection result.

[0115] Step S802: Calculate the inflection point coordinates of the obstacle based on the obstacle detection result.

[0116] Step S803: Determine the spatial outline of the obstacle based on the inflection point coordinates of the obstacle.

[0117] Step S804: Determine the target vehicle's travel path based on the spatial contour of the obstacle.

[0118] Specifically, in actual applications, when the obstacle is a static obstacle, a fixed outline is output for the static obstacle to achieve a similar freezing effect. The outline is represented by the absolute position of the vertices of the polygon surrounding the obstacle. One way to represent the absolute position is the position in the ENU coordinate system. In order to improve the safety of obstacle avoidance, based on several frames of historical data of the obstacle, the outline vertex with the largest possible space occupied by the obstacle is estimated as the fixed value of the outline when freezing the obstacle. The effect is shown in the top view diagram as follows: Figure 8 As shown in FIG, the dotted box represents the bounding box corresponding to the obstacle at a certain moment, and the black solid line represents the outline of the obstacle estimated based on the obstacle bounding boxes at multiple moments.

[0119] The specific process of estimating the largest possible outline of an obstacle based on multi-frame data is as follows:

[0120] (1) Based on the cached obstacle position, size, and orientation angle data, calculate the vertex coordinates corresponding to the obstacle and cache the vertices together. Generally, each obstacle is represented by 8 vertices, representing the front upper left, front lower left, front upper right, front lower right, rear upper left, rear lower left, rear upper right, and rear lower right of the obstacle, a total of 8 vertices. Among them, the coordinates of the upper and lower vertices are different only in the height direction and are the same in the plane. Therefore, only the coordinates of the front lower left, front lower right, rear lower left, and rear lower right vertices need to be calculated. The coordinates of the other 4 vertices only need to add the height of the obstacle in the height direction;

[0121] (2) When it is determined that the obstacle needs to be frozen, the minimum and maximum coordinates of each vertex of the obstacle cached in (1) are obtained: if a certain vertex is closer to the vehicle in a certain direction on the plane, the minimum value of its coordinates in that direction in the history of multiple frames is obtained in that direction; otherwise, the maximum value of its coordinates in that direction in the history of multiple frames is obtained in that direction;

[0122] (3) After (2), the largest possible outline of the obstacle is obtained, which covers the maximum range of the space that the obstacle may occupy, thereby improving the safety of subsequent obstacle avoidance.

[0123] This embodiment of the present invention calculates the maximum possible spatial footprint of an obstacle based on several frames of historical data preceding the moment the obstacle freezes, providing sufficient safety distance for downstream planning modules to generate avoidance paths. This not only effectively reduces the number of obstacle state calculations, saving operating costs, but also further improves the accuracy of vehicle path planning and enhances the safety of the autonomous driving system when avoiding obstacles.

[0124] Specifically, when the obstacle is a non-static obstacle, the spatial contour of the obstacle will be determined based on the current state value of the obstacle.

[0125] It should be noted that there are two types of non-static obstacles in the embodiments of the present invention. One is that the obstacle is always in motion, in which case the state value of the obstacle needs to be the tracking result at the current moment; the other is that the obstacle changes from a stationary state to a moving state, in which case the fixed position, orientation angle, speed, and size of the obstacle are no longer output at this moment, and are switched to a real-time update state to achieve a defrosting effect.

[0126] The embodiments of the present invention use multi-dimensional information such as obstacle category, continuous multi-frame position, orientation angle, speed, and size as input features to identify static obstacles based on a neural network model. This avoids the limitations of methods based on manually designed rules for judgment, improves the accuracy of static obstacle identification, and specifically adjusts the data processing process based on the judgment results to calculate the maximum possible outline of the obstacle's space occupation, providing sufficient safety distance for the downstream planning module to generate an obstacle avoidance path. This not only improves path planning efficiency, but also further enhances the safety of the autonomous driving system when avoiding obstacles, and improves the user's driving experience.

[0127] Specifically, in actual applications, the target vehicle's driving path will be determined based on the current state value of the obstacle. Therefore, when the obstacle is a static obstacle, there is no need to re-plan the target vehicle's driving path multiple times, thereby improving the efficiency of vehicle path planning.

[0128] By executing the above steps, the vehicle path planning method provided by the embodiment of the present invention determines the driving path of the target vehicle based on the obstacle detection results. While ensuring that the planned vehicle path effectively avoids obstacles, it further analyzes and judges the obstacles to avoid misjudgments due to slight jitters of obstacles, significantly reducing the number of invalid vehicle path planning times, improving planning efficiency and the user's driving experience.

[0129] Obstacle avoidance planning, also known as local path planning or dynamic path planning, primarily involves detecting obstacles and tracking their movement trajectories, thereby estimating the next movement position of the obstacles and ultimately creating an obstacle map that includes existing and potential collision risks. Obstacle avoidance planning is a crucial step in the autonomous vehicle's driving planning process. Based on the obstacle avoidance planning results, vehicle trajectory planning is performed. Based on the autonomous vehicle's kinematic model and kinematic constraints, appropriate control commands are sought to convert feasible paths into feasible trajectories, enabling the autonomous vehicle to quickly and smoothly reach its target location. The present invention sequentially performs obstacle detection and vehicle path planning. Based on a deep understanding of the target area, particularly dynamic target areas containing dynamic obstacles, this ensures accurate obstacle detection within the target area where the autonomous vehicle resides. Furthermore, vehicle path planning is performed based on the target vehicle's kinematic model and constraints, and the target path of the target vehicle is plotted on a timeline. This significantly reduces the number of invalid vehicle path planning attempts, allowing the target vehicle to smoothly and efficiently reach its designated target location. This completes the entire process of obstacle avoidance planning, trajectory planning, and path planning, which is of great significance for the global navigation planning of autonomous vehicles.

[0130] The embodiment of the present invention provides an obstacle detection device, such as Figure 9 As shown, the obstacle detection device includes:

[0131] The acquisition module 101 is used to obtain the historical state value of obstacles in the target area where the target vehicle is located within a certain period of time. The historical state value of the obstacles includes the location information and speed information of the obstacles. For details, please refer to the relevant description of step S101 in the above method embodiment, which will not be repeated here.

[0132] The first processing module 102 is used to determine whether the obstacle is a static obstacle based on the position information and speed information of the obstacle. For details, please refer to the relevant description of step S102 in the above method embodiment, which will not be repeated here.

[0133] The second processing module 103 is used to update the current state value of the obstacle based on the judgment result and the state value of the obstacle. For details, please refer to the relevant description of step S103 in the above method embodiment, which will not be repeated here.

[0134] The third processing module 104 is used to detect the obstacle based on the current state value of the obstacle and obtain the obstacle detection result. For details, please refer to the relevant description of step S104 in the above method embodiment, which will not be repeated here.

[0135] For further description of the above-mentioned obstacle detection device, please refer to the relevant description of the above-mentioned obstacle detection method embodiment, which will not be repeated here.

[0136] Through the collaborative work of the aforementioned components, the obstacle detection device provided by the embodiments of the present invention obtains the historical state value of an obstacle, providing preliminary data support for the subsequent comprehensive judgment of whether the obstacle is a static obstacle. By performing static and dynamic judgments on the obstacle and specifically updating the current state value of the obstacle, the obstacle is detected based on the updated current state value of the obstacle to obtain a detection result. By analyzing and judging the obstacle, misjudgments caused by slight jitter of the obstacle are avoided. While improving the static obstacle perception effect, the number of subsequent invalid vehicle path planning is significantly reduced, further improving production efficiency.

[0137] The embodiment of the present invention provides a vehicle path planning device, such as Figure 10 As shown, the vehicle path planning device includes:

[0138] The detection module 801 is used to perform obstacle detection using the above-mentioned obstacle detection device to obtain the obstacle detection result. For details, please refer to the relevant description of step S801 in the above-mentioned method embodiment, which will not be repeated here.

[0139] The first calculation module 802 is used to calculate the inflection point coordinates of the obstacle based on the obstacle detection result. For details, please refer to the relevant description of step S802 in the above method embodiment, which will not be repeated here.

[0140] The second calculation module 803 is used to determine the spatial contour of the obstacle based on the inflection point coordinates of the obstacle. For details, please refer to the relevant description of step S803 in the above method embodiment, which will not be repeated here.

[0141] The third calculation module 804 is used to determine the driving path of the target vehicle based on the spatial contour of the obstacle. For details, please refer to the relevant description of step S804 in the above method embodiment, which will not be repeated here.

[0142] For further description of the above-mentioned vehicle path planning device, please refer to the relevant description of the above-mentioned vehicle path planning method embodiment, which will not be repeated here.

[0143] Through the coordinated cooperation of the above-mentioned components, the vehicle path planning device provided in the embodiment of the present invention determines the driving path of the target vehicle based on the obstacle detection results. While ensuring that the planned vehicle path effectively avoids obstacles, it further analyzes and judges the obstacles to avoid misjudgments due to slight shaking of obstacles, significantly reducing the number of invalid vehicle path planning times, improving planning efficiency and the user's driving experience.

[0144] An embodiment of the present invention provides an electronic device, which can be a domain controller mounted on a vehicle or a cloud server. When the electronic device is a domain controller mounted on a vehicle, the domain controller collects the current driving scene corresponding to the vehicle, performs global reference path planning based on the current driving scene, and can then control the vehicle to travel based on the global reference path. When the electronic device is a cloud server, the cloud server can communicate with the vehicle, obtain the vehicle's current driving scene through a current driving scene collection device provided on the vehicle, and then perform global reference path planning in the cloud. The planned global reference path is then sent back to the vehicle controller, so that the vehicle controller can control the vehicle's travel based on the global reference path.

[0145] like Figure 11 As shown, the electronic device includes a processor 901 and a memory 902, wherein the memory 902 and the processor 901 are communicatively connected to each other, wherein the processor 901 and the memory 902 can be connected via a bus or other means. Figure 11 The bus connection is taken as an example.

[0146] The processor 901 may be a central processing unit (CPU). The processor 901 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0147] Memory 902, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of the present invention. Processor 901 executes the non-transitory software programs, instructions, and modules stored in memory 902 to perform various functional applications and data processing of processor 901, thereby implementing the methods in the above-mentioned method embodiments.

[0148] The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created by the processor 901, etc. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 902 may optionally include a memory remotely located relative to the processor 901, and these remote memories may be connected to the processor 901 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.

[0149] One or more modules are stored in the memory 902 and, when executed by the processor 901 , perform the method in the above method embodiment.

[0150] The specific details of the above electronic device can be understood by referring to the corresponding descriptions and effects in the above method embodiments, and will not be repeated here.

[0151] Those skilled in the art will appreciate that all or part of the processes in the above-described embodiments can be implemented by instructing related hardware through a computer program. The implemented program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-described embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD). The storage medium can also include a combination of the above-mentioned types of memory.

[0152] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. An obstacle detection method, characterized in that: include: Obtaining historical state values ​​of obstacles in a target area where the target vehicle is located within a certain period of time, wherein the historical state values ​​of the obstacles include position information and speed information of the obstacles; Determining whether the obstacle is a static obstacle based on the position information and speed information of the obstacle; Based on the judgment result and the historical state value of the obstacle, updating the current state value of the obstacle; The state value further includes a size and an orientation angle of the obstacle. Updating the current state value of the obstacle based on the judgment result and the state value of the obstacle includes: when the obstacle is a static obstacle, calculating an average of the position information, the speed information, the size, and the orientation angle based on historical state values ​​of the obstacle over a certain period of time; determining the average of the position information, the speed information, the size, and the orientation angle as the static state value at a current moment, updating the static state value as the current state value of the obstacle, and outputting a fixed outline for the static obstacle; The obstacle is detected based on the current state value of the obstacle to obtain an obstacle detection result.

2. The method according to claim 1, characterized in that Before determining whether the obstacle is a static obstacle based on the position information and speed information of the obstacle, the method further includes: Calculating a distance of the obstacle from the target vehicle based on the position information of the obstacle and the current position of the target vehicle; Determining whether the distance between the obstacle and the target vehicle is less than a preset distance threshold; When the distance between the obstacle and the target vehicle is less than the preset distance threshold, the position information of the obstacle is subjected to coordinate axis conversion to be converted into position information based on an absolute coordinate system.

3. The method according to claim 2, characterized in that Determining whether the obstacle is a static obstacle based on the converted position information and speed information of the obstacle includes: Extracting the position information and speed information of the obstacle at the starting time and the current time within the certain period of time respectively to obtain a first position, a second position, a first speed, and a second speed; Calculating a first absolute position change of the obstacle based on the first position and the second position; Calculating an average absolute speed of the obstacle based on the first speed and the second speed; When the first absolute position change is smaller than the target position change, and / or the absolute speed average is smaller than the target speed average, it is determined that the obstacle is a static obstacle.

4. The method according to claim 1, wherein When the obstacle is not a static obstacle, the position information, speed information, size, and orientation angle corresponding to the current moment are determined as the current state value of the obstacle.

5. The method according to claim 2, characterized in that The calculating the distance of the obstacle relative to the target vehicle based on the position information of the obstacle and the current position of the target vehicle includes: Extracting the coordinates of each inflection point of the space occupied by the obstacle from the position information of the obstacle; Determining the distance of the obstacle relative to the target vehicle based on the distance between the current position and the coordinates of each inflection point of the space occupied by the obstacle; determining the distance of the obstacle relative to the target vehicle based on the distance between the current position and the coordinates of each inflection point of the space occupied by the obstacle, including: sorting the distances between the current position and the coordinates of each inflection point of the space occupied by the obstacle; The minimum distance in the sorting results is determined as the distance of the obstacle relative to the target vehicle.

6. A vehicle path planning method, characterized in that: include: Performing obstacle detection using the obstacle detection method according to any one of claims 1 to 5 to obtain an obstacle detection result; Calculating the inflection point coordinates of the obstacle based on the obstacle detection result; Determining a spatial contour of the obstacle based on the inflection point coordinates of the obstacle; Based on the spatial contour of the obstacle, a driving path of the target vehicle is determined.

7. An obstacle detection device, characterized in that: include: An acquisition module is used to obtain the historical state value of the obstacle in the target area where the target vehicle is located within a certain period of time, wherein the historical state value of the obstacle includes the position information and speed information of the obstacle; a first processing module, configured to determine whether the obstacle is a static obstacle based on the position information and speed information of the obstacle; a second processing module, configured to update a current state value of the obstacle based on the judgment result and the historical state value of the obstacle; The state value further includes a size and an orientation angle of the obstacle. Updating the current state value of the obstacle based on the judgment result and the state value of the obstacle includes: when the obstacle is a static obstacle, calculating an average of the position information, the speed information, the size, and the orientation angle based on historical state values ​​of the obstacle over a certain period of time; determining the average of the position information, the speed information, the size, and the orientation angle as the static state value at a current moment, updating the static state value as the current state value of the obstacle, and outputting a fixed outline for the static obstacle; The third processing module is configured to detect the obstacle based on the current state value of the obstacle to obtain an obstacle detection result.

8. A vehicle path planning device, characterized in that: include: a detection module, configured to perform obstacle detection using the obstacle detection device according to claim 7 and obtain an obstacle detection result; A first calculation module, configured to calculate the coordinates of the inflection point of the obstacle based on the obstacle detection result; a second calculation module, configured to determine a spatial contour of the obstacle based on the inflection point coordinates of the obstacle; The third calculation module is used to determine the driving path of the target vehicle based on the spatial contour of the obstacle.

9. An electronic 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 method according to any one of claims 1 to 6 by executing the computer instructions.

10. 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 method according to any one of claims 1 to 6.

Images