CN112686281A - Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression - Google Patents
Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression Download PDFInfo
- Publication number
- CN112686281A CN112686281A CN202011423100.7A CN202011423100A CN112686281A CN 112686281 A CN112686281 A CN 112686281A CN 202011423100 A CN202011423100 A CN 202011423100A CN 112686281 A CN112686281 A CN 112686281A
- Authority
- CN
- China
- Prior art keywords
- time
- space
- information
- vehicle
- sample
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 230000014509 gene expression Effects 0.000 title claims abstract description 11
- 238000010586 diagram Methods 0.000 claims abstract description 38
- 238000012549 training Methods 0.000 claims abstract description 10
- 230000002123 temporal effect Effects 0.000 claims description 14
- 238000003860 storage Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 7
- 230000003993 interaction Effects 0.000 claims description 7
- 239000002356 single layer Substances 0.000 claims description 3
- 230000006870 function Effects 0.000 description 9
- 230000007246 mechanism Effects 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 239000010410 layer Substances 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000001902 propagating effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 229910052802 copper Inorganic materials 0.000 description 1
- 239000010949 copper Substances 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000008846 dynamic interplay Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a vehicle track prediction method based on space-time attention and multi-stage LSTM information expression. The method comprises the following steps: acquiring vehicle running track information and space-time map information of road conditions where the vehicle runs, wherein the space-time map information is used for representing position information of the vehicle and surrounding targets in a preset time period and a spatial relation between the vehicle and the surrounding targets along with the time; and inputting the vehicle running track information and the space-time diagram information into a track prediction model to obtain a predicted running track of the vehicle at a subsequent moment, wherein the track prediction model is obtained by pre-training sample space-time diagram information based on a plurality of sample vehicles in the same time period and sample running tracks corresponding to the sample vehicles respectively, and the track prediction model comprises a multistage LSTM codec and a space-time attention module. The invention improves the track prediction accuracy and the robustness under the complex environment, so that the vehicle track prediction can be better used for unmanned driving.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a vehicle track prediction method based on space-time attention and multi-level LSTM information expression.
Background
Vehicle trajectory prediction aims to let the vehicle know how to move next. If the track is accurately predicted, the vehicle can reasonably and effectively infer the future motion of the surrounding vehicle, and make the best planning decision for the driving operation of the vehicle. At present, vehicle track prediction focuses on solving extraction and fusion of interaction information (interaction of data in time sequence and space and semantic relation) between a target vehicle and surrounding vehicles, but the moving track of an object in an actual scene changes due to mutual influence among various targets, so that an independent prediction method simply considering the historical track of the object is difficult to cope with influence caused by association among multiple targets.
The existing vehicle track prediction methods mainly comprise the following three methods:
(1) many of the most advanced LSTM-based trajectory prediction methods model the non-linear temporal dependencies in sequence learning and generation tasks. Are often used for speech recognition and text analysis. While LSTM has the ability to learn and copy a vehicle's historical track, it cannot capture the dependencies between multiple related historical tracks. Therefore, the accuracy of trajectory prediction using LSTM alone is not high.
(2) The Social-LSTM based approach, which processes each track in a scene separately by multiple LSTMs. The LSTMs are then interconnected by Social-LSTM. Unlike conventional LSTM, this pooling layer allows spatially close LSTM to share information with each other, and for one pooling layer of target objects, the hidden states of all LSTM within a certain radius are aggregated together and used as input in the next time step. But this method is not high in capturing capability of mutual information between time series data.
(3) Deep learning approaches attempt to learn data representations from raw data through a multi-layer neural network, learning these features during the course of training. Research has shown that the Convolutional Neural Network (CNN) based method works better than the partial LSTM method, but the Convolutional Neural Network (CNN) belongs to a local extraction feature, i.e., a feature at a short distance. Such methods have insufficient spatial information-interactive capabilities to capture data.
It has been analyzed that in practical situations, autonomously driven vehicles are less likely to be present alone on a road, but are surrounded by vehicles that are not fixed in their surroundings. Existing vehicle trajectory prediction methods typically only consider the state of the vehicle itself and the navigation requirements, and do not fully capture the problem of relevant interactivity between the vehicle and surrounding vehicles. Furthermore, existing research also tends to ignore the correlation between temporal and spatial information.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a vehicle track prediction method based on space-time attention and multi-level LSTM information expression.
The technical scheme of the invention is as follows: a vehicle trajectory prediction method based on spatiotemporal attention and multi-level LSTM information expression is provided. The method comprises the following steps:
acquiring vehicle running track information and space-time map information of road conditions where the vehicle runs, wherein the space-time map information is used for representing position information of the vehicle and surrounding targets in a preset time period and a spatial relation between the vehicle and the surrounding targets along with the time;
and inputting the vehicle running track information and the space-time diagram information into a track prediction model to obtain a predicted running track of the vehicle at a subsequent moment, wherein the track prediction model is obtained by pre-training sample space-time diagram information based on a plurality of sample vehicles in the same time period and sample running tracks corresponding to the sample vehicles respectively, and the track prediction model comprises a multistage LSTM codec and a space-time attention module.
In one embodiment, the space-time diagram information is a dynamic space-time diagram represented by a diagram structure, and is labeled G ═ V, Σ S, Σ T, where V is a set of nodes, Σ S is a set of spatial edges, Σ T is a set of temporal edges, node V represents each individual target, spatial edges represent two inter-node relationships to reflect the interaction between the targets, and temporal edges connect the same node in consecutive time steps to represent a connected graph over time.
In one embodiment, each stage of the multi-stage LSTM codec respectively corresponds to a time edge LSTM for capturing position information between nodes; the space edge LSTM is used for capturing relative space distance information between the node and the neighbor node along with the time; the node LSTM is used for capturing the position information of each node changing along with the time; the spatiotemporal attention module is configured to perform spatial attention weighting and temporal attention weighting.
In one embodiment, the spatiotemporal attention module cascades and averages all edge features of each node with soft attention and implements the spatial domain features and the temporal domain features of common attention with a single layer.
In one embodiment, the sample data set is constructed according to the following steps:
obtaining sample running tracks corresponding to a plurality of sample vehicles in a preset time period;
and processing the sample track data of the plurality of sample vehicles to obtain corresponding sample space-time diagram information, wherein the sample space-time diagram information comprises a sample space-time diagram corresponding to each moment arranged in time sequence in the same time period.
In one embodiment, the sample data set consists of trajectories of real road traffic, covering light, medium and congested road conditions.
Compared with the prior art, the vehicle trajectory prediction method based on the space-time attention mechanism and the multi-stage LSTM information expression and association has the advantages that the space-time attention mechanism and the multi-stage LSTM information expression and association are fully combined, the trajectory prediction accuracy and the robustness in a complex environment are improved, and the vehicle trajectory prediction can be better used for unmanned driving.
Other features of the present invention and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a flow diagram of a vehicle trajectory prediction method based on spatiotemporal attention and multi-level LSTM information expression in accordance with one embodiment of the present invention;
FIG. 2 is an overall architecture diagram based on a spatiotemporal attention model and multi-stage LSTM trajectory prediction, according to one embodiment of the present invention.
Detailed Description
Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that: the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.
The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.
Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate.
In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting. Thus, other examples of the exemplary embodiments may have different values.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
Aiming at the problem that the correlation interactivity between a vehicle and surrounding vehicles is difficult to be comprehensively captured by the current vehicle track prediction and the problem that the correlation between time and space is insufficient in the existing research, the invention provides a vehicle track prediction method based on the information representation and the correlation prediction of space-time attention and multi-stage long-short term memory (LSTM). In short, the invention comprehensively considers the space positions and the moving tracks of the obstacles and the vehicles, constructs a graph convolution network model based on a space-time attention mechanism, designs a coder-decoder based on an LSTM algorithm as a main part, analyzes the moving track of the obstacles, and combines the historical track of the obstacles and the interaction relation between each object and the vehicles to achieve the effect of accurately predicting different tracks. Because the space-time attention mechanism can simultaneously represent the time and space reasoning activities of the area needing attention and the node (node) sharing parameters, excessive parameters are avoided, and the prediction precision and the running speed are improved.
Specifically, referring to fig. 1, the vehicle trajectory prediction method provided by this embodiment includes the following steps:
step S110, a sample data set is constructed based on vehicle running track information and the space-time diagram information of the road condition where the vehicle runs, wherein the space-time diagram information is used for representing the position information of the vehicle and the surrounding targets in a preset time period and the spatial relationship of the vehicle and the surrounding targets along with the time.
The invention takes the starting point that the prediction of the running track of a target vehicle is obtained by training a track prediction model based on the space-time diagram information and the corresponding motion information among multiple targets, and preferably takes the prediction model combining a space-time attention mechanism and a multi-stage LSTM codec as an example for explanation.
For example, a trajectory prediction model is trained and validated on a sample data set consisting of trajectories of real highway traffic. Specifically, each data set was captured at a frequency of 10hz within 45 minutes. Each data set had 15 minutes of light, medium and congested traffic conditions. The data set provides vehicle coordinates projected to a local coordinate system, dividing the entire data set into a training set and a test set. The test set is from 3 subsets of the US-101 and I-80 data sets, and one-fourth of the trace data is selected from each subset as the test set. In one embodiment, the edge set and the point set are semantically divided based on the two data sets, so that the data sets are constructed into a space-time diagram, and then the characteristics are extracted through a network model, and finally the trajectory prediction of the vehicle is realized.
In the description herein, trajectory data is data sampled from the course of motion of a moving object (including vehicles and obstacles) and is used to characterize the position, time, speed of motion, etc. of the moving object. The motion tracks are obtained by sequencing according to the running time sequence and can be used for reflecting the motion trends of time dimension and space dimension.
Step S120, training a track prediction model by using the sample data set, wherein the track prediction model comprises a multi-stage LSTM codec and a space-time attention module.
In order to predict the multi-target future motion trajectory around the moving vehicle more accurately, the importance of the interaction among various targets needs to be clarified, and the more influential sources need to be paid attention to. The preferred embodiment of the invention estimates the corresponding track by adopting a mode of combining a space-time attention mechanism and a multi-stage LSTM codec based on the space-time diagram information and the corresponding motion information among multiple targets, and the overall frame diagram is shown in figure 2.
First, vehicle motion information obtained during trajectory tracking is directly used as an input to the network architecture. And converting the obtained data into space-time diagram information. For example, a data format for processing data in an image coordinate system into a dynamic space-time diagram is represented by G ═ V, Σ S, Σ T, where V represents a node set V, Σ S represents a spatial edge set, and Σ T represents a temporal edge set. Node V represents each individual target (e.g., vehicle or other obstacle); the spatial edge represents the relationship between two nodes, namely, the relationship is used for reflecting the interaction between the nodes; the time edges then connect the same node in successive time steps, representing the connected graph over time.
With reference to fig. 2, a multi-stage LSTM codec is constructed, each stage corresponding to a temporal edge LSTM, a spatial edge LSTM and a node LSTM, respectively. The relation between two nodes being characterised by their relative coordinates, e.g. XV2V3Representing the spatial distance, X, between nodes v2 and v3V2V2Is node V2 that changes over timeLocation. Defining a spatial edge LSTM embedding function phi whose input isAnd embedding the weight matrix WsExpressed as:
whereinRepresenting node XV2All relative spatial distances to its neighbors (e.g., including X)V2V3)。
Spatial edge LSTM obtains embedded input features and previous spatial hidden states from all relevant nodesAnd using normally initialized weight matricesThey are converted. Outputting hidden state vectorsExpressed as:
then, a time-edge LSTM embedding function φ is defined, which is based on the time position of the nodeAnd embedding the weight matrix WtAs input, expressed as:
for an LSTM cell and its inputs: previous temporal hidden stateEmbedded input featuresNormally initialized transition of these inputs to the current hidden state
And then, respectively inputting the space-time diagram information and the attitude information of each object into corresponding multistage LSTM codecs in sequence, and realizing the dependency prediction through a space-time attention module.
For example, for the spatiotemporal attention module, a variant of multi-attention is employed, with soft attention initially set to two simple operations, cascading and averaging all the edge features of each node. Furthermore, a multi-node note is created that has only a single layer, collectively notes the features of the spatial and temporal domains, and stores the note coefficients into a single vector for each time step of node v 2.
Next, the object position coordinates are determinedBy embedding the layer phi, which is then used as an input node to the LSTM, a vector is outputWith the previous hidden stateConnected and then passed to node V2, and the weight matrix W is transformedlstmTo generate a current hidden stateAnd finally estimating the future position of the target from the binary normal distribution N.
In summary, the multi-stage LSTM codec can capture the time-series positional relationship of each individual moving vehicle, and also can capture the relative positional relationship between vehicles, so that the dynamic interaction situation between vehicles or between a vehicle and other obstacles can be better analyzed. And, by combining the spatiotemporal attention module (or component), the corresponding attention can be calculated for each stage of the LSTM module, thereby improving the accuracy, flexibility and applicability of the prediction.
Step S130, inputting the vehicle running track information acquired in real time and the time-space diagram information of the road condition where the vehicle runs into the trained track prediction model to obtain the predicted running track of the vehicle at the subsequent moment.
After the training is completed, in a prediction stage, the vehicle running track information obtained in real time and the corresponding space-time diagram information are used as input, and the predicted vehicle running track can be obtained, wherein the data obtaining process is similar to that in the training stage and is not repeated herein.
In conclusion, the data structure obtained by utilizing the space-time attention mechanism is ingenious in form, the relation between points and edges in the graph structure is effectively utilized, space-time information is represented and predicted through the multi-stage LSTM module, and a reasonable structural framework is combined, so that training parameters are reduced, the accuracy of vehicle trajectory prediction and the flexibility of a model are improved, and the robustness of the model is enhanced. Compared with the prior art, the method has remarkable advantages in the aspects of efficiency and robustness.
The present invention may be a system, method and/or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied therewith for causing a processor to implement various aspects of the present invention.
The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or electrical signals transmitted through electrical wires.
The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.
The computer program instructions for carrying out operations of the present invention may be assembler instructions, Instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present invention are implemented by personalizing an electronic circuit, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA), with state information of computer-readable program instructions, which can execute the computer-readable program instructions.
Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.
These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, by software, and by a combination of software and hardware are equivalent.
Having described embodiments of the present invention, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.
Claims (8)
1. A vehicle track prediction method based on space-time attention and multi-level LSTM information expression comprises the following steps:
acquiring vehicle running track information and space-time map information of road conditions where the vehicle runs, wherein the space-time map information is used for representing position information of the vehicle and surrounding targets in a preset time period and a spatial relation between the vehicle and the surrounding targets along with the time;
and inputting the vehicle running track information and the space-time diagram information into a track prediction model to obtain a predicted running track of the vehicle at a subsequent moment, wherein the track prediction model is obtained by pre-training sample space-time diagram information based on a plurality of sample vehicles in the same time period and sample running tracks corresponding to the sample vehicles respectively, and the track prediction model comprises a multistage LSTM codec and a space-time attention module.
2. The method of claim 1, wherein the space-time graph information is a dynamic space-time graph represented in a graph structure, denoted G ═ V, Σ S, Σ T, where V is a set of nodes, Σ S is a set of spatial edges, Σ T is a set of temporal edges, node V represents each individual target, spatial edges represent two inter-node relationships to reflect interactions between targets, and temporal edges connect the same node in consecutive time steps to represent a connected graph over time.
3. The method of claim 2, wherein each stage of the multi-stage LSTM codec respectively corresponds to a time edge LSTM for capturing position information between nodes; the space edge LSTM is used for capturing relative space distance information between the node and the neighbor node along with the time; the node LSTM is used for capturing the position information of each node changing along with the time; the spatiotemporal attention module is configured to perform spatial attention weighting and temporal attention weighting.
4. The method of claim 2, wherein the spatiotemporal attention module cascades and averages all edge features of each node with soft attention and implements the spatial domain features and the temporal domain features of common attention with a single layer.
5. The method of claim 1, wherein the sample data set is constructed according to the following steps:
obtaining sample running tracks corresponding to a plurality of sample vehicles in a preset time period;
and processing the sample track data of the plurality of sample vehicles to obtain corresponding sample space-time diagram information, wherein the sample space-time diagram information comprises a sample space-time diagram corresponding to each moment arranged in time sequence in the same time period.
6. The method of claim 1, wherein the sample data set consists of trajectories of real road traffic, covering light, medium and congested road conditions.
7. 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 according to any one of claims 1 to 6.
8. A computer device comprising a memory and a processor, on which memory a computer program is stored which is executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 6 when executing the program.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011423100.7A CN112686281A (en) | 2020-12-08 | 2020-12-08 | Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011423100.7A CN112686281A (en) | 2020-12-08 | 2020-12-08 | Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112686281A true CN112686281A (en) | 2021-04-20 |
Family
ID=75446374
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011423100.7A Pending CN112686281A (en) | 2020-12-08 | 2020-12-08 | Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112686281A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112859883A (en) * | 2021-04-25 | 2021-05-28 | 北京三快在线科技有限公司 | Control method and control device of unmanned equipment |
CN113204718A (en) * | 2021-04-22 | 2021-08-03 | 武汉大学 | Vehicle track destination prediction method considering space-time semantics and driving state |
CN113256681A (en) * | 2021-05-26 | 2021-08-13 | 北京易航远智科技有限公司 | Pedestrian trajectory prediction method based on space-time attention mechanism |
CN113291321A (en) * | 2021-06-16 | 2021-08-24 | 苏州智加科技有限公司 | Vehicle track prediction method, device, equipment and storage medium |
CN113435356A (en) * | 2021-06-30 | 2021-09-24 | 吉林大学 | Track prediction method for overcoming observation noise and perception uncertainty |
CN113762473A (en) * | 2021-08-25 | 2021-12-07 | 江苏大学 | Complex scene driving risk prediction method based on multi-space-time diagram |
CN113870319A (en) * | 2021-12-03 | 2021-12-31 | 宁波大学 | Trajectory prediction system and method based on graph convolution characteristic coding and decoding |
CN114372116A (en) * | 2021-12-30 | 2022-04-19 | 华南理工大学 | Vehicle track prediction method based on LSTM and space-time attention mechanism |
CN114461931A (en) * | 2021-12-21 | 2022-05-10 | 国家计算机网络与信息安全管理中心 | User trajectory prediction method and system based on multi-relation fusion analysis |
CN114495036A (en) * | 2021-12-27 | 2022-05-13 | 浙江零跑科技股份有限公司 | Vehicle track prediction method based on three-stage attention mechanism |
CN114648551A (en) * | 2022-05-19 | 2022-06-21 | 武汉深图智航科技有限公司 | Trajectory prediction method and apparatus |
CN114692941A (en) * | 2021-12-30 | 2022-07-01 | 江南大学 | Multi-attention-based company financial prediction method |
CN117216614A (en) * | 2023-09-22 | 2023-12-12 | 哈尔滨工业大学 | Track characterization mining method based on space-time information extraction |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190184561A1 (en) * | 2017-12-15 | 2019-06-20 | The Regents Of The University Of California | Machine Learning based Fixed-Time Optimal Path Generation |
CN111783960A (en) * | 2020-07-09 | 2020-10-16 | 中国人民解放军国防科技大学 | Ship track prediction method and system based on automatic encoder and bidirectional LSTM |
-
2020
- 2020-12-08 CN CN202011423100.7A patent/CN112686281A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190184561A1 (en) * | 2017-12-15 | 2019-06-20 | The Regents Of The University Of California | Machine Learning based Fixed-Time Optimal Path Generation |
CN111783960A (en) * | 2020-07-09 | 2020-10-16 | 中国人民解放军国防科技大学 | Ship track prediction method and system based on automatic encoder and bidirectional LSTM |
Non-Patent Citations (2)
Title |
---|
GUANGXI CHEN 等: "ST-LSTM: SPATIO-TEMPORAL GRAPH BASED LONG SHORT-TERM MEMORY NETWORK FOR VEHICLE TRAJECTORY PREDICTION", 《2020 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)》, 28 October 2020 (2020-10-28) * |
YINGFAN HUANG 等: "STGAT: Modeling Spatial-Temporal Interactions for Human Trajectory Prediction", 《2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV)》, 31 December 2019 (2019-12-31) * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113204718A (en) * | 2021-04-22 | 2021-08-03 | 武汉大学 | Vehicle track destination prediction method considering space-time semantics and driving state |
CN112859883A (en) * | 2021-04-25 | 2021-05-28 | 北京三快在线科技有限公司 | Control method and control device of unmanned equipment |
CN113256681A (en) * | 2021-05-26 | 2021-08-13 | 北京易航远智科技有限公司 | Pedestrian trajectory prediction method based on space-time attention mechanism |
CN113291321A (en) * | 2021-06-16 | 2021-08-24 | 苏州智加科技有限公司 | Vehicle track prediction method, device, equipment and storage medium |
CN113435356A (en) * | 2021-06-30 | 2021-09-24 | 吉林大学 | Track prediction method for overcoming observation noise and perception uncertainty |
CN113762473B (en) * | 2021-08-25 | 2024-04-12 | 江苏大学 | Complex scene driving risk prediction method based on multi-time space diagram |
CN113762473A (en) * | 2021-08-25 | 2021-12-07 | 江苏大学 | Complex scene driving risk prediction method based on multi-space-time diagram |
CN113870319A (en) * | 2021-12-03 | 2021-12-31 | 宁波大学 | Trajectory prediction system and method based on graph convolution characteristic coding and decoding |
CN113870319B (en) * | 2021-12-03 | 2022-03-08 | 宁波大学 | Trajectory prediction system and method based on graph convolution characteristic coding and decoding |
CN114461931A (en) * | 2021-12-21 | 2022-05-10 | 国家计算机网络与信息安全管理中心 | User trajectory prediction method and system based on multi-relation fusion analysis |
CN114495036A (en) * | 2021-12-27 | 2022-05-13 | 浙江零跑科技股份有限公司 | Vehicle track prediction method based on three-stage attention mechanism |
CN114372116A (en) * | 2021-12-30 | 2022-04-19 | 华南理工大学 | Vehicle track prediction method based on LSTM and space-time attention mechanism |
CN114692941A (en) * | 2021-12-30 | 2022-07-01 | 江南大学 | Multi-attention-based company financial prediction method |
CN114372116B (en) * | 2021-12-30 | 2023-03-21 | 华南理工大学 | Vehicle track prediction method based on LSTM and space-time attention mechanism |
CN114648551A (en) * | 2022-05-19 | 2022-06-21 | 武汉深图智航科技有限公司 | Trajectory prediction method and apparatus |
CN117216614A (en) * | 2023-09-22 | 2023-12-12 | 哈尔滨工业大学 | Track characterization mining method based on space-time information extraction |
CN117216614B (en) * | 2023-09-22 | 2024-03-08 | 哈尔滨工业大学 | Track characterization mining method based on space-time information extraction |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112686281A (en) | Vehicle track prediction method based on space-time attention and multi-stage LSTM information expression | |
US11003923B2 (en) | Spatial and temporal information for semantic segmentation | |
EP3535705B1 (en) | Reinforcement learning with auxiliary tasks | |
Haddad et al. | Situation-aware pedestrian trajectory prediction with spatio-temporal attention model | |
Tao et al. | Dynamic and static context-aware lstm for multi-agent motion prediction | |
US11100646B2 (en) | Future semantic segmentation prediction using 3D structure | |
CN113850363A (en) | Techniques for applying driving norms to automated vehicle behavior predictions | |
US11328183B2 (en) | Imagination-based agent neural networks | |
Akan et al. | Stretchbev: Stretching future instance prediction spatially and temporally | |
Xie et al. | Congestion-aware multi-agent trajectory prediction for collision avoidance | |
Zhang et al. | Explainable multimodal trajectory prediction using attention models | |
JP2022539696A (en) | A method for on-device learning of a machine learning network of an autonomous vehicle through multi-stage learning using adaptive hyperparameter sets and an on-device learning device using the same | |
CN113139446A (en) | End-to-end automatic driving behavior decision method, system and terminal equipment | |
Zhang et al. | Pedestrian behavior prediction using deep learning methods for urban scenarios: A review | |
US20200160060A1 (en) | System and method for multiple object tracking | |
CN112115744B (en) | Point cloud data processing method and device, computer storage medium and electronic equipment | |
Wu et al. | Continual pedestrian trajectory learning with social generative replay | |
Stearns et al. | Spot: Spatiotemporal modeling for 3d object tracking | |
US20180197080A1 (en) | Learning apparatus and method for bidirectional learning of predictive model based on data sequence | |
Singh et al. | Multi-scale graph-transformer network for trajectory prediction of the autonomous vehicles | |
Liang et al. | Sequential action-induced invariant representation for reinforcement learning | |
Han | Multimodal intelligent logistics robot combining 3D CNN, LSTM, and visual SLAM for path planning and control | |
Yan et al. | A Survey of Generative AI for Intelligent Transportation Systems | |
Julka et al. | Conditional generative adversarial networks for speed control in trajectory simulation | |
CN115222769A (en) | Trajectory prediction method, device and agent |
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 |