CN113837297A - AI-based behavior prediction method and system for intelligently driving vehicles to circulate - Google Patents
AI-based behavior prediction method and system for intelligently driving vehicles to circulate Download PDFInfo
- Publication number
- CN113837297A CN113837297A CN202111144946.1A CN202111144946A CN113837297A CN 113837297 A CN113837297 A CN 113837297A CN 202111144946 A CN202111144946 A CN 202111144946A CN 113837297 A CN113837297 A CN 113837297A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- output
- encoder
- prediction
- intelligent
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides an AI-based behavior prediction method for the week of an intelligent driving vehicle, which comprises the following steps: step S1: acquiring motion information data of a cycle of the intelligent driving vehicle to obtain acquired data; step S2: processing the input collected data; step S3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction; step S4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction. The invention can improve the prediction time and the prediction accuracy; the method can improve the prediction time and the prediction accuracy, and has important significance for the track prediction of the intelligent vehicle.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a behavior prediction method and a behavior prediction system for the week of an intelligent driving vehicle based on AI.
Background
The intelligent driving is an important component of a strategic emerging industry, the intelligent driving is developed, the intelligent driving has great significance for promoting national science and technology, economy and social life and integrating national strength, the intelligent driving can make up for the defects of human drivers, the intelligent driving is realized, the traffic efficiency can be improved, the safety rate is ensured, and the problem of labor shortage is solved.
Target monitoring and behavior prediction can be seen as two main functions of the perception system of an autonomous vehicle, although they both rely on sensor data, the former aiming at locating and classifying targets in the surroundings of the autonomous vehicle, while the latter provides an understanding of the dynamics of the surrounding objects and predicts their future behavior, behavior prediction playing a key role in autonomous driving applications, as it supports efficient decisions and enables risk assessment.
The LSTM is a special recurrent neural network, can learn long-term dependence, shows strong information acquisition capability, and can show inherent characteristics when processing time series, and in addition, the LSTM can reliably predict the track in a complex scene, and the track belongs to the time series problem, so that it is reasonable to predict the future track by using the LSTM through historical motion data, and the structure of selecting an encoder-decoder as a track prediction model is beneficial to solving the problem that the output of different time steps is irrelevant, processing the seq-seq problem, and adding an attention mechanism can improve the prediction accuracy.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a behavior prediction method and a behavior prediction system for intelligently driving a vehicle to operate based on AI.
The invention provides an AI-based behavior prediction method for the week of an intelligently driven vehicle, which comprises the following steps:
step S1: acquiring motion information data of a cycle of the intelligent driving vehicle to obtain acquired data;
step S2: processing the input collected data;
step S3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction;
step S4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
Preferably, the step S2 includes:
and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
Preferably, the step S3 includes the steps of:
step S3.1: inputting the motion information of the target vehicle into an encoder, converting an input sequence into an intermediate semantic representation through nonlinear transformation, and marking the intermediate semantic representation as Context which is the summary of hidden information;
step S3.2: combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses a vector after Context and attention as an initial state to obtain a predicted position of the next time step;
step S3.3: and circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future.
Preferably, the conversion process of the encoder includes:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
The corresponding elements are added.
Preferably, the conversion process of the decoder comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
The invention also provides an AI-based behavior prediction system for the week of the intelligent driving vehicle, which comprises the following modules:
module M1: acquiring motion information data of a cycle of the intelligent driving vehicle to obtain acquired data;
module M2: processing the input collected data;
module M3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction;
module M4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
Preferably, said module M2 comprises:
and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
Preferably, the module M3 includes the following modules:
module M3.1: inputting the motion information of the target vehicle into an encoder, converting an input sequence into an intermediate semantic representation through nonlinear transformation, and marking the intermediate semantic representation as Context which is the summary of hidden information;
module M3.2: combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses a vector after Context and attention as an initial state to obtain a predicted position of the next time step;
module M3.3: and circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future.
Preferably, the conversion process of the encoder includes:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
The corresponding elements are added.
Preferably, the conversion process of the decoder comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention can improve the prediction time and the prediction accuracy;
2. the method can improve the prediction time and the prediction accuracy, and has important significance for the prediction of the track of the intelligent vehicle.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a schematic flow chart of a behavior prediction method for intelligent AI-based vehicle cycling;
FIG. 2 is a diagram of a trajectory prediction model.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
Referring to fig. 1, the present invention provides an AI-based behavior prediction method for the week of an intelligent driving vehicle, comprising:
Step 2: processing the input collected data; and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates (X, Y) of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
And step 3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction; inputting the motion information of the target vehicle into an encoder, converting an input sequence into an intermediate semantic representation through nonlinear transformation, and marking the intermediate semantic representation as Context which is the summary of hidden information; combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses a vector after Context and attention as an initial state to obtain a predicted position of the next time step; circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future; the number of hidden cells of the LSTM cell bodies of the long-short term memory network is 128, the deep-layer circulating neural network structure comprises two circulating layers, and the learning rate alpha is 0.0005 by using an Adam optimizer.
And 4, step 4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
The conversion process of the encoder comprises:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
The corresponding elements are added.
The attention binding process includes: h ═ H1, H2, …, ht }, is the output of the encoder, M ═ tanh (H), wTIs a parameter to be learned, and the dimension is dW×1，dWRepresenting the dimension of the input vector, the dimension of M being dW×T,α＝softmax(wTM),τ＝HαTThe encoder output after final combination with attention is h*＝tanh(τ)。
The conversion process of the decoder comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
The scene of the data set needs to be complete and has higher sampling frequency; the scene of the data set is set when no motor vehicle passes through frequently, the surrounding road is a multi-lane motor vehicle road, and no other non-motor vehicle exists. To ensure safety, there is a limit to the maximum vehicle speed.
Specifically, the information data collection process is as follows: and acquiring coordinate vectors by acquiring coordinate origin of other motor vehicles relative to the vehicle position through the vehicle-mounted laser radar to establish a coordinate system.
The bi-directional LSTM trajectory prediction for an attention-based encoder-decoder architecture, as shown in fig. 2, includes:
step S10: the encoder and decoder networks are initialized.
Step S20: the encoder input X is initialized.
Step S30: acquiring an output H of an encoder; the output Y of the decoder is obtained.
Step S40: minimizing the loss function of the update network:
and repeating the steps S20-S40 to obtain the final result.
The effect of the two-way LSTM trajectory prediction for the attention-based encoder-decoder architecture can be evaluated by:
the performance of the prediction result can be quantitatively evaluated in the time domain by using the Euclidean metric epsilon, and the method comprises the following steps:
the predicted position is used to determine the Euclidean metric of the position relative to the Ground Truth (GT), which is generated by the lidar. At the predicted time step t, the predicted position of the rider is (X)pre(t),Ypre(t)), GT position is XGT(t),YGT(t))。
The specific prediction time of the Average Euclidean Error (AEE) may be considered. When one of all predictions in the test data with a particular prediction time is given, the AEE shows the average value at each time step. Different prediction times are selected to evaluate the performance of different step sizes.
The invention also provides an AI-based behavior prediction system for the week of the intelligent driving vehicle, which comprises the following modules:
module M1: the motion information data of the whole cycle of the intelligent driving vehicle is collected to obtain collected data.
Module M2: processing the input collected data; and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
Module M3: from the processed data, a bi-directional LSTM with an attention encoder-decoder structure is used for track prediction.
Module M3.1: the motion information of the target vehicle is input into an encoder, and the input sequence is converted into an intermediate semantic representation through nonlinear transformation, which is denoted as Context, which is a summary of hidden information.
Module M3.2: and combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses the vector after Context and attention as an initial state to obtain the predicted position of the next time step.
Module M3.3: and circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future.
Module M4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
The conversion process of the encoder comprises:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
The corresponding elements are added.
The conversion process of the decoder comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
The invention can improve the prediction time and the prediction accuracy; the method can improve the prediction time and the prediction accuracy, and has important significance for the track prediction of the intelligent vehicle.
Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, units provided by the present invention as pure computer readable program code, the system and its various devices, modules, units provided by the present invention can be fully implemented by logically programming method steps in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system and various devices, modules and units thereof provided by the invention can be regarded as a hardware component, and the devices, modules and units included in the system for realizing various functions can also be regarded as structures in the hardware component; means, modules, units for performing the various functions may also be regarded as structures within both software modules and hardware components for performing the method.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.
Claims (10)
1. An AI-based behavior prediction method for the week of an intelligently driven vehicle, characterized by comprising the following steps:
step S1: acquiring motion information data of a cycle of the intelligent driving vehicle to obtain acquired data;
step S2: processing the input collected data;
step S3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction;
step S4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
2. The AI-based behavior prediction method for intelligent vehicle cycling of claim 1, wherein the step S2 comprises:
and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
3. The AI-based behavior prediction method for intelligent vehicle cycling of claim 1, wherein the step S3 comprises the following steps:
step S3.1: inputting the motion information of the target vehicle into an encoder, converting an input sequence into an intermediate semantic representation through nonlinear transformation, and marking the intermediate semantic representation as Context which is the summary of hidden information;
step S3.2: combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses a vector after Context and attention as an initial state to obtain a predicted position of the next time step;
step S3.3: and circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future.
4. The AI-based behavior prediction method for intelligent vehicle cycling of claim 1, wherein the conversion process of the encoder comprises:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
And ≧ is the corresponding element addition.
5. The AI-based behavior prediction method for intelligent vehicle cycling of claim 1, wherein the switching process of the decoder comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
6. An AI-based behavior prediction system for intelligent vehicle cycling, the system comprising the following modules:
module M1: acquiring motion information data of a cycle of the intelligent driving vehicle to obtain acquired data;
module M2: processing the input collected data;
module M3: according to the processed data, adopting bidirectional LSTM with attention encoder-decoder structure to make track prediction;
module M4: and outputting the predicted position of the intelligent driving vehicle according to the track prediction.
7. An AI-based intelligent vehicular activity prediction system as claimed in claim 6, wherein said module M2 includes:
and (3) constructing a coordinate system by taking the intelligent driving vehicle as the origin of coordinates and the advancing direction as the positive direction of the Y axis, and obtaining the position coordinates of the surrounding vehicle after the surrounding vehicle appears around the intelligent driving vehicle.
8. An AI-based intelligent vehicle cycling behavior prediction system according to claim 6, characterized in that said module M3 comprises the following modules:
module M3.1: inputting the motion information of the target vehicle into an encoder, converting an input sequence into an intermediate semantic representation through nonlinear transformation, and marking the intermediate semantic representation as Context which is the summary of hidden information;
module M3.2: combining the output result of the encoder with attention and inputting the output result into a decoder, wherein the decoder uses a vector after Context and attention as an initial state to obtain a predicted position of the next time step;
module M3.3: and circularly inputting the result output in the previous step to obtain a new position, obtaining a prediction sequence of the track, and predicting the position information of k time steps in the future.
9. An AI-based intelligent vehicle cycling behavior prediction system as claimed in claim 6, wherein the encoder switching process comprises:
inputting X ═ { X1, X2, …, Xt }, where Xt ═ Xt, yt }, the length of the vector X indicates the length of inputtable information, and (Xt, yt) indicates the position of the target vehicle at time t in a coordinate system with the host vehicle as the origin of coordinates and the traveling direction as the y-axis; the positive output obtained after the Xt is input into the bidirectional LSTM is
The backward output is
The output of the encoder is
And ≧ is the corresponding element addition.
10. The AI-based behavior prediction system for intelligent weekly vehicle driving according to claim 6, wherein the decoder switching process comprises:
inputting hidden information h*Hiding information h*Outputting the result after attention combination for the encoder; the output Y ═ { Yt +1, Yt + +2, …, Yt + k }, where Yt + k ═ { xt + k, Yt + k } represents the position of the target vehicle at time t + k in a coordinate system with the host vehicle as the origin of coordinates and the direction of travel as the Y-axis, and subscript k represents the length of the output information; and a loss function
Wherein Y't+iAnd k represents the length of the output prediction sequence.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111144946.1A CN113837297A (en) | 2021-09-28 | 2021-09-28 | AI-based behavior prediction method and system for intelligently driving vehicles to circulate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111144946.1A CN113837297A (en) | 2021-09-28 | 2021-09-28 | AI-based behavior prediction method and system for intelligently driving vehicles to circulate |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113837297A true CN113837297A (en) | 2021-12-24 |
Family
ID=78967129
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111144946.1A Pending CN113837297A (en) | 2021-09-28 | 2021-09-28 | AI-based behavior prediction method and system for intelligently driving vehicles to circulate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113837297A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112015842A (en) * | 2020-09-02 | 2020-12-01 | 中国科学技术大学 | Bicycle track prediction automatic driving vehicle risk assessment method and system |
| CN112949597A (en) * | 2021-04-06 | 2021-06-11 | 吉林大学 | Vehicle track prediction and driving manipulation identification method based on time mode attention mechanism |
-
2021
- 2021-09-28 CN CN202111144946.1A patent/CN113837297A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112015842A (en) * | 2020-09-02 | 2020-12-01 | 中国科学技术大学 | Bicycle track prediction automatic driving vehicle risk assessment method and system |
| CN112949597A (en) * | 2021-04-06 | 2021-06-11 | 吉林大学 | Vehicle track prediction and driving manipulation identification method based on time mode attention mechanism |
Non-Patent Citations (1)
| Title |
|---|
| CRIS_LEE卡卡卡: "用于关系提取的基于注意力机制的双向LSTM网络[ACL 2016]", pages 1 - 7, Retrieved from the Internet <URL:https://blog.csdn.net/lrs1353281004/article/details/97970659> * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dai et al. | Modeling vehicle interactions via modified LSTM models for trajectory prediction | |
| Chen et al. | Interpretable end-to-end urban autonomous driving with latent deep reinforcement learning | |
| Veres et al. | Deep learning for intelligent transportation systems: A survey of emerging trends | |
| Mozaffari et al. | Vehicle speed prediction via a sliding-window time series analysis and an evolutionary least learning machine: A case study on San Francisco urban roads | |
| CN109910909B (en) | Automobile track internet interactive prediction method for multi-automobile motion state | |
| CN114495527A (en) | Method and system for collaborative optimization of vehicle paths at internet intersection in mixed traffic environment | |
| Wang et al. | An effective dynamic spatiotemporal framework with external features information for traffic prediction | |
| CN112015842A (en) | Bicycle track prediction automatic driving vehicle risk assessment method and system | |
| Tang et al. | Trajectory prediction for autonomous driving based on multiscale spatial‐temporal graph | |
| Hu et al. | Driver identification using 1D convolutional neural networks with vehicular CAN signals | |
| CN113159403A (en) | Method and device for predicting pedestrian track at intersection | |
| Pisarov | Smart Cars as a Solution for Overpopulation | |
| Guillen-Perez et al. | Learning from Oracle demonstrations—a new approach to develop autonomous intersection management control algorithms based on multiagent deep reinforcement learning | |
| Shao et al. | Failure detection for motion prediction of autonomous driving: An uncertainty perspective | |
| Wu et al. | Graph-based interaction-aware multimodal 2D vehicle trajectory prediction using diffusion graph convolutional networks | |
| CN113837297A (en) | AI-based behavior prediction method and system for intelligently driving vehicles to circulate | |
| CN116795720A (en) | Unmanned driving system credibility evaluation method and device based on scene | |
| CN113962460B (en) | Urban fine granularity flow prediction method and system based on space-time comparison self-supervision | |
| Farajiparvar et al. | Deep learning techniques for traffic speed forecasting with side information | |
| CN115719478A (en) | End-to-end automatic driving method for accelerated reinforcement learning independent of irrelevant information | |
| Naitmalek et al. | Embedded real-time speed forecasting for electric vehicles: a case study on RSK urban roads | |
| Cao et al. | City buses’ future velocity prediction for multiple driving cycle: A meta supervised learning solution | |
| Fu et al. | LSTM-based lane change prediction using Waymo open motion dataset: The role of vehicle operating space | |
| Pradhan et al. | Applications of deep learning in severity prediction of traffic accidents | |
| Li et al. | Personalized trajectory prediction for driving behavior modeling in ramp-merging scenarios |
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 |