CN114387307A - Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving - Google Patents

Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving Download PDF

Info

Publication number
CN114387307A
CN114387307A CN202210020394.1A CN202210020394A CN114387307A CN 114387307 A CN114387307 A CN 114387307A CN 202210020394 A CN202210020394 A CN 202210020394A CN 114387307 A CN114387307 A CN 114387307A
Authority
CN
China
Prior art keywords
moving object
predicted
preset
predicted motion
trajectory
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210020394.1A
Other languages
Chinese (zh)
Inventor
陈红丽
李荣华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
FAW Group Corp
Original Assignee
FAW Group Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by FAW Group Corp filed Critical FAW Group Corp
Priority to CN202210020394.1A priority Critical patent/CN114387307A/en
Publication of CN114387307A publication Critical patent/CN114387307A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G06T7/251—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments involving models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G06T7/75—Determining position or orientation of objects or cameras using feature-based methods involving models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30241—Trajectory
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Traffic Control Systems (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)

Abstract

The application relates to a method, a device, equipment and a medium for predicting a track of a moving object in automatic driving. The method comprises the following steps: determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by the automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set; acquiring existing predicted motion trajectories of first moving objects in a first set to be predicted, which are obtained by a preset trajectory prediction model before the current frame data, and determining predicted motion trajectories of the first moving objects in a preset time period after the current frame data based on the existing predicted motion trajectories; and determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through a preset track prediction model. The method improves the real-time performance of the track prediction.

Description

Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving
Technical Field
The present application relates to the field of automatic driving technologies, and in particular, to a method, an apparatus, a device, and a medium for predicting a trajectory of a moving object in automatic driving.
Background
In the driving process, driving safety is always a very important thing. In order to ensure the driving safety of the automatic driving vehicle, it is necessary to predict the track of the moving object in the visual field of the automatic driving vehicle.
However, when the number of moving objects in the field of view of the autonomous vehicle is large, the real-time prediction of the trajectory of the moving object may not meet the actual requirements due to the limited computing resources of the on-board computing unit of the autonomous vehicle.
Disclosure of Invention
Based on this, it is necessary to provide a method, an apparatus, a device, and a medium for predicting a trajectory of a moving object in automatic driving, aiming at the technical problem that the real-time performance of the trajectory prediction of the moving object in the conventional manner may not meet the actual requirement.
In a first aspect, an embodiment of the present application provides a method for predicting a trajectory of a moving object in automatic driving, including:
determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
acquiring the existing predicted motion trail of each first moving object in the first to-be-predicted set before the current frame data through the preset trail prediction model, and determining the predicted motion trail of each first moving object in a preset time period after the current frame data based on each existing predicted motion trail;
and determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
In a second aspect, an embodiment of the present application provides an apparatus for predicting a trajectory of a moving object in automatic driving, including:
the automatic prediction system comprises a determining module, a calculating module and a calculating module, wherein the determining module is used for determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
an obtaining module, configured to obtain an existing predicted motion trajectory, which is predicted by each first moving object in the first set to be predicted through the preset trajectory prediction model before the frame data;
the first prediction module is used for determining the predicted motion trail of each first motion object in a preset time period after the current frame data based on each existing predicted motion trail;
and the second prediction module is used for determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
In a third aspect, an embodiment of the present application provides a computing device, including a memory and a processor, where the memory stores a computer program, and the processor implements, when executing the computer program, the steps of the method for predicting a trajectory of a moving object in automatic driving provided by the first aspect of the embodiment of the present application.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the trajectory prediction method for a moving object in automatic driving provided by the first aspect of the present application.
The technical scheme provided by the embodiment of the application determines a first moving object and a second moving object in a moving object set corresponding to the frame data acquired by the automatic driving vehicle, for each first moving object processed by the preset track prediction model before the current frame data, determining the predicted motion track of each first moving object in the preset time period after the current frame data based on the existing predicted motion track of each first moving object, for each second moving object which is not processed by the preset track prediction model, the predicted moving track of each second moving object in the preset time period after the current frame data is determined by the preset track prediction model, that is, the number of the moving objects processed by the preset track prediction model is reduced, the calculation pressure of a calculation unit of the automatic driving vehicle is relieved, and therefore the real-time performance of track prediction of each moving object corresponding to the current frame data is improved. Meanwhile, the predicted motion trail of part of the moving objects can be determined by the predicted existing motion trail before the current frame data, so that the number of the predicted motion trail output by the calculating unit is increased on the whole, and the accuracy of the output predicted motion trail is ensured.
Drawings
Fig. 1 is a schematic flowchart of a method for predicting a trajectory of a moving object in automatic driving according to an embodiment of the present disclosure;
fig. 2 is a flowchart illustrating a process of determining a predicted motion trajectory of a first moving object according to an embodiment of the present application;
fig. 3 is a flowchart illustrating a process of determining a predicted motion trajectory of a second moving object according to an embodiment of the present application;
fig. 4 is another schematic flowchart of a process of determining a predicted motion trajectory of a second moving object according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a trajectory prediction apparatus for a moving object in automatic driving according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of a computing device according to an embodiment of the present application.
Detailed Description
In the related art, the motion trail of the moving object in the visual field of the automatic driving vehicle is predicted by the artificial intelligence model, but the prediction time of the complex artificial intelligence model may be long, so that when the number of the moving objects in the visual field of the automatic driving vehicle is large, the prediction is limited by the computing resource of the vehicle-mounted computing unit of the automatic driving vehicle, and the real-time performance of the prediction of the track of each moving object may not meet the actual requirement.
In view of the above problems, the technical solution provided in the embodiments of the present application can not only improve the real-time performance of the trajectory prediction for each moving object, but also improve the number of output predicted movement trajectories and the accuracy of the predicted movement trajectories, thereby improving the driving safety of the autonomous vehicle.
It should be noted that the technical solution provided in the embodiment of the present application may be executed by an onboard computing device in an autonomous vehicle, or may be executed by a cloud device, which is not limited in the embodiment of the present application.
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application are further described in detail by the following embodiments in combination with the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Fig. 1 is a schematic flowchart of a method for predicting a trajectory of a moving object in automatic driving according to an embodiment of the present disclosure. As shown in fig. 1, the method may include:
s101, determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by the automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set.
The automatic driving vehicle can be provided with one or more environment sensing devices, and the surrounding environment information of the vehicle, the driving information of the vehicle and the like are collected in the driving process of the vehicle. The driving information of the vehicle includes, but is not limited to, a geographical location, a driving speed, a driving direction, and the like of the vehicle. Alternatively, the environment sensing device may be a camera, various sensors, and the like.
After the frame data is collected, a preset detection algorithm can be adopted to detect a moving object existing in the frame data. Alternatively, the moving object may be another vehicle or a pedestrian, and the like, and the other vehicle may be a motor vehicle as well as a non-motor vehicle. Alternatively, the detection algorithm may be a yolo (young Only Look one) detection network, or may be a full-volume area-based network (R-CNN) or a ssd (single Shot multi box detector).
After obtaining the moving object set corresponding to the frame of data, the computing device may traverse each moving object one by one to obtain the service tag corresponding to each moving object, and divide each moving object based on the service tag corresponding to each moving object. The service mark is used for identifying whether the moving object is processed by a preset track prediction model before the frame data. Optionally, the preset trajectory prediction model may be a pre-trained recurrent neural network model, that is, a trajectory prediction model may be constructed by using a deep learning algorithm, the established trajectory prediction model is trained by using a pre-obtained training data set, and the trained trajectory prediction model is finally obtained by continuously iterating until convergence, and is stored in the computing device. The preset track prediction model can predict the motion track of the moving object in a future preset time period based on the historical motion track of the moving object.
For example, when the service flag of the moving object is "future", which indicates that the moving object has been processed by the preset trajectory prediction model before the current frame data, the moving object may be determined as the first moving object, and all the first moving objects form the first set to be predicted. When the service mark of the moving object is "ase", it indicates that the moving object is not processed by the preset trajectory prediction model, and the moving object may be determined as a second moving object, and all the second moving objects form a second set to be predicted. Of course, other characters or values may be used to represent the service mark, such as the strings "yes" and "no", such as the values "1" and "0".
It is understood that, as the autonomous vehicle travels and the surrounding environment changes, the moving object in the frame data may include a moving object in the previous frame data and may also include a newly added moving object, that is, the second moving object that has not been processed by the preset trajectory prediction model may include a newly added moving object in the frame data and also include a moving object that has not been processed by the preset trajectory prediction model in the previous frame data.
S102, obtaining the existing predicted motion trail of each first moving object in the first to-be-predicted set before the current frame data through the preset trail prediction model, and determining the predicted motion trail of each first moving object in the preset time period after the current frame data based on each existing predicted motion trail.
It is considered that the computing device already predicts the motion trail of the first moving object by using the preset trail prediction model in the process of processing the previous frame data, namely, the computing device already has the existing predicted motion trail of the first moving object before the data of the current frame is processed. Based on this, when predicting the motion trajectory of the moving object in the present frame data, in order to relieve the computational pressure of the computing device, for a first moving object having an existing predicted motion trajectory, the computing device may directly acquire the existing predicted motion trajectory of the first moving object before the present frame data, which is obtained through a preset trajectory prediction model, and determine the predicted motion trajectory of the first moving object in a preset time period after the present frame data by using the existing predicted motion trajectory as a reference, instead of predicting the future motion trajectory of the first moving object by using the preset trajectory prediction model.
S103, determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
For a second moving object which is not processed by the preset trajectory prediction model, the computing device may obtain historical motion trajectory data of the second moving object, input the historical motion trajectory data of the second moving object to the preset trajectory prediction model, and perform model prediction by using the preset trajectory prediction model, so as to obtain a predicted motion trajectory of the second moving object in a preset time period after the current frame of data.
Therefore, for the computing equipment, the motion tracks of part of the moving objects in the frame data are predicted only by adopting the preset track prediction model, that is, the number of the moving objects predicted by adopting the preset track prediction model is relatively reduced, the processing speed of the computing equipment is greatly improved, and the real-time performance of track prediction is improved.
The trajectory prediction method for moving objects in automatic driving provided by the embodiment of the present application determines a first moving object and a second moving object in a moving object set corresponding to a current frame of data acquired by an automatic driving vehicle, determines, for each first moving object processed by a preset trajectory prediction model before the current frame of data, a predicted motion trajectory of each first moving object in a preset time period after the current frame of data based on an existing predicted motion trajectory of each first moving object, determines, for each second moving object not processed by the preset trajectory prediction model, a predicted motion trajectory of each second moving object in a preset time period after the current frame of data by using the preset trajectory prediction model, that is, reduces the number of moving objects processed by the preset trajectory prediction model, and relieves the computational pressure of a computing unit of the automatic driving vehicle, therefore, the real-time performance of the track prediction of each moving object corresponding to the frame data is improved. Meanwhile, the predicted motion trail of part of the moving objects can be determined by the predicted existing motion trail before the current frame data, so that the number of the predicted motion trail output by the calculating unit is increased on the whole, and the accuracy of the output predicted motion trail is ensured.
In practical applications, for the same moving object, multiple moving trajectories are usually predicted by using a preset trajectory prediction model. Based on this, optionally, the process of obtaining the existing predicted motion trajectory of each first moving object in the first to-be-predicted set before the current frame data through the preset trajectory prediction model in S102 may be: and acquiring a plurality of existing predicted motion tracks of each first moving object in the first set to be predicted, which are obtained by a preset track prediction model before the frame data.
Correspondingly, as shown in fig. 2, the process of determining the predicted motion trajectory of each first moving object in the preset time period after the current frame data based on each existing predicted motion trajectory in S102 may be:
s201, determining a target existing predicted motion trail from a plurality of existing predicted motion trails of each first moving object based on actual position information of each first moving object at the current moment and predicted position information of the existing predicted motion trails at the current moment.
The target existing predicted motion track can be regarded as the predicted motion track which is most matched with the actual running condition of the first moving object. After the frame data is processed by using a preset detection algorithm to obtain the actual position information of the first moving object at the current time, the computing device may compare the actual position information of the first moving object at the current time with the predicted position information of the first moving object at the current time in each existing predicted movement track obtained in the previous frame data processing process, select an existing predicted movement track most matched with the current actual running condition of the first moving object from the existing predicted movement tracks based on the comparison result, and determine the existing predicted movement track as the target existing predicted movement track.
As an alternative embodiment, the computing device may employ the following process to determine the existing predicted motion trajectory of the target. Optionally, the step S201 may include the following steps:
s2011, for each first moving object, calculating a euclidean distance between the actual position information of the first moving object at the current time and the predicted position information of the existing predicted motion trajectory at the current time.
Specifically, the calculation device may calculate the euclidean distance trj between the actual position information of the first moving object at the current time and the predicted position information at the current time in each existing predicted motion trajectory according to the following formula 1 or a modification of formula 1.
Figure BDA0003462280510000091
Wherein (X)Prediction，YPrediction) For the predicted position information of the first moving object in the existing predicted motion trail at the current moment, (X)Practice of，YPractice of) Is the actual position information of the first moving object at the current moment.
And S2012, determining the existing predicted motion track corresponding to the predicted position information with the minimum Euclidean distance as the existing predicted motion track of the target.
The smaller the euclidean distance is, the higher the matching degree of the corresponding existing predicted motion trajectory with the actual operation condition of the first motion object is, whereas, the larger the euclidean distance is, the lower the matching degree of the corresponding existing predicted motion trajectory with the actual operation condition of the first motion object is, and the existing predicted motion trajectory corresponding to the predicted position information with the minimum euclidean distance may be determined as the existing predicted motion trajectory of the target.
S202, updating the time stamp of the track point in the existing predicted motion track of each target to obtain the predicted motion track of each first motion object in the preset time period after the current frame of data.
After the existing predicted motion track of the target is obtained, the first track point of the existing predicted motion track of the target can be deleted, the timestamps of the rest track points in the existing predicted motion track of the target are updated according to the time interval between the track points, and the predicted motion track of the first motion object in the preset time period after the frame data is formed based on the updated track points.
For example, assuming that the time interval between the track points is 100ms, the timestamp of the rest of the track points in the target predicted motion track can be updated to Told-100. Wherein, Told is the original time stamp of the rest track points.
In this embodiment, the computing device may determine, based on the actual position information of the first moving object at the current time and the predicted position information of the first moving object at the current time in the existing predicted movement trajectories, an existing target predicted movement trajectory with the highest accuracy from among a plurality of existing predicted movement trajectories of the first moving object, and update the existing target predicted movement trajectory, so as to obtain a predicted movement trajectory of the first moving object in a preset time period after the current frame of data, thereby improving the accuracy of the predicted movement trajectory.
In practical applications, there may be a case where it is determined that the number of second moving objects that have not been processed by the preset trajectory prediction model is greater than the upper limit of the processing capacity of the computing device, and for this case, the trajectory prediction of the second moving object may be performed by referring to the process of the following embodiment. On the basis of the foregoing embodiment, as an optional implementation manner, as shown in fig. 3, the process of S103 may be:
s301, screening out second moving objects located in a preset range of the automatic driving vehicle from the second set to be predicted based on the position information of the second moving objects at the current moment, and forming a third set to be predicted.
In general, moving objects closer to the autonomous vehicle have a greater degree of influence on the autonomous vehicle, and moving objects farther from the autonomous vehicle have a smaller degree of influence on the autonomous vehicle. Therefore, in order to improve the real-time performance of the trajectory prediction, second moving objects located within the preset range of the autonomous vehicle can be screened out from the second to-be-predicted set based on the position information of each second moving object at the current moment, the second moving objects are processed by the preset trajectory prediction model preferentially, and other second moving objects outside the preset range can be processed when the next frame of data arrives. The preset range may be set based on actual conditions, for example, the preset range may be set to be within 30 meters.
S302, determining the predicted motion track of each second motion object in the third set to be predicted in a preset time period after the frame data through the preset track prediction model.
The computing device may input historical motion trajectory data of each second moving object in the third set to be predicted to the preset trajectory prediction model, and perform model prediction through the preset trajectory prediction model to obtain a predicted motion trajectory of each second moving object in a preset time period after the current frame of data.
In this embodiment, the computing device may preferentially adopt the preset trajectory prediction model to predict the future travel trajectory of each second moving object located within the preset range of the autonomous vehicle, that is, preferentially adopt the preset trajectory prediction model to predict the future travel trajectory of the second moving object having a higher influence on the autonomous vehicle, so as to improve the real-time performance and accuracy of the prediction of the trajectory of the second moving object having a higher influence on the autonomous vehicle, thereby assisting the travel decision of the autonomous vehicle and contributing to improve the safety of the autonomous vehicle.
As another alternative implementation, as shown in fig. 4, the process of S103 may be:
s401, the priority of each second moving object is determined based on the traveling information of each second moving object.
The driving information may include, but is not limited to, a geographical location of the second moving object, a driving speed of the second moving object, a driving direction, a driving acceleration, a distance between the second moving object and the autonomous vehicle, and the like. Generally, a second moving object which is far away from the autonomous vehicle and has a relatively low driving speed and relatively low driving acceleration may be determined to have a lower priority as the influence on the autonomous vehicle is smaller, and the motion trajectory of the second moving object may be predicted when the next frame of data arrives; conversely, the closer the second moving object is to the autonomous vehicle, and the higher the traveling speed and the traveling acceleration are, the greater the influence of the second moving object on the autonomous vehicle is, the higher the priority of the second moving object is determined, and the motion trajectory thereof needs to be predicted as soon as possible.
S402, determining each second moving object with the priority meeting the preset requirement as an object to be predicted.
After the priority of each second moving object is obtained, each second moving object whose priority meets the preset requirement may be determined as an object to be predicted. For example, if the priority of each second moving object is represented by a numerical value, each second moving object whose priority exceeds a preset threshold may be determined as an object to be predicted.
And S403, determining the predicted motion track of each object to be predicted in a preset time period after the frame data through a preset track prediction model corresponding to the type of each object to be predicted.
Wherein, each object to be predicted can be a pedestrian or other vehicles. Different types of moving objects can adopt different preset track prediction models to perform track prediction. For example, when the object to be predicted is a pedestrian, the computing device may perform the trajectory prediction using a pre-trained pedestrian trajectory prediction model, and when the object to be predicted is a vehicle, the computing device may perform the trajectory prediction using a pre-trained vehicle trajectory prediction model.
That is to say, the computing device may input the historical motion trajectory data of each object to be predicted into the preset trajectory prediction model corresponding to the type of each object to be predicted, and perform model prediction through the corresponding preset trajectory prediction model to obtain the predicted motion trajectory of each object to be predicted in the preset time period after the current frame of data.
In this embodiment, the computing device may determine the priority of each second moving object based on the driving information of each second moving object, and preferentially adopt the corresponding preset trajectory prediction model to predict the future driving trajectory of each second moving object whose priority meets the preset requirement, that is, preferentially adopt the preset trajectory prediction model to predict the future driving trajectory of the second moving object having a higher influence degree on the autonomous vehicle, so as to improve the real-time performance and accuracy of the prediction of the trajectory of the second moving object having a higher influence degree on the autonomous vehicle, thereby assisting the driving decision of the autonomous vehicle, and contributing to improve the safety of the autonomous vehicle.
For the understanding of those skilled in the art, the following describes the trajectory prediction process of the moving object in automatic driving in detail:
for the first frame of data acquired by the autonomous vehicle, when it is detected that the number of moving objects in the first frame of data exceeds the upper limit of the processing capacity of the computing device, the computing device may determine, in the first frame of data, a moving object that has a large influence on the autonomous vehicle (for example, the moving object may be determined based on the driving information of each moving object), and preferentially perform trajectory prediction on the moving object by using a preset trajectory prediction model. Then, when second frame data arrives, the computing device detects a moving object in the second frame data to obtain a moving object set corresponding to the second frame data, and determines a first moving object which is processed by a preset track prediction model before the second frame data and a second moving object which is not processed by the preset track prediction model from the moving object set; further, for each first moving object, the computing device obtains an existing predicted motion trajectory of each first moving object before the second frame data through a preset trajectory prediction model, and determines a predicted motion trajectory of each first moving object after the second frame data for a preset time period based on each existing predicted motion trajectory. For each second moving object, the computing device determines a predicted motion track of each second moving object in a preset time period after the second frame data through a preset track prediction model. By analogy, the computing equipment adopts the method to predict the future motion trail of the motion object in each frame of data, so that the real-time prediction of each motion object trail is improved, the number of the predicted motion trails output by the computing equipment is improved on the whole, and the accuracy of the output predicted motion trail is ensured.
Fig. 5 is a schematic structural diagram of an apparatus for predicting a trajectory of a moving object in automatic driving according to an embodiment of the present disclosure. As shown in fig. 5, the apparatus may include: a determination module 501, an acquisition module 502, a first prediction module 503, and a second prediction module 504.
Specifically, the determining module 501 is configured to determine, from a moving object set corresponding to a current frame of data acquired by an autonomous vehicle, a first moving object that has been processed by a preset trajectory prediction model before the current frame of data to form a first set to be predicted, and determine a second moving object that has not been processed by the preset trajectory prediction model to form a second set to be predicted;
the obtaining module 502 is configured to obtain an existing predicted motion trajectory of each first moving object in the first set to be predicted, which is predicted by the preset trajectory prediction model before the current frame data;
the first prediction module 503 is configured to determine, based on each existing predicted motion trajectory, a predicted motion trajectory of each first moving object in a preset time period after the current frame of data;
the second prediction module 504 is configured to determine, through the preset trajectory prediction model, a predicted motion trajectory of each second moving object in the second set to be predicted in a preset time period after the current frame of data.
The trajectory prediction apparatus for moving objects in automatic driving provided in the embodiment of the present application determines a first moving object and a second moving object in a moving object set corresponding to a current frame of data acquired by an automatic driving vehicle, determines, for each first moving object processed by a preset trajectory prediction model before the current frame of data, a predicted motion trajectory of each first moving object in a preset time period after the current frame of data based on an existing predicted motion trajectory of each first moving object, determines, for each second moving object not processed by the preset trajectory prediction model, a predicted motion trajectory of each second moving object in a preset time period after the current frame of data by using the preset trajectory prediction model, that is, reduces the number of moving objects processed by the preset trajectory prediction model, and relieves the computational pressure of a computing unit of the automatic driving vehicle, therefore, the real-time performance of the track prediction of each moving object corresponding to the frame data is improved. Meanwhile, the predicted motion trail of part of the moving objects can be determined by the predicted existing motion trail before the current frame data, so that the number of the predicted motion trail output by the calculating unit is increased on the whole, and the accuracy of the output predicted motion trail is ensured.
On the basis of the foregoing embodiment, optionally, the obtaining module 502 is specifically configured to obtain a plurality of existing predicted motion trajectories, obtained by the first moving objects in the first set to be predicted through the preset trajectory prediction model before the frame data;
correspondingly, the first prediction module 503 may include: a determining unit and an updating unit.
Specifically, the determining unit is configured to determine an existing predicted motion trajectory of the target from a plurality of existing predicted motion trajectories of each first moving object based on actual position information of each first moving object at a current time and predicted position information of the existing predicted motion trajectories at the current time;
the updating unit is used for updating the time stamp of the track point in the existing predicted motion track of each target to obtain the predicted motion track of each first moving object in the preset time period after the current frame data.
On the basis of the foregoing embodiment, optionally, the determining unit is specifically configured to calculate, for each first moving object, a euclidean distance between actual position information of the first moving object at the current time and predicted position information of the existing predicted motion trajectory at the current time; and determining the existing predicted motion track corresponding to the predicted position information with the minimum Euclidean distance as the existing predicted motion track of the target.
On the basis of the foregoing embodiment, optionally, the second prediction module 504 is specifically configured to, based on the position information of each second moving object at the current time, screen out, from the second set to be predicted, a second moving object located within a preset range of the autonomous vehicle, and form a third set to be predicted; and determining the predicted motion track of each second motion object in the third set to be predicted in a preset time period after the frame data through the preset track prediction model.
On the basis of the foregoing embodiment, optionally, the second prediction module 504 is specifically configured to determine the priority of each second moving object based on the driving information of each second moving object; determining each second moving object with the priority meeting the preset requirement as an object to be predicted; and determining the predicted motion track of each object to be predicted in a preset time period after the frame data through a preset track prediction model corresponding to the type of each object to be predicted.
Optionally, the preset trajectory prediction model is a recurrent neural network model.
In one embodiment, a computing device is provided that may be a device installed in an autonomous vehicle or a cloud-based device. Referring to fig. 6, the computing device may include a processor and memory, a network interface, and a database connected by a system bus. Wherein the processor of the computing device is configured to provide computing and control capabilities. The memory of the computing device includes a non-volatile storage medium, an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computing device is used for storing data in a process of predicting the track of a moving object in automatic driving. The network interface of the computing device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of trajectory prediction of a moving object in autonomous driving.
Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, there is provided a computing device comprising a memory and a processor, the memory having stored therein a computer program that when executed by the processor performs the steps of:
determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
acquiring the existing predicted motion trail of each first moving object in the first to-be-predicted set before the current frame data through the preset trail prediction model, and determining the predicted motion trail of each first moving object in a preset time period after the current frame data based on each existing predicted motion trail;
and determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
acquiring the existing predicted motion trail of each first moving object in the first to-be-predicted set before the current frame data through the preset trail prediction model, and determining the predicted motion trail of each first moving object in a preset time period after the current frame data based on each existing predicted motion trail;
and determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
The trajectory prediction device, the apparatus and the storage medium for a moving object in automatic driving provided in the above embodiments may perform the trajectory prediction method for a moving object in automatic driving provided in any embodiment of the present disclosure, and have corresponding functional modules and beneficial effects for performing the method. Technical details that are not described in detail in the above embodiments may be referred to a trajectory prediction method of a moving object in automatic driving provided in any embodiment of the present disclosure.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (10)

1. A method for predicting a trajectory of a moving object in automatic driving, comprising:
determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
acquiring the existing predicted motion trail of each first moving object in the first to-be-predicted set before the current frame data through the preset trail prediction model, and determining the predicted motion trail of each first moving object in a preset time period after the current frame data based on each existing predicted motion trail;
and determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
2. The method according to claim 1, wherein the obtaining of the existing predicted motion trajectory of each first moving object in the first set to be predicted, which is obtained by the preset trajectory prediction model before the current frame data, comprises:
obtaining a plurality of existing predicted motion tracks of each first moving object in the first set to be predicted, which are obtained by the preset track prediction model before the frame data;
correspondingly, the determining a predicted motion trajectory of each first moving object in a preset time period after the current frame data based on each existing predicted motion trajectory includes:
determining an existing target predicted motion trajectory from a plurality of existing predicted motion trajectories of each first moving object based on actual position information of each first moving object at the current time and predicted position information of each existing predicted motion trajectory at the current time;
and updating the time stamp of the track point in the existing predicted motion track of each target to obtain the predicted motion track of each first motion object in a preset time period after the current frame of data.
3. The method of claim 2, wherein determining a target existing predicted motion trajectory from a plurality of existing predicted motion trajectories for each first moving object based on actual position information for each first moving object at a current time and predicted position information for each existing predicted motion trajectory at the current time comprises:
calculating the Euclidean distance between the actual position information of the first moving object at the current moment and the predicted position information of the existing predicted motion trail at the current moment aiming at each first moving object;
and determining the existing predicted motion track corresponding to the predicted position information with the minimum Euclidean distance as the existing predicted motion track of the target.
4. The method according to any one of claims 1 to 3, wherein the determining, by the preset trajectory prediction model, the predicted motion trajectory of each second moving object in the second set to be predicted in a preset time period after the current frame data includes:
screening out second moving objects located in a preset range of the automatic driving vehicle from the second set to be predicted based on the position information of each second moving object at the current moment to form a third set to be predicted;
and determining the predicted motion track of each second motion object in the third set to be predicted in a preset time period after the frame data through the preset track prediction model.
5. The method according to any one of claims 1 to 3, wherein the determining, by the preset trajectory prediction model, the predicted motion trajectory of each second moving object in the second set to be predicted in a preset time period after the current frame data includes:
determining the priority of each second moving object based on the driving information of each second moving object;
determining each second moving object with the priority meeting the preset requirement as an object to be predicted;
and determining the predicted motion track of each object to be predicted in a preset time period after the frame data through a preset track prediction model corresponding to the type of each object to be predicted.
6. The method according to any one of claims 1 to 3, wherein the preset trajectory prediction model is a recurrent neural network model.
7. An apparatus for predicting a trajectory of a moving object in automatic driving, comprising:
the automatic prediction system comprises a determining module, a calculating module and a calculating module, wherein the determining module is used for determining a first moving object which is processed by a preset track prediction model before the frame data from a moving object set corresponding to the frame data acquired by an automatic driving vehicle to form a first to-be-predicted set, and determining a second moving object which is not processed by the preset track prediction model to form a second to-be-predicted set;
an obtaining module, configured to obtain an existing predicted motion trajectory, which is predicted by each first moving object in the first set to be predicted through the preset trajectory prediction model before the frame data;
the first prediction module is used for determining the predicted motion trail of each first motion object in a preset time period after the current frame data based on each existing predicted motion trail;
and the second prediction module is used for determining the predicted motion track of each second motion object in the second set to be predicted in a preset time period after the frame data through the preset track prediction model.
8. The apparatus according to claim 7, wherein the obtaining module is specifically configured to obtain a plurality of existing predicted motion trajectories obtained by each first moving object in the first set to be predicted through the preset trajectory prediction model before the current frame data;
correspondingly, the first prediction module comprises:
a determination unit configured to determine a target existing predicted motion trajectory from among a plurality of existing predicted motion trajectories of each first moving object, based on actual position information of each first moving object at a current time and predicted position information of the existing predicted motion trajectories at the current time;
and the updating unit is used for updating the time stamp of the track point in the existing predicted motion track of each target to obtain the predicted motion track of each first motion object in the preset time period after the current frame data.
9. A computing device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 6 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
CN202210020394.1A 2022-01-10 2022-01-10 Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving Pending CN114387307A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210020394.1A CN114387307A (en) 2022-01-10 2022-01-10 Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210020394.1A CN114387307A (en) 2022-01-10 2022-01-10 Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving

Publications (1)

Publication Number Publication Date
CN114387307A true CN114387307A (en) 2022-04-22

Family

ID=81200309

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210020394.1A Pending CN114387307A (en) 2022-01-10 2022-01-10 Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving

Country Status (1)

Country Link
CN (1) CN114387307A (en)

Similar Documents

Publication Publication Date Title
CN109572550B (en) Driving track prediction method, system, computer equipment and storage medium
EP3680877A1 (en) Method for determining the location of an ego-vehicle
CN110550030B (en) Lane changing control method and device for unmanned vehicle, computer equipment and storage medium
CN111060125B (en) Collision detection method and device, computer equipment and storage medium
CN112839855A (en) Trajectory prediction method and device
CN111256687A (en) Map data processing method and device, acquisition equipment and storage medium
CN111915878B (en) Method and device for predicting road traffic state, computer device and storage medium
CN112298194B (en) Lane changing control method and device for vehicle
CN113379099B (en) Machine learning and copula model-based highway traffic flow self-adaptive prediction method
CN113239719A (en) Track prediction method and device based on abnormal information identification and computer equipment
CN113942524A (en) Vehicle running control method and system and computer readable storage medium
CN112530159B (en) Self-calibration type multi-lane-level traffic flow detection method and electronic equipment
CN115140060A (en) Data processing method and device, electronic equipment and storage medium
CN114663804A (en) Driving area detection method, device, mobile equipment and storage medium
CN114387307A (en) Method, apparatus, device, and medium for predicting trajectory of moving object in automatic driving
CN116443032A (en) Method, system, equipment and storage medium for predicting future long-term vehicle speed
CN116907523A (en) Path planning method, path planning device, computer equipment and storage medium
CN114228746B (en) Method and device for predicting motion trail of vehicle
CN113442949B (en) Vehicle control method, device, equipment and storage medium
KR102427366B1 (en) Lane estimation method and apparatus using deep neural network
CN115320572A (en) Vehicle control method and device
CN114945961B (en) Lane changing prediction regression model training method, lane changing prediction method and apparatus
CN114037967A (en) Fusion method and device of multi-source lane lines, vehicle and storage medium
CN113447037A (en) Stroke yaw detection method and device
KR101847113B1 (en) Estimation method and apparatus for information corresponding camera orientation by using image

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination