CN113200086A - Intelligent vehicle steering control system and control method thereof - Google Patents
Intelligent vehicle steering control system and control method thereof Download PDFInfo
- Publication number
- CN113200086A CN113200086A CN202110688887.8A CN202110688887A CN113200086A CN 113200086 A CN113200086 A CN 113200086A CN 202110688887 A CN202110688887 A CN 202110688887A CN 113200086 A CN113200086 A CN 113200086A
- Authority
- CN
- China
- Prior art keywords
- model
- neural network
- steering
- steering control
- layer
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
- B62D—MOTOR VEHICLES; TRAILERS
- B62D6/00—Arrangements for automatically controlling steering depending on driving conditions sensed and responded to, e.g. control circuits
- B62D6/001—Arrangements for automatically controlling steering depending on driving conditions sensed and responded to, e.g. control circuits the torque NOT being among the input parameters
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Combined Controls Of Internal Combustion Engines (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses an intelligent vehicle steering control system and a control method thereof, wherein the control system comprises a data acquisition module, a deep learning module and a neural network steering model actual measurement module, the data acquisition module is used for acquiring driving information of an intelligent vehicle, the data acquisition module comprises image information processing, a data acquisition vehicle, seven paths of electromagnetic intensity data and a steering engine steering angle, the deep learning module comprises a data set established through the data acquisition module, the deep learning module also comprises a neural network steering control model, a training model and a testing model, and the neural network steering model actual measurement module comprises a neural network model, seven paths of real-time electromagnetic intensity data, a model testing vehicle, a steering angle and a steering engine. The control system establishes a data set, trains and tests a neural network model on the basis, designs a novel software control algorithm, forms a closed-loop feedback control mode, and greatly improves the running stability of the intelligent vehicle.
Description
Technical Field
The invention relates to the field of intelligent vehicle steering control, in particular to an intelligent vehicle steering control system and a control method thereof.
Background
The background art related to the present invention will be described below, but these descriptions do not necessarily constitute the prior art of the present invention.
The intelligent vehicle is built by taking an electromagnetic, laser or camera as a sensor and can automatically identify a route and track the route, and the speed of the intelligent vehicle is higher and higher, so that the control requirement on the vehicle is higher and higher. In the control system of the intelligent vehicle, steering control is a complex problem because the steering engine of the control object is a nonlinear hysteresis system. By adopting the traditional control method, the phenomena of step-type jump change of steering adjustment, insensitive response to path change, easy generation of overshoot and oscillation and the like are caused, and a satisfactory result is difficult to achieve.
The steering control system is the basis for embodying the intelligent behavior of the intelligent vehicle and is a hotspot and difficulty in the research field of the intelligent vehicle. Along with the development of an intelligent control theory, more and more intelligent control methods are applied to a steering control system of an intelligent vehicle, so that how to select a proper intelligent control method according to different working conditions and road environments becomes a new subject. The steering control method based on classical PID control, fuzzy control, optimal control and adaptive control, sliding mode control and predictive control has the advantages of simple principle, easiness in realization, strong adaptability and the like, and is widely applied to the field of steering system control. Although the PID control scheme is simple and easy to implement, the adaptability to complex working conditions is lacked, the robustness is poor, and the precise control is difficult to realize; the control object of the optimal control is generally a linear time-invariant system, the control accuracy is higher under the condition that the control model is accurate and has no interference, but the robustness to external interference is poor, and the stability of a steering system is easily reduced; membership degree parameters and control rule parameters of the fuzzy control are mainly determined by an expert experience method and a heuristic method, subjectivity is high, and steady-state errors are easy to generate. The control algorithms have the common point that the control accuracy is greatly reduced as long as the tire enters a nonlinear region, and the stability of the system is difficult to ensure, so that the application range of the intelligent automobile is limited.
Disclosure of Invention
In order to solve the technical problem, the invention provides an intelligent vehicle steering control system and a control method thereof.
An intelligent vehicle steering control system comprises a data acquisition module, a deep learning module and a neural network steering model actual measurement module,
the data acquisition module is used for acquiring the driving information of the intelligent vehicle, the data acquisition module comprises image information processing, a data acquisition vehicle, seven paths of electromagnetic intensity data and a steering engine steering angle,
the deep learning module comprises a data set created by the data acquisition module, a neural network steering control model, a training model and a testing model,
the actual measurement module of the neural network steering model comprises a neural network model, seven paths of real-time electromagnetic intensity data, a model test vehicle, a steering angle and a steering engine.
Further, the neural network steering control model comprises selection of hyper-parameters, and the selection of the hyper-parameters comprises learning rate Alpha, size of Mini-batch, Dropout, number of neural network layers, activation function and learning rate attenuation.
An intelligent vehicle steering control method comprises the following steps:
s1, creating a data set, finishing data acquisition by a data acquisition vehicle, acquiring data under two modes of autonomous driving and manual driving respectively, acquiring image information of a driving road by adopting a vision sensor, identifying a path, acquiring steering engine steering angle data, acquiring electromagnetic intensity values on the driving road by adopting seven paths of electromagnetic sensors, constructing the data set, wherein the autonomous driving mode and the manual driving mode are respectively used for acquiring data of a vehicle body in the middle of the driving road and at the edge of the driving road,
s2, constructing a neural network steering control model, wherein the steering control neural network model has 5 layers, the hidden layer is 3 layers, each layer adopts a full-connection layer, the number of nerve cells is 140, 100 and 40 respectively, the 1 st layer is an input layer, 7 paths of input are electromagnetic intensity data, and x is used for inputting the data1，x2，x3…x7Expressed, vector is expressed as X ═ X1，x2，x3…x7]TThe 2 nd layer to the 4 th layer are hidden layers, the 5 th layer is an output layer, and Y is setmThe output of the m layer neuron is
Ym＝f(Wm-1，mYm-1+bm)
In the formula, Wm-1，mWeight matrix representing the m-1 th to m layers, bmRepresenting the bias vector of the m-th layer, f representing the activation function, 2 ≦ m ≦ 5, Y1X denotes an input layer, Y5The output layer is represented by a number of layers,
s3, training a steering control model, training the neural network steering control model by using the acquired steering engine value as a label value and using a manufactured data set and a gradient descent method, calculating a partial derivative (gradient) of a loss function to each weight by using chain derivation, updating the weight according to a gradient descent formula,
s4, testing a steering control model, wherein the model test vehicle autonomously runs along a running road at different speeds, the vibration condition of the vehicle body and the times of road shoulder pressing are observed while the number of running circles is recorded, the stability of the neural network is judged, and the optimal speed at which the neural network model can stably run is obtained.
Further, the method comprises the selection of learning rate Alpha, a proper range is selected for the hyper-parameter by using random Search, the parameter is enabled to uniformly fall in each interval, the probability that the parameter falls between 0.0001-0.001, 0.001-0.1 and 0.1-1 is the same, namely the Alpha values are selected to be 0.001, 0.01, 0.1 and 1 respectively, the Loss value after training and the Accuracy value are compared, and the Alpha value is selected.
And further, selecting 16, 32 and 128 for the Mini-batch value, comparing the trained Loss value with the Accuracy value, and selecting the Mini-batch value.
Further, the method also comprises an excitation function, and the function is selected by comparing the Loss value and the Accuracy value corresponding to the three different excitation functions Sigmoid, Tanh and Relu.
Further, the method also comprises a Dropout value, wherein for the Dropout value, 0, 0.2, 0.3 and 0.5 are selected, and the Dropout value is selected by comparing the trained Loss value with the Accuracy value.
The invention has the beneficial effects that: the control system is characterized in that a data acquisition process is completed by a data acquisition vehicle, a data set for deep learning is created by acquiring data such as electromagnetic intensity, images and steering angles, a neural network model is trained and tested on the basis, and the model test vehicle adopts the trained and tested neural network model and combines real-time electromagnetic intensity data to realize real-time control on steering of the model test vehicle. Through actual measurement, the control system can effectively control steering, so that the intelligent vehicle can complete the driving task under complex road conditions.
Drawings
FIG. 1 is a logic block diagram of a steering control system of an intelligent vehicle according to the present invention,
FIG. 2 is a schematic view of a steering control system of an intelligent vehicle according to the present invention,
FIG. 3 is a schematic diagram of a steering control neural network model of the intelligent vehicle steering control method according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1 and 2, the intelligent vehicle steering control system comprises a data acquisition module, a deep learning module and a neural network steering model actual measurement module,
the data acquisition module is used for acquiring the driving information of the intelligent vehicle, the data acquisition module comprises image information processing, a data acquisition vehicle, seven paths of electromagnetic intensity data and a steering engine steering angle,
the deep learning module comprises a data set created by the data acquisition module, a neural network steering control model, a training model and a testing model,
the actual measurement module of the neural network steering model comprises a neural network model, seven paths of real-time electromagnetic intensity data, a model test vehicle, a steering angle and a steering engine,
in practical application, the neural network steering control model comprises selection of hyper-parameters, and the selection of the hyper-parameters comprises learning rate Alpha, size of Mini-batch, Dropout, number of neural network layers, activation function and learning rate attenuation.
An intelligent vehicle steering control method comprises the following steps:
s1, creating a data set, wherein the data set for deep learning is important, and the richness and integrity of the data have great influence on the completion effect of the driving task. The data acquisition vehicle finishes data acquisition and respectively acquires data in an autonomous driving mode and a manual driving mode,
(1) the vision sensor is adopted to collect the image information of the driving road, the path recognition is carried out, the autonomous driving is realized,
(2) steering angle data of the steering engine is collected, electromagnetic intensity values on a driving road are collected by adopting seven paths of electromagnetic sensors to construct a data set,
(3) the autonomous driving mode and the manual driving mode are respectively used for acquiring data of the vehicle body in the middle of a driving road and at the edge of the driving road,
s2, as shown in figure 3, constructing a neural network steering control model, wherein the steering control neural network model has 5 layers, the hidden layer is 3 layers, each layer adopts a full connection layer, the number of neurons is respectively 140, 100 and 40,
(1) the layer 1 is an input layer, 7 paths of input are used as electromagnetic intensity data, and x is used1，x2，x3…x7Expressed, vector is expressed as X ═ X1，x2，x3…x7]T，
(2) The 2 nd to 4 th layers are hidden layers,
(3) the 5 th layer is an output layer and is provided with YmThe output of the m layer neuron is
Ym＝f(Wm-1，mYm-1+bm)
In the formula, Wm-1，mWeight matrix representing the m-1 th to m layers, bmRepresents the bias vector of the mth layer, f represents the activation function, 2 ≦ m ≦ 5. Y is1X denotes an input layer. Y is5The output layer is represented.
S3, training a steering control model, training the neural network steering control model by using the acquired steering engine value as a label value and using a manufactured data set and a gradient descent method, calculating a partial derivative (gradient) of a loss function to each weight by using chain derivation, updating the weight according to a gradient descent formula,
Train 101times
Epoch 1/1
104572/104572[---------------------------------]-8s 72us/step
loss：0.0182-acc：0.1232
104572/104572[----------------------------------]-2s 20us/step
overFit rate＝0％
times＝101，loss＝0.0108，accu＝0.1235，ir＝0.07408，decay＝0.000002overfit rate＝0
new ir_rat＝0.007386
Train 102times
Epoch 1/1
104572/104572[----------------------------------]-8s 76us/step
loss：0.0172-acc：0.1234
104572/104572[----------------------------------]-2s 18us/step
overFit rate＝5％
times＝102，loss＝0.0108，accu＝0.1235，ir＝0.07386，decay＝0.000002overfit rate＝5％
new ir_rat＝0.007363
s4, testing a steering control model, wherein the model test vehicle autonomously runs along a running road at different speeds, the vibration condition of the vehicle body and the times of pressing the road shoulder are observed while the running turns are recorded, the stability of the neural network is judged, the optimal speed at which the neural network model can stably run is obtained,
further, the step S2 further includes selecting a learning rate Alpha, and when a plurality of different hyper-parameters are tried, there are two methods, i.e., grid Search and random Search. The gird Search is grid Search, a group of candidate values can be given for the hyper-parameters to be adjusted, and the grid Search sequentially combines the candidate values to select the optimal parameter combination; random Search is to randomly Search for hyper-parameters, and for the hyper-parameters to be adjusted, a parameter range is given, and a plurality of hyper-parameters are randomly selected to try in the parameter range, and finally, the optimal hyper-parameter combination is selected. The key to using random Search is to select a proper range for the hyper-parameter, to make the parameter fall in each interval uniformly, to make the probability that the parameter falls between 0.0001-0.001, 0.001-0.1, 0.1-1 equal, namely to select the Alpha values as 0.001, 0.01, 0.1, 1, to compare the trained Loss value and the Accuracy value, the Alpha value is 0.01 best.
Further, the step S2 includes the size of the Mini-batch, and if the Mini-batch is too small, the computing resources are not fully utilized; if Mini-batch is too large, updating the weights and bias is slower. For the Mini-batch value, 16, 32 and 128 are selected, and the Mini-batch value is 32 optimal after the trained Loss value and the Accuracy value are compared.
Further, the step S2 includes an excitation function, which is theoretically optional, but not any function is suitable in practical applications. Common excitation functions are: sigmoid, Tanh, Relu. And selecting a Relu function by comparing the Loss value and the Accuracy value corresponding to the three different excitation functions.
Further, step S2 includes a Dropout value, where the Dropout value is that, in the training process of the deep learning network, the neural network unit is temporarily discarded from the network according to a certain probability. For random gradient descent, each Mini-batch is training a different net because of random discard. In linear space, it is sufficient to learn a feature set of the whole space, but when the data is distributed in a nonlinear discontinuous space, it is better to learn a feature set of the local space. For Dropout values, 0, 0.2, 0.3 and 0.5 are selected, and the Dropout value is 0 best after comparison of the trained Loss value and the Accuracy value.
The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It should be understood by those skilled in the art that the foregoing embodiments are merely illustrative of the technical spirit and features of the present invention, and the present invention is not limited thereto but may be implemented by those skilled in the art.
Claims (7)
1. The utility model provides an intelligence car steering control system which characterized in that: the control system comprises a data acquisition module, a deep learning module and a neural network steering model actual measurement module,
the data acquisition module is used for acquiring the driving information of the intelligent vehicle, the data acquisition module comprises image information processing, a data acquisition vehicle, seven paths of electromagnetic intensity data and a steering engine steering angle,
the deep learning module comprises a data set created by the data acquisition module, a neural network steering control model, a training model and a testing model,
the actual measurement module of the neural network steering model comprises a neural network model, seven paths of real-time electromagnetic intensity data, a model test vehicle, a steering angle and a steering engine.
2. The intelligent vehicle steering control system of claim 1, wherein: the neural network steering control model comprises selection of hyper-parameters, wherein the selection of the hyper-parameters comprises learning rate Alpha, Mini-batch size, Dropout, neural network layer number, activation function and learning rate attenuation.
3. The intelligent vehicle steering control method is characterized by comprising the following steps: the method comprises the following steps:
s1, creating a data set, finishing data acquisition by a data acquisition vehicle, respectively acquiring data in an autonomous driving mode and a manual driving mode, acquiring image information of a driving road by adopting a vision sensor, identifying a path, acquiring steering engine steering angle data, acquiring electromagnetic intensity values on the driving road by adopting seven paths of electromagnetic sensors, constructing the data set, wherein the autonomous driving mode and the manual driving mode are respectively used for acquiring data of a vehicle body in the middle of the driving road and at the edge of the driving road,
s2, constructing a neural network steering control model, wherein the steering control neural network model comprises 5 layers, the hidden layer is 3 layers, each layer adopts a full-connection layer, the number of nerve cells is 140, 100 and 40 respectively, the 1 st layer is an input layer, 7 paths of input are electromagnetic intensity data, and x is used for inputting the electromagnetic intensity data1，x2，x3…x7Expressed, vector is expressed as X ═ X1，x2，x3…x7]TThe 2 nd layer to the 4 th layer are hidden layers, the 5 th layer is an output layer, and Y is setmThe output of the m layer neuron is
Ym＝f(Wm-1，mYm-1+bm)
In the formula, Wm-1，mWeight matrix representing the m-1 th to m layers, bmRepresenting the bias vector of the m-th layer, f representing the activation function, 2 ≦ m ≦ 5, Y1X denotes an input layer, Y5The output layer is represented by a number of layers,
s3, training a steering control model, taking the collected steering engine value as a label value, training the neural network steering control model by using the manufactured data set and a gradient descent method, calculating a partial derivative of a loss function to each weight by using chain type derivation, updating the weight according to a gradient descent formula,
and S4, testing the steering control model, wherein the model test vehicle autonomously runs along a running road at different speeds, the vibration condition of the vehicle body and the times of pressing the road shoulder are observed while the running turns are recorded, the stability of the neural network is judged, and the optimal speed at which the neural network model can stably run is obtained.
4. The intelligent vehicle steering control method according to claim 3, wherein: selecting learning rate Alpha, selecting a proper range for the hyper-parameter by using random Search, enabling the parameter to uniformly fall in each interval, enabling the probability that the parameter falls between 0.0001-0.001, 0.001-0.1 and 0.1-1 to be the same, namely selecting the Alpha values to be 0.001, 0.01, 0.1 and 1 respectively, comparing the trained Loss value with the Accuracy value, and selecting the Alpha value.
5. The intelligent vehicle steering control method according to claim 4, wherein: and the size of the Mini-batch is also included, 16, 32 and 128 are selected for the Mini-batch value, the trained Loss value and the Accuracy value are compared, and the Mini-batch value is selected.
6. The intelligent vehicle steering control method according to claim 5, wherein: the method also comprises an excitation function, and the function is selected by comparing the Loss value and the Accuracy value corresponding to the three different excitation functions Sigmoid, Tanh and Relu.
7. The intelligent vehicle steering control method according to claim 6, wherein: and the Dropout value is selected from 0, 0.2, 0.3 and 0.5, and is selected by comparing the trained Loss value with the Accuracy value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110688887.8A CN113200086A (en) | 2021-06-21 | 2021-06-21 | Intelligent vehicle steering control system and control method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110688887.8A CN113200086A (en) | 2021-06-21 | 2021-06-21 | Intelligent vehicle steering control system and control method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113200086A true CN113200086A (en) | 2021-08-03 |
Family
ID=77022633
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110688887.8A Pending CN113200086A (en) | 2021-06-21 | 2021-06-21 | Intelligent vehicle steering control system and control method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113200086A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114056425A (en) * | 2021-11-23 | 2022-02-18 | 东软集团股份有限公司 | Vehicle steering control method and device, vehicle and storage medium |
| CN115035105A (en) * | 2022-08-10 | 2022-09-09 | 山东西曼克技术有限公司 | Multidimensional data fusion and decision-making method for AGV trolley steering control |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202453734U (en) * | 2011-12-28 | 2012-09-26 | 长安大学 | Automatic-tracking intelligent car controller based on electromagnetic guidance |
| CN106054893A (en) * | 2016-06-30 | 2016-10-26 | 江汉大学 | Intelligent vehicle control system and method |
| DE102016205152A1 (en) * | 2016-03-29 | 2017-10-05 | Avl List Gmbh | Driver assistance system for assisting a driver in driving a vehicle |
| CN110516380A (en) * | 2019-08-30 | 2019-11-29 | 的卢技术有限公司 | A kind of deeply test method and system based on vehicle driving emulation data |
| CN111553188A (en) * | 2020-03-24 | 2020-08-18 | 张印祺 | Deep learning-based end-to-end automatic driving vehicle steering control system |
| CN112278072A (en) * | 2020-11-06 | 2021-01-29 | 吉林大学 | Intelligent vehicle steering safety control system and control method |
-
2021
- 2021-06-21 CN CN202110688887.8A patent/CN113200086A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202453734U (en) * | 2011-12-28 | 2012-09-26 | 长安大学 | Automatic-tracking intelligent car controller based on electromagnetic guidance |
| DE102016205152A1 (en) * | 2016-03-29 | 2017-10-05 | Avl List Gmbh | Driver assistance system for assisting a driver in driving a vehicle |
| CN106054893A (en) * | 2016-06-30 | 2016-10-26 | 江汉大学 | Intelligent vehicle control system and method |
| CN110516380A (en) * | 2019-08-30 | 2019-11-29 | 的卢技术有限公司 | A kind of deeply test method and system based on vehicle driving emulation data |
| CN111553188A (en) * | 2020-03-24 | 2020-08-18 | 张印祺 | Deep learning-based end-to-end automatic driving vehicle steering control system |
| CN112278072A (en) * | 2020-11-06 | 2021-01-29 | 吉林大学 | Intelligent vehicle steering safety control system and control method |
Non-Patent Citations (1)
| Title |
|---|
| 孙艺铭等: "基于BP神经网络的智能车电磁导航控制算法", 《科学技术创新》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114056425A (en) * | 2021-11-23 | 2022-02-18 | 东软集团股份有限公司 | Vehicle steering control method and device, vehicle and storage medium |
| CN115035105A (en) * | 2022-08-10 | 2022-09-09 | 山东西曼克技术有限公司 | Multidimensional data fusion and decision-making method for AGV trolley steering control |
| CN115035105B (en) * | 2022-08-10 | 2022-10-25 | 山东西曼克技术有限公司 | Multidimensional data fusion and decision-making method for AGV trolley steering control |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111899510B (en) | Intelligent traffic system flow short-term prediction method and system based on divergent convolution and GAT | |
| Wang et al. | Capturing car-following behaviors by deep learning | |
| CN106529818B (en) | Water quality assessment Forecasting Methodology based on Fuzzy Wavelet Network | |
| CN111047085B (en) | Hybrid vehicle working condition prediction method based on meta-learning | |
| CN110991027A (en) | Robot simulation learning method based on virtual scene training | |
| CN108010029A (en) | Fabric defect detection method based on deep learning and support vector data description | |
| Li et al. | Data augmented deep behavioral cloning for urban traffic control operations under a parallel learning framework | |
| CN113200086A (en) | Intelligent vehicle steering control system and control method thereof | |
| Liu et al. | Modeling and simulation of robot inverse dynamics using LSTM-based deep learning algorithm for smart cities and factories | |
| CN110000781B (en) | Development network-based mobile robot motion direction pre-decision method | |
| CN108960421A (en) | The unmanned surface vehicle speed of a ship or plane online forecasting method based on BP neural network of improvement | |
| CN111967087A (en) | Neural network-based online vehicle decision control model establishing and evaluating method | |
| CN115952736A (en) | Multi-agent target collaborative search method and system | |
| CN108073978A (en) | A kind of constructive method of the ultra-deep learning model of artificial intelligence | |
| Zhao et al. | A driving risk prediction algorithm based on PCA-BP neural network in vehicular communication | |
| CN117370766A (en) | Satellite mission planning scheme evaluation method based on deep learning | |
| Khidhir et al. | Comparative Transfer Learning Models for End-to-End Self-Driving Car | |
| CN108647784A (en) | A kind of lifelong machine learning method based on depth belief network | |
| Li et al. | Road pavement identification based on acceleration signals of off-road vehicles using the batch normalized recurrent neural network | |
| CN114881308A (en) | Internet vehicle speed prediction method based on meta-learning | |
| CN112580702A (en) | Multi-sensor cooperative sensing method and device | |
| Toubeh et al. | Risk-aware planning by confidence estimation using deep learning-based perception | |
| CN112052905B (en) | Method for extracting multi-operation fatigue features of driver based on recurrent neural network | |
| Gu et al. | Interpretable Deep Knowledge Tracing with Graph Relationship Information | |
| CN108074007A (en) | A kind of Prediction of Stock Index method of the ultra-deep study of artificial intelligence |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210803 |