Unmanned driving control method and system combined with dynamic obstacle trajectory prediction

Abstract

The invention relates to an unmanned driving control method and system based on dynamic obstacle trajectory prediction, and relates to the technical field of unmanned driving, and the method comprises the steps: detecting radar point cloud time sequence information in a preset radius range through a millimeter wave radar disposed on a vehicle after the vehicle is started; performing moving target identification on the radar point cloud time sequence information to obtain moving target positioning time sequence information; performing moving state analysis on the moving target positioning time sequence information to obtain a moving rule bipartite parameter, if the moving rule bipartite parameter is equal to 0, summarizing the moving target as a type of moving target, and calling a terrain grid model; and carrying out terrain guide trajectory analysis on the moving target based on the terrain grid model, and obtaining predicted position time sequence information of the moving target to carry out unmanned driving control. The technical problem that the avoidance rate of unpredictable obstacles is not high due to the lack of training standard data in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of unmanned driving technology, and in particular to an unmanned driving control method and system combined with dynamic obstacle trajectory prediction. Background Art

[0002] In the field of autonomous driving technology, vehicles need to process a large amount of environmental information in real time during driving to ensure safe driving. Traditionally, autonomous vehicles rely on a variety of sensors, such as cameras, LiDAR, and millimeter-wave radar, to detect and identify surrounding obstacles. However, for dynamic obstacles, especially those that do not follow traffic rules or are unpredictable (such as stones), traditional prediction methods may not provide sufficient avoidance rates.

[0003] Existing unmanned driving systems often rely on machine learning and pattern recognition technologies when dealing with dynamic obstacles, which require a large amount of labeled data to train the model. However, it is very difficult to obtain sufficient training data for unpredictable obstacles such as stones on mountain roads, resulting in a low avoidance rate for unpredictable obstacles. Summary of the invention

[0004] The present invention aims to solve the technical problem in the prior art that the avoidance rate of unpredictable obstacles is low due to the lack of training standard data, and provides an unmanned driving control method and system combined with dynamic obstacle trajectory prediction to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: In the first aspect, the present invention provides an unmanned driving control method combined with dynamic obstacle trajectory prediction, including: when the vehicle is started, the radar point cloud timing information within a preset radius is detected by a millimeter wave radar deployed on the vehicle; the mobile target is identified on the radar point cloud timing information to obtain the mobile target positioning timing information; the mobile state analysis is performed on the mobile target positioning timing information to obtain the movement law binary parameter; when the movement law binary parameter is equal to 0, the mobile target is classified into a type of mobile target, and at the same time, the terrain grid model within the preset radius is retrieved through the road network map; based on the terrain grid model, the mobile target is analyzed for terrain guidance trajectory to obtain the mobile target predicted position timing information; and the unmanned driving control is performed according to the mobile target predicted position timing information.

[0006] In the second aspect, the present invention provides an unmanned driving control system combined with dynamic obstacle trajectory prediction, including: a radar detection unit, which is used to detect radar point cloud timing information within a preset radius through a millimeter-wave radar deployed on the vehicle after the vehicle is started; a mobile target identification unit, which is used to perform mobile target identification on the radar point cloud timing information to obtain mobile target positioning timing information; a mobile state analysis unit, which is used to perform mobile state analysis on the mobile target positioning timing information to obtain a mobile law binary parameter; a terrain model retrieval unit, which is used to classify the mobile target into a type of mobile target when the mobile law binary parameter is equal to 0, and at the same time, retrieve the terrain grid model within a preset radius through a road network map; a target movement prediction unit, which is used to perform terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain mobile target predicted position timing information; an unmanned driving control unit, which is used to perform unmanned driving control according to the mobile target predicted position timing information.

[0007] In a third aspect, the present invention provides an electronic device comprising: a memory for storing a computer software program; and a processor for reading and executing the computer software program, thereby implementing an unmanned driving control method combined with dynamic obstacle trajectory prediction as described in the first aspect.

[0008] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium, in which a computer software program is stored. When the computer software program is executed by a processor, an unmanned driving control method combined with dynamic obstacle trajectory prediction described in the first aspect is implemented.

[0009] The beneficial effects of the present invention are: by using millimeter wave radar to detect the surrounding environment, dynamic obstacles are detected and identified in real time, and for irregular moving targets, such as stones, the possible position time series information is predicted through terrain guidance trajectory analysis. This method does not rely on a large amount of training data, but uses a terrain grid model to assist in prediction, thereby improving the avoidance rate of unpredictable obstacles. In this way, it can respond to emergencies more effectively, achieving the technical effect of improving the safety and reliability of unmanned vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic flow chart of an unmanned driving control method combined with dynamic obstacle trajectory prediction provided by the present invention; Figure 2 A schematic diagram of the structure of an unmanned driving control system combined with dynamic obstacle trajectory prediction provided by the present invention; Figure 3 A schematic diagram of the structure of an electronic device provided by the present invention; Figure 4A schematic diagram of the structure of a computer-readable storage medium provided by the present invention.

[0011] In the accompanying drawings, the components represented by the reference numerals are described as follows: Electronic device 500 , memory 510 , processor 520 , first computer program 511 , computer-readable storage medium 600 , second computer program 611 . DETAILED DESCRIPTION

[0012] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0013] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0014] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0015] Embodiment 1: like Figure 1 As shown, an embodiment of the present invention provides an unmanned driving control method combined with dynamic obstacle trajectory prediction, comprising the steps of: S10: When the vehicle is started, the radar point cloud time series information within a preset radius is detected by a millimeter wave radar deployed on the vehicle; Specifically, the preset radius range refers to the distance threshold for screening point clouds. When the point cloud is larger than the preset radius range, it is deemed to be temporarily posing no threat to the vehicle and does not need to be considered at present. In detail, the radar point cloud timing information refers to the point cloud distribution data detected by the millimeter-wave radar at continuous moments within the preset radius range.

[0016] The detection steps are as follows: After the vehicle is started, the millimeter-wave radar system starts working and prepares to collect data; the radar transmits millimeter waves and receives the reflected signals. By processing the received signals, the radar can calculate the distance, speed, angle and other information of the object and convert them into point cloud data; according to the preset radius range, the point cloud whose distance from the radar is greater than the preset radius range will be deleted; the remaining point clouds are stored in time sequence to obtain the radar point cloud time sequence information.

[0017] Millimeter-wave radar provides unmanned vehicles with precise environmental perception capabilities, enabling them to understand and predict dynamic changes in the surrounding environment. Through the point cloud time series information collected by millimeter-wave radar, the vehicle can detect and track dynamic obstacles and obtain the movement information of obstacles, which is crucial for the subsequent prediction of obstacle movement trajectories and the generation of obstacle avoidance paths.

[0018] S20: performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information; Specifically, mobile target recognition refers to the process of distinguishing stationary objects from moving objects from these continuous point cloud data, with the aim of identifying which point clouds represent moving targets; mobile target positioning timing information refers to the data obtained by continuously recording the changes in the position and motion status of these targets over time after the mobile targets are identified.

[0019] The specific process of mobile target recognition includes: extracting features that help identify mobile targets, such as speed, acceleration, shape and size, from point cloud data; using the extracted features to analyze the mobile targets and static background; tracking the detected mobile targets, recording their position and motion state changes over time, and generating mobile target positioning timing information.

[0020] Providing accurate information about surrounding moving targets to unmanned vehicles is crucial for safe driving of the vehicle. Through mobile target recognition and positioning, the vehicle can identify and distinguish dynamic and static obstacles around the vehicle, providing a basis for obstacle avoidance and path planning.

[0021] S30: Analyze the movement state of the moving target positioning time sequence information to obtain the binary parameters of the movement law; Furthermore, the moving state analysis is performed on the moving target positioning time sequence information to obtain the moving law binary parameters, including: Analyze the first speed direction and the first speed magnitude of the first moment positioning information and the second moment positioning information of the mobile target positioning timing information; Until analyzing the N-1th moment positioning information of the mobile target positioning timing information and the N-1th speed direction and the N-1th speed magnitude of the N-1th moment positioning information; Calculating a first deviation angle between the first speed direction and the second speed direction, until calculating an N-2th deviation angle between the N-2th speed direction and the N-1th speed direction, and counting first variance parameters of the first deviation angles until the N-2th deviation angle; Calculating a first deviation between the first speed magnitude and the second speed magnitude, until calculating an N-2th deviation between the N-2th speed magnitude and the N-1th speed magnitude, and counting second variance parameters of the first deviation magnitudes until the N-2th deviation magnitude; When the first variance parameter is greater than a first variance parameter threshold, or / and the second variance parameter is greater than a second variance parameter threshold, the moving law binary parameter is equal to 0; Otherwise, the movement law binary parameter is equal to 1.

[0022] Specifically, the moving law binary parameter refers to a parameter obtained from the moving state analysis, which divides the moving law of the moving target into two categories, for example, one category is the target that follows specific traffic rules, and the other category is the target that does not follow the rules or the target whose behavior pattern is difficult to predict (such as obstacles such as stones). Moving state analysis refers to analyzing the moving regularity of the moving target based on the positioning timing information of the moving target to determine whether it is a predictable dynamic obstacle. Preferably, when it is a target that does not follow the rules or the target whose behavior pattern is difficult to predict, the moving law binary parameter is equal to 0, otherwise, the moving law binary parameter is equal to 1.

[0023] When the binary parameter of the movement law is equal to 1, the existing algorithm for predicting the trajectory of the target can be used to predict and avoid the target trajectory. When the binary parameter of the movement law is equal to 0, the algorithm steps set in the embodiment of the present application need to be executed to predict and avoid the target trajectory.

[0024] In detail, the steps of mobile status analysis include: By comparing the positioning information of two consecutive moments, the speed direction and speed magnitude of the moving target are analyzed. The speed direction refers to the direction of the target's movement, while the speed magnitude refers to the distance the target moves per unit time. The speed direction of the positioning information of two consecutive moments refers to the azimuth vector pointing from the target position of the previous moment to the target position of the next moment. The speed magnitude of the positioning information of two consecutive moments refers to the absolute value of the distance between the target position of the previous moment and the target position of the next moment.

[0025] Based on the above principle, the first moment positioning information and the second moment positioning information of the mobile target positioning timing information are extracted, and then the first speed direction and the first speed magnitude from the first moment positioning information to the second moment positioning information are analyzed; further, the second moment positioning information and the third moment positioning information of the mobile target positioning timing information are extracted, and then the second speed direction and the second speed magnitude from the second moment positioning information to the third moment positioning information are analyzed… until the N-1th moment positioning information and the N-1th speed direction and the N-1th speed magnitude of the N-1th moment positioning information of the mobile target positioning timing information are extracted, where N represents the total number of moments.

[0026] Then, a first deviation angle between the first speed direction and the second speed direction is calculated until an N-2nd deviation angle between the N-2nd speed direction and the N-1st speed direction is calculated, and variance calculation is performed on the first deviation angles until the N-2nd deviation angles to obtain a first variance parameter.

[0027] And, calculate the first deviation size between the first speed size and the second speed size, until calculate the N-2nd deviation size between the N-2nd speed size and the N-1st speed size, and calculate the variance of the first deviation size to the N-2nd deviation size to obtain a second variance parameter.

[0028] By counting a series of variance parameters of deviation angles and deviation sizes, the stability and regularity of the target motion state can be quantified. The variance parameter is a statistic that measures the degree of data dispersion. The larger the variance, the greater the data fluctuation. The greater the fluctuation, the weaker the movement regularity. When any of the first variance parameter is greater than the first variance parameter threshold and the second variance parameter is greater than the second variance parameter threshold, the movement regularity binary parameter is equal to 0. The first variance parameter threshold and the second variance parameter threshold are parameters that can be customized by the user for specific applications. The larger the first variance parameter threshold and the second variance parameter threshold, the greater the user's tolerance for irregularly moving targets, and vice versa. Preferably, In urban and plain areas, the first variance parameter threshold and the second variance parameter threshold are set larger because the probability of uncontrollable dynamic obstacles appearing in these areas is low, while in rural and mountainous areas, the first variance parameter threshold and the second variance parameter threshold are set smaller to ensure driving safety. In the above screening scheme, some non-threatening moving targets will also be classified as uncontrollable dynamic obstacles. This is to sacrifice accuracy to ensure the safety of unmanned driving. In other cases, that is, when the first variance parameter is less than or equal to the first variance parameter threshold, and the second variance parameter is less than or equal to the second variance parameter threshold, the moving law binary parameter is equal to 1, indicating that the moving target is a regular moving obstacle, and the existing trajectory prediction algorithm can be executed.

[0029] By providing the unmanned driving system with quantitative analysis of the motion status of surrounding mobile targets, the system can make reasonable path planning and obstacle avoidance decisions based on the motion laws of the targets.

[0030] S40: when the binary parameter of the movement rule is equal to 0, the moving target is classified into one type of moving target, and at the same time, a terrain grid model within a preset radius is retrieved through a road network map; S50: performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain time series information of the predicted position of the mobile target; Further, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information, which includes: Performing terrain guidance weight analysis on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient; When the terrain guidance weight coefficient is greater than or equal to the terrain guidance weight coefficient threshold, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information; When the terrain guidance weight coefficient is less than the terrain guidance weight coefficient threshold, a safety warning signal is generated according to the real-time position of the mobile target to warn the central control platform.

[0031] Further, a terrain guidance weight analysis is performed on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient, including: Performing terrain guidance trajectory analysis on the mobile target positioning information at the first moment of the mobile target positioning time series information based on the terrain grid model to obtain the mobile target terrain guidance positioning information at the second moment; Until the terrain guidance trajectory analysis is performed on the mobile target positioning information at the N-1th moment of the mobile target positioning time series information based on the terrain grid model, to obtain the mobile target terrain guidance positioning information at the Nth moment; Calculate a first positioning distance between the terrain guidance positioning information of the mobile target at the second moment and the positioning information of the mobile target at the second moment; Until calculating the Nth positioning distance between the terrain guidance positioning information of the mobile target at the Nth moment and the positioning information of the mobile target at the Nth moment; The proportion of the first positioning distances to the Nth positioning distance that is less than or equal to the positioning distance threshold is counted and set as the terrain guidance weight coefficient.

[0032] Furthermore, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information, including: Obtain the real-time position and speed of the moving target; Inputting the real-time position of the mobile target and the terrain grid model into a mobile trajectory matching model for analysis to obtain a mobile prediction trajectory; Obtaining a terrain height sequence of the predicted movement trajectory; Inputting the real-time speed of the moving target and the terrain height sequence into a moving speed matching model for analysis to obtain a moving speed sequence; The moving speed sequence and the moving predicted trajectory are fused to obtain the moving target predicted position time series information.

[0033] Specifically, when the binary parameter of the moving rule is equal to 0, it indicates that the moving target is classified as a special type of moving target, which may not follow conventional traffic rules or behavior patterns, such as obstacles such as stones. In view of such a situation, the embodiment of the present application analyzes a type of moving target through a customized trajectory prediction algorithm, as follows: The road network map represents a big data map, which stores the terrain map of the vehicle driving area and selects the terrain grid model within the preset radius. The terrain grid model refers to the terrain map model obtained by dividing the terrain map according to the preset grid. Since most of the obstacles that do not follow the convention are objects and cannot lead the direction of movement by themselves, the only trajectory influence comes from gravity and terrain. Therefore, the movement of the moving target can be analyzed based on the terrain and gravity, and then the predicted position time series information of the moving target in the future time interval can be specified.

[0034] Specifically, the terrain-guided trajectory analysis process is as follows: The real-time position of the mobile target and the real-time position of the mobile target refer to the real-time dynamic data obtained by continuously tracking and locating the mobile target through the millimeter wave radar. The mobile trajectory matching model is used to combine the real-time position and historical trajectory of the mobile target with the terrain grid model to predict the future movement trajectory of the target; the mobile prediction trajectory is the future movement trajectory of the mobile target.

[0035] The terrain height sequence refers to the terrain height data sequence that corresponds one-to-one to the position of the mobile prediction trajectory; the mobile speed matching model is used to analyze the relationship between the real-time speed of the mobile target and the terrain height sequence to predict the speed change of the target on different terrains; the mobile speed sequence is the speed change sequence of the mobile target on the predicted trajectory obtained by analyzing the mobile speed matching model.

[0036] The predicted position time series information of the moving target refers to the predicted position information of the moving target at each future time point obtained by fusing the moving speed sequence and the moving predicted trajectory. If the speed information of the moving target at any two points is known, the predicted position information of the moving target on the trajectory can be analyzed.

[0037] Specifically, the mobile trajectory matching model and the mobile speed matching model are both long short-term memory neural network topology structures. Preferably, the training steps of the mobile trajectory matching model include: collecting mobile target position record data, terrain grid record model and mobile trajectory identification data; constructing a mobile trajectory matching loss function: , a>1, Characterize the matching trajectory, Characterize the identification trajectory, Characterize the area enclosed by the matching trajectory and the identification trajectory, Characterize the mobile trajectory matching loss value, use the mobile trajectory identification data as supervision, use the mobile target position record data and the terrain grid record model as input, and train the long short-term memory neural network topology structure. The number of consecutive losses is less than or equal to The loss threshold is used to generate a mobile trajectory matching model. The training steps of the mobile speed matching model include: collecting mobile speed sequence record data, mobile prediction trajectory record data and mobile target prediction position time series identification data; constructing a mobile speed matching loss function: , Characterizes the mobile speed matching loss value, Represents the distance between the predicted position at the i-th moment and the marked position at the i-th moment, represents the predicted position at the i-th moment, Represents the position of the marker at the i-th moment, Represents the total number of preset moments, b>1. The mobile speed sequence record data and mobile prediction trajectory record data are used as the input of the long short-term memory neural network, and the mobile target prediction position time series identification data is used as supervision for training. The number of consecutive losses is less than or equal to The loss threshold is used to generate a mobile speed matching model. The mobile speed matching model and the mobile trajectory matching model are used to realize the fusion analysis of the motion state of the moving target and the terrain, and the motion trajectory time sequence information of obstacles such as stones is obtained, which provides a basis for effective obstacle avoidance in the next step.

[0038] In a preferred embodiment, the terrain guidance weight coefficient refers to the degree of influence of terrain factors on the prediction of the trajectory of the mobile target. If the terrain guidance weight coefficient is less than the terrain guidance weight coefficient threshold preset by the user, it means that the mobile target and the terrain are not closely related. At this time, it is meaningless to use the terrain guidance prediction trajectory. In this case, the nature of the mobile target is unclear, and a safety warning signal is generated according to the real-time position of the mobile target to warn the central control platform and remind the user to make manual decisions. When the terrain guidance weight coefficient is greater than or equal to the terrain guidance weight coefficient threshold, the terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information.

[0039] In a preferred embodiment, the detailed configuration process of the terrain guidance weight coefficient is as follows: Performing the terrain guidance trajectory analysis on the mobile target positioning information at the first moment of the mobile target positioning time series information based on the terrain grid model to obtain the mobile target terrain guidance positioning information at the second moment; until performing the terrain guidance trajectory analysis on the mobile target positioning information at the N-1th moment of the mobile target positioning time series information based on the terrain grid model to obtain the mobile target terrain guidance positioning information at the Nth moment; Then, the Euclidean distance between the terrain-guided positioning information of the mobile target at the second moment and the positioning information of the mobile target at the second moment is calculated, and stored as the first positioning distance; until the Euclidean distance between the terrain-guided positioning information of the mobile target at the Nth moment and the positioning information of the mobile target at the Nth moment is calculated, and stored as the Nth positioning distance; the proportion of the positioning distances less than or equal to the positioning distance threshold in the first positioning distance until the Nth positioning distance is counted, and set as the terrain guidance weight coefficient.

[0040] S60: Performing unmanned driving control according to the predicted position timing information of the moving target.

[0041] Specifically, after determining the timing information of the predicted position of the moving target, when it intersects with the preset path of the unmanned driving, deceleration, acceleration or detour avoidance path planning is performed for control.

[0042] Further, performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information includes: Traversing the radar point cloud time series information to perform outlier analysis to obtain a dense radar point cloud time series diagram; Segmenting the dense radar point cloud time series graph according to a first segmentation step to obtain a first dense radar point cloud time series graph; Segmenting the dense radar point cloud time series graph according to one eighth of the first segmentation step length to obtain a second dense radar point cloud time series graph; The first dense radar point cloud timing diagram is input into a mobile target positioning channel of a mobile target recognition network, and the second dense radar point cloud timing diagram is input into a mobile target segmentation channel of the mobile target recognition network for processing to obtain the mobile target positioning timing information.

[0043] Specifically, in order to achieve rapid sorting of lightning point cloud data, the embodiment of the present application processes the radar point cloud through the SLOWFAST neural network model to identify the positioning timing information of the mobile target. The preferred process is as follows: The outliers of each point at each moment of the radar point cloud time series information are deleted according to the position distribution, and the concentrated point cloud at each moment is obtained and stored as a dense radar point cloud time series diagram. The outlier analysis algorithm can be implemented using the local outlier factor (LOF) algorithm, which calculates the outlier factor of each point at a certain moment, deletes the points whose outlier factor is greater than the outlier factor threshold, and the remaining points are regarded as the dense radar point cloud at that moment.

[0044] Then, the dense radar point cloud timing diagram is segmented according to one eighth of the first segmentation step length to obtain a second dense radar point cloud timing diagram. Specifically, the unit of the first segmentation step length is milliseconds. The dense radar point cloud timing diagram is segmented according to one eighth of the first segmentation step length to obtain a second dense radar point cloud timing diagram. The first dense radar point cloud timing diagram is input into the mobile target positioning channel of the mobile target recognition network, and the second dense radar point cloud timing diagram is input into the mobile target segmentation channel of the mobile target recognition network for processing to obtain the mobile target positioning timing information. The mobile target positioning channel is a SLOW channel to achieve the positioning of the target, and the mobile target segmentation channel is a FAST channel to achieve fast tracking of the target.

[0045] Furthermore, the steps of constructing the mobile target recognition network include: Collect radar point cloud time series diagram record data and target position identification time series information; Segmenting the radar point cloud time sequence diagram record data according to a first segmentation step length to obtain first radar point cloud time sequence diagram record data; Segmenting the radar point cloud time series graph record data according to one eighth of the first segmentation step length to obtain second radar point cloud time series graph record data; Input the first radar point cloud time series diagram record data into the SLOW channel, and input the second radar point cloud time series diagram record data into the FAST channel to obtain target position prediction time series information; The proportion of moments in which the position deviation distance in the position deviation distance time series information of the target position identification time series information and the target position prediction time series information is greater than or equal to the position deviation distance threshold is taken as the mobile target recognition loss value; When the mobile target recognition loss value for a preset number of consecutive trainings is less than or equal to the mobile target recognition loss value threshold, the mobile target recognition network is generated.

[0046] Specifically, the radar point cloud timing diagram record data and the target position identification timing information are both training data pre-identified by the user. The radar point cloud timing diagram record data is segmented according to the first segmentation step length to obtain the first radar point cloud timing diagram record data; the radar point cloud timing diagram record data is segmented according to one eighth of the first segmentation step length to obtain the second radar point cloud timing diagram record data. The first radar point cloud timing diagram record data is input into the SLOW channel, and the second radar point cloud timing diagram record data is input into the FAST channel to obtain the target position prediction timing information; the proportion of the moments in which the position deviation distance in the position deviation distance timing information of the target position identification timing information and the target position prediction timing information is greater than or equal to the position deviation distance threshold is taken as the mobile target recognition loss value; when the mobile target recognition loss value of at least the preset number of times in the continuous preset number of trainings is less than or equal to the mobile target recognition loss value threshold, it is considered to be converged, and a mobile target recognition network is generated. The mobile target recognition network can quickly realize the recognition and sorting of mobile targets, and improve timeliness.

[0047] An unmanned driving control method combined with dynamic obstacle trajectory prediction provided by an embodiment of the present invention has at least the following technical effects: By using millimeter-wave radar to detect the surrounding environment, dynamic obstacles are detected and identified in real time. For irregular moving targets, such as stones, the possible position timing information is predicted through terrain-guided trajectory analysis. This method does not rely on a large amount of training data, but uses the terrain grid model to assist in prediction, thereby improving the avoidance rate of unpredictable obstacles. In this way, it can respond to emergencies more effectively, achieving the technical effect of improving the safety and reliability of unmanned vehicles.

[0048] Embodiment 2: like Figure 2 As shown, based on the same inventive concept of an unmanned driving control method combined with dynamic obstacle trajectory prediction provided in embodiment 1, an embodiment of the present invention further provides an unmanned driving control system combined with dynamic obstacle trajectory prediction, comprising: A radar detection unit is used to detect radar point cloud timing information within a preset radius through a millimeter-wave radar deployed on the vehicle after the vehicle is started; A mobile target identification unit, used to perform mobile target identification on the radar point cloud time series information to obtain mobile target positioning time series information; A mobile state analysis unit, used to perform mobile state analysis on the mobile target positioning time sequence information to obtain a mobile law binary parameter; A terrain model retrieving unit, used for classifying the moving target as a type of moving target when the binary parameter of the moving rule is equal to 0, and retrieving a terrain grid model within a preset radius through a road network map; A target movement prediction unit, used for performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the predicted position time series information of the mobile target; An unmanned driving control unit is used to perform unmanned driving control according to the predicted position timing information of the moving target.

[0049] Further, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information, which includes: Performing terrain guidance weight analysis on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient; When the terrain guidance weight coefficient is greater than or equal to the terrain guidance weight coefficient threshold, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information; When the terrain guidance weight coefficient is less than the terrain guidance weight coefficient threshold, a safety warning signal is generated according to the real-time position of the mobile target to warn the central control platform.

[0050] Further, a terrain guidance weight analysis is performed on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient, including: Performing terrain guidance trajectory analysis on the mobile target positioning information at the first moment of the mobile target positioning time series information based on the terrain grid model to obtain the mobile target terrain guidance positioning information at the second moment; Until the terrain guidance trajectory analysis is performed on the mobile target positioning information at the N-1th moment of the mobile target positioning time series information based on the terrain grid model, to obtain the mobile target terrain guidance positioning information at the Nth moment; Calculate a first positioning distance between the terrain guidance positioning information of the mobile target at the second moment and the positioning information of the mobile target at the second moment; Until calculating the Nth positioning distance between the terrain guidance positioning information of the mobile target at the Nth moment and the positioning information of the mobile target at the Nth moment; The proportion of the first positioning distances to the Nth positioning distance that is less than or equal to the positioning distance threshold is counted and set as the terrain guidance weight coefficient.

[0051] Further, performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information includes: Traversing the radar point cloud time series information to perform outlier analysis to obtain a dense radar point cloud time series diagram; Segmenting the dense radar point cloud time series graph according to a first segmentation step to obtain a first dense radar point cloud time series graph; Segmenting the dense radar point cloud time series graph according to one eighth of the first segmentation step length to obtain a second dense radar point cloud time series graph; The first dense radar point cloud timing diagram is input into a mobile target positioning channel of a mobile target recognition network, and the second dense radar point cloud timing diagram is input into a mobile target segmentation channel of the mobile target recognition network for processing to obtain the mobile target positioning timing information.

[0052] Furthermore, the steps of constructing the mobile target recognition network include: Collect radar point cloud time series diagram record data and target position identification time series information; Segmenting the radar point cloud time sequence diagram record data according to a first segmentation step length to obtain first radar point cloud time sequence diagram record data; Segmenting the radar point cloud time series graph record data according to one eighth of the first segmentation step length to obtain second radar point cloud time series graph record data; Input the first radar point cloud time series diagram record data into the SLOW channel, and input the second radar point cloud time series diagram record data into the FAST channel to obtain target position prediction time series information; The proportion of moments in which the position deviation distance in the position deviation distance time series information of the target position identification time series information and the target position prediction time series information is greater than or equal to the position deviation distance threshold is taken as the mobile target recognition loss value; When the mobile target recognition loss value for a preset number of consecutive trainings is less than or equal to the mobile target recognition loss value threshold, the mobile target recognition network is generated.

[0053] Furthermore, the moving state analysis is performed on the moving target positioning time sequence information to obtain the moving law binary parameters, including: Analyze the first speed direction and the first speed magnitude of the first moment positioning information and the second moment positioning information of the mobile target positioning timing information; Until analyzing the N-1th moment positioning information of the mobile target positioning timing information and the N-1th speed direction and the N-1th speed magnitude of the N-1th moment positioning information; Calculating a first deviation angle between the first speed direction and the second speed direction, until calculating an N-2th deviation angle between the N-2th speed direction and the N-1th speed direction, and counting first variance parameters of the first deviation angles until the N-2th deviation angle; Calculating a first deviation between the first speed magnitude and the second speed magnitude, until calculating an N-2th deviation between the N-2th speed magnitude and the N-1th speed magnitude, and counting second variance parameters of the first deviation magnitudes until the N-2th deviation magnitude; When the first variance parameter is greater than a first variance parameter threshold, or / and the second variance parameter is greater than a second variance parameter threshold, the moving law binary parameter is equal to 0; Otherwise, the movement law binary parameter is equal to 1.

[0054] Furthermore, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information, including: Obtain the real-time position and speed of the moving target; Inputting the real-time position of the mobile target and the terrain grid model into a mobile trajectory matching model for analysis to obtain a mobile prediction trajectory; Obtaining a terrain height sequence of the predicted movement trajectory; Inputting the real-time speed of the moving target and the terrain height sequence into a moving speed matching model for analysis to obtain a moving speed sequence; The moving speed sequence and the moving predicted trajectory are fused to obtain the moving target predicted position time series information.

[0055] Embodiment three: See also Figure 3 , Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a first computer program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the first computer program 511, the following steps are implemented: When the vehicle starts, the millimeter-wave radar deployed on the vehicle detects the radar point cloud time series information within a preset radius; Performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information; Performing a moving state analysis on the moving target positioning time sequence information to obtain a moving law binary parameter; When the binary parameter of the movement rule is equal to 0, the moving target is classified as a type of moving target, and at the same time, a terrain grid model within a preset radius is retrieved through a road network map; Performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the predicted position time series information of the mobile target; Unmanned driving control is performed according to the predicted position timing information of the moving target.

[0056] Embodiment 4: See also Figure 4 , Figure 4A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 4 As shown, this embodiment provides a computer-readable storage medium 600, on which a second computer program 611 is stored. When the second computer program 611 is executed by a processor, the following steps are implemented: When the vehicle starts, the millimeter-wave radar deployed on the vehicle detects the radar point cloud time series information within a preset radius; Performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information; Performing a moving state analysis on the moving target positioning time sequence information to obtain a moving law binary parameter; When the binary parameter of the movement rule is equal to 0, the moving target is classified as a type of moving target, and at the same time, a terrain grid model within a preset radius is retrieved through a road network map; Performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the predicted position time series information of the mobile target; Unmanned driving control is performed according to the predicted position timing information of the moving target.

[0057] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0058] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0062] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0063] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. An unmanned driving control method combined with dynamic obstacle trajectory prediction, characterized in that: include: When the vehicle is started, the millimeter-wave radar deployed on the vehicle detects the radar point cloud time series information within a preset radius; Performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information; Performing a moving state analysis on the moving target positioning time sequence information to obtain a moving law binary parameter; When the binary parameter of the movement rule is equal to 0, the moving target is classified as a type of moving target, and at the same time, a terrain grid model within a preset radius is retrieved through a road network map; Performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information; Unmanned driving control is performed according to the predicted position timing information of the moving target.

2. The method according to claim 1, characterized in that The terrain-guided trajectory of the mobile target is analyzed based on the terrain grid model to obtain the predicted position time series information of the mobile target, which includes: Performing terrain guidance weight analysis on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient; When the terrain guidance weight coefficient is greater than or equal to the terrain guidance weight coefficient threshold, terrain guidance trajectory analysis is performed on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information; When the terrain guidance weight coefficient is less than the terrain guidance weight coefficient threshold, a safety warning signal is generated according to the real-time position of the mobile target to warn the central control platform.

3. The method according to claim 2, characterized in that Performing terrain guidance weight analysis on the mobile target positioning time series information based on the terrain grid model to obtain a terrain guidance weight coefficient includes: Performing terrain guidance trajectory analysis on the mobile target positioning information at the first moment of the mobile target positioning time series information based on the terrain grid model to obtain the mobile target terrain guidance positioning information at the second moment; Until the terrain guidance trajectory analysis is performed on the mobile target positioning information at the N-1th moment of the mobile target positioning time series information based on the terrain grid model, to obtain the mobile target terrain guidance positioning information at the Nth moment; Calculate a first positioning distance between the terrain guidance positioning information of the mobile target at the second moment and the positioning information of the mobile target at the second moment; Until calculating the Nth positioning distance between the terrain guidance positioning information of the mobile target at the Nth moment and the positioning information of the mobile target at the Nth moment; The proportion of the first positioning distances to the Nth positioning distance that is less than or equal to the positioning distance threshold is counted and set as the terrain guidance weight coefficient.

4. The method according to claim 1, characterized in that Performing mobile target recognition on the radar point cloud time series information to obtain mobile target positioning time series information includes: Traversing the radar point cloud time series information to perform outlier analysis to obtain a dense radar point cloud time series diagram; Segmenting the dense radar point cloud time series graph according to a first segmentation step to obtain a first dense radar point cloud time series graph; Segmenting the dense radar point cloud time series graph according to one eighth of the first segmentation step length to obtain a second dense radar point cloud time series graph; The first dense radar point cloud timing diagram is input into a mobile target positioning channel of a mobile target recognition network, and the second dense radar point cloud timing diagram is input into a mobile target segmentation channel of the mobile target recognition network for processing to obtain the mobile target positioning timing information.

5. The method according to claim 4, characterized in that The steps of constructing the mobile target recognition network include: Collect radar point cloud time series diagram record data and target position identification time series information; Segmenting the radar point cloud time sequence diagram record data according to a first segmentation step length to obtain first radar point cloud time sequence diagram record data; Segmenting the radar point cloud time series graph record data according to one eighth of the first segmentation step length to obtain second radar point cloud time series graph record data; Input the first radar point cloud time series diagram record data into the SLOW channel, and input the second radar point cloud time series diagram record data into the FAST channel to obtain target position prediction time series information; The proportion of moments in which the position deviation distance in the position deviation distance time series information of the target position identification time series information and the target position prediction time series information is greater than or equal to the position deviation distance threshold is taken as the mobile target recognition loss value; When the mobile target recognition loss value for a preset number of consecutive trainings is less than or equal to the mobile target recognition loss value threshold, the mobile target recognition network is generated.

6. The method according to claim 1, characterized in that Performing a mobile state analysis on the mobile target positioning time sequence information to obtain a binary parameter of the mobile law includes: Analyze the first speed direction and the first speed magnitude of the first moment positioning information and the second moment positioning information of the mobile target positioning timing information; Until analyzing the N-1th moment positioning information of the mobile target positioning timing information and the N-1th speed direction and the N-1th speed magnitude of the N-1th moment positioning information; Calculating a first deviation angle between the first speed direction and the second speed direction, until calculating an N-2th deviation angle between the N-2th speed direction and the N-1th speed direction, and counting first variance parameters of the first deviation angles until the N-2th deviation angle; Calculating a first deviation between the first speed magnitude and the second speed magnitude, until calculating an N-2th deviation between the N-2th speed magnitude and the N-1th speed magnitude, and counting second variance parameters of the first deviation magnitudes until the N-2th deviation magnitude; When the first variance parameter is greater than a first variance parameter threshold, or / and the second variance parameter is greater than a second variance parameter threshold, the moving law binary parameter is equal to 0; Otherwise, the movement law binary parameter is equal to 1.

7. The method according to claim 1, characterized in that Performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the mobile target predicted position time series information includes: Obtain the real-time position and speed of the moving target; Inputting the real-time position of the mobile target and the terrain grid model into a mobile trajectory matching model for analysis to obtain a mobile prediction trajectory; Obtaining a terrain height sequence of the predicted movement trajectory; Inputting the real-time speed of the moving target and the terrain height sequence into a moving speed matching model for analysis to obtain a moving speed sequence; The moving speed sequence and the moving predicted trajectory are fused to obtain the moving target predicted position time series information.

8. An unmanned driving control system combined with dynamic obstacle trajectory prediction, characterized in that: Used to implement the method according to any one of claims 1 to 7, comprising: A radar detection unit is used to detect radar point cloud timing information within a preset radius through a millimeter-wave radar deployed on the vehicle after the vehicle is started; A mobile target identification unit, used to perform mobile target identification on the radar point cloud time series information to obtain mobile target positioning time series information; A mobile state analysis unit, used to perform mobile state analysis on the mobile target positioning time sequence information to obtain a mobile law binary parameter; A terrain model retrieving unit, used for classifying the moving target as a type of moving target when the binary parameter of the moving rule is equal to 0, and retrieving a terrain grid model within a preset radius through a road network map; A target movement prediction unit, used for performing terrain guidance trajectory analysis on the mobile target based on the terrain grid model to obtain the predicted position time series information of the mobile target; An unmanned driving control unit is used to perform unmanned driving control according to the predicted position timing information of the mobile target.

9. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing an unmanned driving control method combined with dynamic obstacle trajectory prediction as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, and when the computer software program is executed by the processor, it implements an unmanned driving control method combined with dynamic obstacle trajectory prediction as described in any one of claims 1 to 7.

Images