CN111391831B - A vehicle following speed control method and system based on the speed prediction of the preceding vehicle - Google Patents
A vehicle following speed control method and system based on the speed prediction of the preceding vehicle Download PDFInfo
- Publication number
- CN111391831B CN111391831B CN202010124684.1A CN202010124684A CN111391831B CN 111391831 B CN111391831 B CN 111391831B CN 202010124684 A CN202010124684 A CN 202010124684A CN 111391831 B CN111391831 B CN 111391831B
- Authority
- CN
- China
- Prior art keywords
- data
- speed
- feature
- preceding vehicle
- features
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000007781 pre-processing Methods 0.000 claims abstract description 18
- 238000013528 artificial neural network Methods 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 24
- 238000005070 sampling Methods 0.000 claims description 16
- 238000013527 convolutional neural network Methods 0.000 claims description 13
- 238000011176 pooling Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 11
- 238000012546 transfer Methods 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 9
- 238000004140 cleaning Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 6
- 238000002360 preparation method Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 18
- 238000010586 diagram Methods 0.000 description 7
- 230000006399 behavior Effects 0.000 description 5
- 230000001133 acceleration Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 230000010354 integration Effects 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000007787 long-term memory Effects 0.000 description 3
- 230000006403 short-term memory Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003534 oscillatory effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/14—Adaptive cruise control
- B60W30/143—Speed control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/14—Adaptive cruise control
- B60W30/16—Control of distance between vehicles, e.g. keeping a distance to preceding vehicle
- B60W30/165—Automatically following the path of a preceding lead vehicle, e.g. "electronic tow-bar"
Landscapes
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Traffic Control Systems (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
本发明提供了一种基于前车速度预测的汽车跟驰速度控制方法及系统，该系统包括：数据预处理模块、特征选择模块、前车速度预测模块、和后车速度控制模块。该方法包括：步骤S1：预处理前车历史速度数据、交通信号数据和路况数据；步骤S2：根据前车历史速度数据、交通信号数据和路况数据，按照基于预设特征选择规则生成选择方案；步骤S3：根据选择方案，输出选择特征至速度预测模型来进行前车速度预测；步骤S4:输入前车预测速度至后车速度控制模型进行速度控制。本发明相比于现有的汽车跟驰速度控制不仅模型结构简单，而且控制效果很好。同时，通过前车速度的预测，可以对后车在将来时刻进行有效的跟驰驾驶策略调整。
The invention provides a vehicle following speed control method and system based on the speed prediction of the preceding vehicle. The system includes a data preprocessing module, a feature selection module, a preceding vehicle speed prediction module, and a rear vehicle speed control module. The method includes: step S1: preprocessing the preceding vehicle historical speed data, traffic signal data and road condition data; step S2: generating a selection scheme based on a preset feature selection rule according to the preceding vehicle historical speed data, traffic signal data and road condition data; Step S3: According to the selection scheme, output the selected feature to the speed prediction model to predict the speed of the preceding vehicle; Step S4: Input the predicted speed of the preceding vehicle to the speed control model of the following vehicle for speed control. Compared with the existing vehicle following speed control, the present invention not only has a simple model structure, but also has a good control effect. At the same time, by predicting the speed of the preceding vehicle, the following vehicle can be effectively adjusted in the future.
Description
Technical Field
The invention relates to the field of intelligent traffic, in particular to an automobile following speed control method and system based on preceding automobile speed prediction.
Background
In the following situation of the automobile, because the following automobile needs to keep the speed change and the safe following distance running which are more consistent with those of the front automobile, ideally, a driver of the rear automobile can predict and estimate the short-term speed of the front automobile in a future period more accurately, and can control the speed of the rear automobile according to the speed and the state of the front automobile. However, the current vehicle-mounted sensor and the global positioning system or the geographic information system can only provide the current speed of the front vehicle and other current driving states and partial environmental information, and the variation trend of the driving speed of the front vehicle cannot be estimated. If the future speed of the front vehicle is not predicted, the change trend of the driving speed of the front vehicle cannot be known, the rear vehicle cannot have sufficient driving decision time, errors in control decision can be caused, driving safety and driving comfort are seriously affected, and energy waste can be caused by behaviors brought by unreasonable driving strategies.
Disclosure of Invention
The invention provides a method and a system for controlling the car following speed based on the prediction of the speed of the front car, aiming at the existing problems, a group of efficient characteristics are selected through a characteristic selection scheme based on mutual information, the efficient characteristics are input into a front car speed prediction model to predict the speed of the front car, and the prediction result is input into a following control system to control the speed. The method and the system provided by the invention improve the safety and driving comfort of the following vehicle and reduce the energy consumption, compared with the prior automobile following control, the method and the system not only have simple sources of prediction information and only need common speed and environmental information, but also have good prediction effect. Meanwhile, the driving behavior of the following vehicle can be continuously adjusted through the PID control based on the speed following.
The technical scheme of the invention is as follows:
in one aspect, a vehicle following speed control system based on a prediction of a preceding vehicle speed, comprising:
the data preprocessing module is used for preprocessing the original acquired data of the front vehicle and extracting characteristics, wherein the original acquired data of the front vehicle comprises speed data of the front vehicle, traffic signal data and road condition data; each feature contains N data collected uniformly over a sampling interval T;
features contained in the raw acquisition data: acceleration in the direction of the X, Y, Z axis of the vehicle, kalman filter value of acceleration in the direction X, Y, Z, altitude, longitude, latitude, pitch, yaw, roll, remaining time of possible rear-end collision, road width, body angle, vehicle angle with respect to curvature of lane, number of currently detected vehicles, distance to preceding vehicle, lane detection state, current number of lanes, speed type, speed, time;
wherein the lane detection state comprises calibration, being initialized, not detected, or running; the speed types comprise a speed greater than a set speed and a speed less than or equal to the set speed;
the characteristic selection module is used for outputting the selected characteristics to the front vehicle speed prediction module according to preset characteristic selection rules according to the characteristics preprocessed by the data preprocessing module;
the front vehicle speed prediction module is used for training a front vehicle speed prediction model according to the selected characteristics output by the characteristic selection module and the front vehicle speed after the set interval time, and outputting a predicted value of the speed of the front vehicle in the future time period;
the rear vehicle speed following control module is used for adjusting and controlling a rear vehicle speed target by taking the safe distance as a control target according to the predicted value of the speed of the front vehicle in the future time period;
the preset feature selection rule is that greedy search is carried out from a set F of features to be selected, the searched features are sequentially added into a set A of target features, and features to which data in the set A belong serve as selected features; the candidate feature set F comprises preceding vehicle speed data, traffic signal data and road condition data, the initial element contained in the target feature set A is the preceding vehicle speed, and the target feature selection standard is the feature with the minimum sum SMI of mutual information values of all elements in the candidate feature set F and the target feature set A.
Further, the greedy search from the set F of the features to be selected means that SMIs of each candidate feature in the set F of the features to be selected and each feature in the set a of the target features are sequentially calculated, and the maximum SMI and the current maximum SMI are selected*To carry outComparing, if the selected maximum SMI is greater than the current maximum SMI*Then the feature f corresponding to the selected maximum SMI is selectedkIs added to A while f is addedkRemove from F and get the current maximum SMI*Update the value of to the selected maximum SMI; the operation is circulated again until each feature in the F and each feature in the A calculate mutual information values; wherein the current maximum SMI*Is 0.
Furthermore, the selected features used for training the preceding vehicle speed prediction model are firstly encoded by the encoder and then decoded by the decoder to be used as input data during the training of the preceding vehicle speed prediction model.
The selected features are the target features, the target features are further encoded and decoded, the encoder compresses the input tensor of the target features into a fixed-dimension vector, and then the decoder reads the vector and converts the vector into a corresponding tensor. In this way, not only is the useful information in the tensor preserved, but the size of the tensor is changed.
Since the target characteristic data is data collected from a vehicle Global Positioning System (GPS), peripheral sensors and a Geographic Information System (GIS). However, the dimensions of the features are not consistent due to different sampling frequencies of various sensors, and the sensors are not suitable for being directly input into a neural network for training. Current processing methods often employ selecting the same sampling frequency, clipping or padding data at different sampling frequencies. The cutting results in a reduction of valid data and a loss of important information in the data. Padding artificially adds information that may introduce noise and erroneous data. Therefore, both cropping and padding may affect the integrity or accuracy of the original data, and therefore the sequence-to-sequence framework proposed in this application, which refers to the structural framework between the encoder and decoder, is adopted to solve this problem.
Further, the encoder comprises a convolutional neural network and an LSTM neural network stacked by linear, wherein the convolutional neural network is composed of one convolutional layer and one pooling layer, and the LSTM neural network is composed of one LSTM layer and one Dropout layer;
the convolution layer in the convolutional neural network is a 1-dimensional convolution layer, the number of convolution kernels is set to be 32, the size of the convolution kernel/the length of a convolution window is set to be 7, filling parameters are set to be same, the shape of an input tensor is set to be (n,1), and n is the number of selected features; the activation function is set to relu, the expression of the relu function: y ═ max (0, x); the pooling layer is a 1-dimensional maximum pooling layer, and the size of a pooling window is set to be 5;
the output dimension of the LSTM layer in the LSTM neural network is set to be 32, the Dropout rate of the Dropout layer is set to be 0.2, and the optimizer is set to be rmsprop.
Further, the decoder is composed of an LSTM layer and two fully-connected layers, wherein the output dimension of the LSTM layer is set to 32; the output dimension of the first fully-connected layer is set to 32 and the activation function is set to relu, and the output dimension of the second fully-connected layer is 1 without using the activation function.
Further, the front vehicle speed prediction model is obtained by training a deep neural network, wherein the deep neural network comprises three linearly stacked fully-connected layers, an optimizer is set to rmsprop, and a loss function is set to mse;
wherein s isreal、spredictionThe real value and the predicted value of the speed are respectively, and m is the number of the predicted time points.
Further, the rear vehicle speed control module comprises a transfer function and a PID controller;
wherein the transfer function is a first order transfer function characterizing the following velocity.
Further, the data preprocessing module comprises:
the data cleaning submodule is used for removing and replacing the abnormal value and the missing value of the input data;
the data normalization submodule is used for carrying out normalization and standardized transformation on the data cleaned by the data cleaning submodule;
and the data preparation submodule is used for dividing the data output by the data normalization submodule into a training set, a verification set and a test set.
The data cleaning submodule is used for screening the characteristics corresponding to the original collected data, and extracting the characteristics which are irrelevant, repeated or have very small data quantity to obtain available characteristics;
the data normalization submodule is used for interpolating the features with the missing amount lower than 1/3 by adopting an interpolation method and supplementing feature data; wherein the standardization treatment adopts a z-score standardization treatment mode;
in another aspect, a method for controlling a follow-up speed of a vehicle based on a prediction of a speed of a preceding vehicle, includes:
step S1: preprocessing the original collected data of the front vehicle and extracting characteristic data; the original collected data of the front vehicle comprises front vehicle speed data, traffic signal data and road condition data; each feature contains N data collected uniformly over a sampling interval T;
step S2: selecting features according to a preset feature selection rule to obtain selected features;
step S3: training a front vehicle speed prediction model by using the selected characteristics and the front vehicle speed after the set interval time, and outputting a predicted value of the speed of the front vehicle in a future time period by using the front vehicle speed prediction model;
step S4: inputting the predicted speed of the front vehicle, and controlling the speed of the rear vehicle by taking the safe distance as a control target;
the preset feature selection rule is that greedy search is carried out from a set F of features to be selected, the searched features are sequentially added into a set A of target features, and features to which data in the set A belong serve as selection features; the candidate feature set F comprises preceding vehicle speed data, traffic signal data and road condition data, the initial element contained in the target feature set A is the preceding vehicle speed, and the target feature selection standard is the feature with the minimum sum of mutual information values of the elements in the candidate feature set F and the elements in the target feature set A.
Further, the front vehicle speed prediction model is obtained by training a deep neural network, wherein the deep neural network comprises three linearly stacked fully-connected layers, an optimizer is set to rmsprop, and a loss function is set to mse;
wherein s isreal、spredictionRespectively representing the real value and the predicted value of the speed, wherein m is the number of predicted time points;
the rear vehicle speed control controls a first order transfer function that characterizes the following speed based on PID.
Advantageous effects
According to the automobile following speed control system and method based on the preceding vehicle speed prediction, a group of efficient features are selected through a feature selection scheme based on mutual information, the efficient features are input into a preceding vehicle speed prediction model to predict the speed of the preceding vehicle, and a prediction result is input into a following speed control system to control the speed. Compared with the prior automobile following control, the method has the advantages that the source of the prediction information is simple, and the prediction effect is good. Meanwhile, the driving behavior of the following vehicle can be continuously adjusted through PID control based on speed following, and errors in control decision are reduced.
Drawings
Fig. 1 is a schematic structural diagram of a following speed control system of an automobile based on a preceding vehicle speed prediction according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a data preprocessing module of a vehicle following speed control system based on a prediction of a preceding vehicle speed according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a feature selection module of a vehicle following speed control system based on a prediction of a preceding vehicle speed according to an embodiment of the invention;
FIG. 4 is a schematic diagram of a rear vehicle speed control module of a vehicle following speed control system based on a prediction of a front vehicle speed according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a preceding vehicle speed prediction module of a following speed control system of an automobile based on preceding vehicle speed prediction according to an embodiment of the present invention;
fig. 6 is a schematic flow chart of a method for controlling a following speed of an automobile based on a prediction of a speed of a preceding automobile according to a second embodiment of the present invention.
Detailed Description
In order to facilitate a better understanding of the present invention, reference will now be made to the following examples.
Example one
Fig. 1 is a schematic structural diagram of an automobile following speed control system based on preceding vehicle speed prediction according to an embodiment of the present invention, including:
the data preprocessing module 100 is configured to preprocess original collected data of a preceding vehicle and extract features, where the original collected data of the preceding vehicle includes speed data of the preceding vehicle, traffic signal data, and road condition data; each feature contains N data collected uniformly over a sampling interval T;
features contained in the raw acquisition data: acceleration in the direction of the X, Y, Z axis of the vehicle, kalman filter value of acceleration in the direction X, Y, Z, altitude, longitude, latitude, pitch, yaw, roll, remaining time of possible rear-end collision, road width, body angle, vehicle angle with respect to curvature of lane, number of currently detected vehicles, distance to preceding vehicle, lane detection state, current number of lanes, speed type, speed, time;
wherein the lane detection state comprises calibration, being initialized, not detected, or running; the speed types comprise a speed greater than a set speed and a speed less than or equal to the set speed;
the feature selection module 200 is used for outputting the selected features to the forward speed prediction module according to the pre-processed features of the data pre-processing module and the preset feature selection rule;
the front vehicle speed prediction module 300 is used for training a front vehicle speed prediction model according to the selected characteristics output by the characteristic selection module and the front vehicle speed after the set interval time, and outputting a predicted value of the speed of the front vehicle in the future time period;
the rear vehicle speed following control module 400 is used for adjusting and controlling the rear vehicle speed target by taking the safe distance as a control target according to the predicted value of the speed of the front vehicle in the future time period;
the preset feature selection rule is that greedy search is carried out from a set F of features to be selected, the searched features are sequentially added into a set A of target features, and features to which data in the set A belong serve as selected features; the candidate feature set F comprises preceding vehicle speed data, traffic signal data and road condition data, the initial element contained in the target feature set A is the preceding vehicle speed, and the target feature selection standard is the feature with the minimum sum of mutual information values of the elements in the candidate feature set F and the elements in the target feature set A.
In this example, the effective range for speed prediction is future speeds within 50s, at least 10 minutes of historical data (sampling frequency up to 10hz) is needed, and the range of speeds in the data sample is 60-120km/h, i.e. after 10 minutes, the speed within 50s is predicted using data with a sampling interval of 10 minutes.
Specifically, the data preprocessing module includes a data cleaning sub-module, a data normalization sub-module and a data preparation sub-module, as shown in fig. 2, in the first embodiment, the data cleaning sub-module is configured to remove and replace an abnormal value and a missing value of input data, for the abnormal value, the data amount is large or the influence of the characteristic on a target characteristic is small, a mode of removing an entire column of characteristics is adopted, otherwise, a mode of replacing by a mean value is adopted, for the missing value, an individual missing value is replaced and filled by a mean value and upper and lower quartiles, and a non-individual missing value is filled by polynomial interpolation; the data normalization submodule is used for carrying out normalization and standardization transformation on the data cleaned by the data cleaning submodule, and carrying out z-score normalization processing after the data is cleaned; and the data preparation submodule is used for dividing the data output by the data normalization submodule into a training set, a verification set and a test set. And dividing the data set into a training set, a verification set and a test set according to the time sequence, wherein the training set is used for model training, the verification set is used for model parameter adjustment, and the test set is used for testing the model effect.
After data preprocessing, feature selection is performed on the data, as shown in fig. 3, performing greedy search from the set F of features to be selected means that SMIs of each candidate feature in the set F of features to be selected and each feature in the set a of target features are sequentially calculated, and the maximum SMI and the current maximum SMI are selected*Comparing, if the selected maximum SMI is greater than the current maximum SMI*Then the feature f corresponding to the selected maximum SMI is selectedkIs added to A while f is addedkRemove from F and get the current maximum SMI*Update the value of to the selected maximum SMI; the operation is circulated again until each feature in the F and each feature in the A calculate mutual information values; wherein the current maximum SMI*Is 0
The mutual information calculation mode is as follows:
wherein, F and a are respectively a certain feature in the candidate feature set F and a certain feature in the target feature set A, and I (F, a) is a mutual information value between F and a. f. ofi，ajThe ith sampling value and the jth sampling value in the sampling time interval T are respectively f and a, and m and n are respectively the number of the f sampling value and the a sampling value.
First, all features (including velocity features) are normalized at 0-1 and calculated, the median of each feature after normalization. The normalized value is 1 or more for the median and the others are 0. Thereby changing the distribution form of the characteristic data into a binomial distribution.
The total number of the samples after the pretreatment is set as r, hkRepresents the number of samples where a-k.
Calculating a joint distribution probability value:
calculating a probability value:
substituting the mutual information value expression to calculate the mutual information value.
The network structure of the preceding vehicle speed prediction module is shown in fig. 4, and the model front end adopts a sequence-to-sequence neural network framework which is composed of an encoder and a decoder. The encoder compresses the input sequence into a fixed length vector, and the decoder decodes the target sequence from the entire vector. A hybrid network of a convolutional neural network and a long and short term memory neural network is used as an encoder, wherein the convolutional neural network is used for preprocessing an input vector, important information is extracted from the convolutional neural network through a convolution and downsampling principle and is used for inputting of a long and short term memory neural network layer, and the long and short term memory neural network layer processes tensors from the convolutional neural network at each time step.
The selected characteristics used for training the front vehicle speed prediction model are firstly coded by the coder and then decoded by the decoder to be used as input data during the training of the front vehicle speed prediction model.
The encoder comprises a convolutional neural network and an LSTM neural network which are linearly stacked, wherein the convolutional neural network consists of a convolutional layer and a pooling layer, and the LSTM neural network consists of an LSTM layer and a Dropout layer;
the convolution layer in the convolutional neural network is a 1-dimensional convolution layer, the number of convolution kernels is set to be 32, the size of the convolution kernel/the length of a convolution window is set to be 7, filling parameters are set to be same, the shape of an input tensor is set to be (n,1), and n is the number of selected features; the activation function is set to relu, the expression of the relu function: y ═ max (0, x); the pooling layer is a 1-dimensional maximum pooling layer, and the size of a pooling window is set to be 5;
the output dimension of the LSTM layer in the LSTM neural network is set to be 32, the Dropout rate of the Dropout layer is set to be 0.2, and the optimizer is set to be rmsprop.
The decoder consists of an LSTM layer and two fully-connected layers, wherein the output dimension of the LSTM layer is set to be 32; the output dimension of the first fully-connected layer is set to 32 and the activation function is set to relu, and the output dimension of the second fully-connected layer is 1 without using the activation function.
The front vehicle speed prediction model is obtained by training a deep neural network, wherein the deep neural network comprises three linearly stacked fully-connected layers, an optimizer is set to rmsprop, and a loss function is set to mse;
wherein s isreal、spredictionThe real value and the predicted value of the speed are respectively, and m is the number of the predicted time points.
The frame of the rear vehicle following speed control model is shown in fig. 5, the input of the control network is a predicted speed value of a front vehicle, the rear vehicle model is simplified into a first-order inertia transfer function, and the control is carried out through a unit speed feedback double closed loop structure. The controller adopts a PID controller to carry out speed following control, and realizes the no-difference tracking of the speed. And (4) parameter identification and adjustment are carried out in an MATLAB simulation environment, and the step response curve is adjusted to be optimal based on prior knowledge. And adjusting the proportional part coefficient according to the set safe distance deviation. Initializing a larger integration time constant randomly, then gradually reducing the constant until the system oscillates, reversely increasing Ti until the system oscillates, and setting the integration time constant of the PID to be 150-180% of the current value by adopting the integration time constant at the moment. The differential is 30% of the non-oscillatory time, as determined by the method of determining the integration time constant.
In summary, according to the automobile following speed control system based on the preceding vehicle speed prediction provided by the invention, a group of efficient features are selected through a feature selection scheme based on mutual information, the efficient features are input into a preceding vehicle speed prediction model to predict the preceding vehicle speed, and a prediction result is input into a following speed control system to control the speed. Compared with the prior automobile following control, the method has the advantages that the source of the prediction information is simple, and the prediction effect is good. Meanwhile, the driving behavior of the following vehicle can be continuously adjusted through PID control based on speed following, and errors in control decision are reduced.
Example two
Figure 6 is a flow chart diagram of a method of controlling the car's follow-up speed with prediction of the speed of the preceding car,
step S1: preprocessing the original collected data of the front vehicle and extracting characteristic data; the original collected data of the front vehicle comprises front vehicle speed data, traffic signal data and road condition data; each feature contains N data collected uniformly over a sampling interval T;
step S2: selecting features according to a preset feature selection rule to obtain selected features;
step S3: training a front vehicle speed prediction model by using the selected characteristics and the front vehicle speed after the set interval time, and outputting a predicted value of the speed of the front vehicle in a future time period by using the front vehicle speed prediction model;
step S4: inputting the predicted speed of the front vehicle, and controlling the speed of the rear vehicle by taking the safe distance as a control target;
the preset feature selection rule is that greedy search is carried out from a set F of features to be selected, the searched features are sequentially added into a set A of target features, and features to which data in the set A belong serve as selection features; the candidate feature set F comprises preceding vehicle speed data, traffic signal data and road condition data, the initial element contained in the target feature set A is the preceding vehicle speed, and the target feature selection standard is the feature with the minimum sum of mutual information values of the elements in the candidate feature set F and the elements in the target feature set A.
The front vehicle speed prediction model is obtained by training a deep neural network, wherein the deep neural network comprises three linearly stacked fully-connected layers, an optimizer is set to rmsprop, and a loss function is set to mse;
wherein s isreal、spredictionRespectively representing the real value and the predicted value of the speed, wherein m is the number of predicted time points;
the rear vehicle speed control controls a first order transfer function that characterizes the following speed based on PID.
For details and description of the implementation of each step in the above method embodiment, reference may be made to the description of the corresponding part of the specific working principle of each module in the above apparatus embodiment, and details are not described here again.
In summary, according to the automobile following speed control method based on the preceding vehicle speed prediction provided by the invention, a group of efficient features are selected through a feature selection scheme based on mutual information, the efficient features are input into a preceding vehicle speed prediction model to predict the preceding vehicle speed, and a prediction result is input into a following speed control system to control the speed. Compared with the prior automobile following control, the method has the advantages that the source of the prediction information is simple, and the prediction effect is good. Meanwhile, the driving behavior of the following vehicle can be continuously adjusted through PID control based on speed following, and errors in control decision are reduced.
The above description is only exemplary of the present invention and should not be taken as limiting the invention, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010124684.1A CN111391831B (en) | 2020-02-27 | 2020-02-27 | A vehicle following speed control method and system based on the speed prediction of the preceding vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010124684.1A CN111391831B (en) | 2020-02-27 | 2020-02-27 | A vehicle following speed control method and system based on the speed prediction of the preceding vehicle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111391831A CN111391831A (en) | 2020-07-10 |
| CN111391831B true CN111391831B (en) | 2021-03-16 |
Family
ID=71419733
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010124684.1A Active CN111391831B (en) | 2020-02-27 | 2020-02-27 | A vehicle following speed control method and system based on the speed prediction of the preceding vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111391831B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112466119B (en) * | 2020-11-26 | 2021-11-16 | 清华大学 | A method and system for predicting vehicle following speed using vehicle-road coordination data |
| CN112550290B (en) * | 2020-12-17 | 2022-07-22 | 江苏大学 | Energy optimal self-adaptive cruise control method and system considering motor energy consumption |
| CN113674525B (en) * | 2021-07-30 | 2022-08-16 | 长安大学 | Signalized intersection vehicle queuing length prediction method based on sparse data |
| CN113723529B (en) * | 2021-09-01 | 2022-11-01 | 清华大学 | Traffic information credible identification method based on speed prediction algorithm |
| CN115788840A (en) * | 2022-09-07 | 2023-03-14 | 上海寻正智能科技有限公司 | An energy-saving method and device for an air compressor station capable of realizing intelligent control |
| CN115384495A (en) * | 2022-09-22 | 2022-11-25 | 中国银行股份有限公司 | A vehicle control method, device, equipment and medium |
| CN116386387B (en) * | 2023-04-19 | 2024-03-08 | 长安大学 | Method and device for predicting car-following behavior of mixed queues of human-driven vehicles |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104271420A (en) * | 2012-03-15 | 2015-01-07 | 谷歌公司 | Modify the behavior of the autonomous vehicle based on the predicted behavior of other vehicles |
| US9002612B2 (en) * | 2012-03-20 | 2015-04-07 | Toyota Motor Engineering & Manufacturing North America, Inc. | Prediction of driver-specific cruise speed using dynamic modeling |
| CN108284836A (en) * | 2018-01-25 | 2018-07-17 | 吉林大学 | A kind of longitudinal direction of car follow-up control method |
| JP2018177148A (en) * | 2017-04-20 | 2018-11-15 | マツダ株式会社 | Driving support device for vehicle |
-
2020
- 2020-02-27 CN CN202010124684.1A patent/CN111391831B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104271420A (en) * | 2012-03-15 | 2015-01-07 | 谷歌公司 | Modify the behavior of the autonomous vehicle based on the predicted behavior of other vehicles |
| US9002612B2 (en) * | 2012-03-20 | 2015-04-07 | Toyota Motor Engineering & Manufacturing North America, Inc. | Prediction of driver-specific cruise speed using dynamic modeling |
| JP2018177148A (en) * | 2017-04-20 | 2018-11-15 | マツダ株式会社 | Driving support device for vehicle |
| CN108284836A (en) * | 2018-01-25 | 2018-07-17 | 吉林大学 | A kind of longitudinal direction of car follow-up control method |
Non-Patent Citations (1)
| Title |
|---|
| 基于贝叶斯网络的联网环境中跟车公况下的前车运动状态预测;张金辉等;《汽车工程》;20190325;第245-251页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111391831A (en) | 2020-07-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111391831B (en) | A vehicle following speed control method and system based on the speed prediction of the preceding vehicle | |
| CN109177982B (en) | An evaluation method of vehicle driving risk considering driving style | |
| US20190287404A1 (en) | Traffic prediction with reparameterized pushforward policy for autonomous vehicles | |
| CN105892471A (en) | Automatic automobile driving method and device | |
| CN112793576B (en) | Lane change decision method and system based on rule and machine learning fusion | |
| CN117325865A (en) | Intelligent vehicle lane change decision method and system for LSTM track prediction | |
| CN113568416B (en) | Unmanned vehicle trajectory planning method, device and computer readable storage medium | |
| EP3640857B1 (en) | Method, vehicle, system, and storage medium for indicating anomalous vehicle scenario using encoder network and discriminator network intermediate layer activation | |
| CN112578419A (en) | GPS data reconstruction method based on GRU network and Kalman filtering | |
| CN120068282B (en) | Real-time wheel centrifugal force prediction method, device and server | |
| CN119691455B (en) | A method for constructing a digital twin system of an intelligent driving vehicle | |
| Long et al. | A method of identifying personalized car-following characteristics for adaptive cruise control system | |
| CN113642114A (en) | Modeling method for humanoid random car following driving behavior capable of making mistakes | |
| CN118722712A (en) | Lane changing decision method for autonomous driving vehicles based on Transformer and improved DQN algorithm | |
| CN118928464A (en) | Method and device for generating automatic driving decision based on hybrid expert model | |
| CN117755325A (en) | Method and device for determining vehicle control parameters, vehicle and storage medium | |
| EP3654246B1 (en) | Method, vehicle, system, and storage medium for indicating anomalous vehicle scenario using encoder network and replay buffer | |
| CN118545054A (en) | Vehicle assisted lane change training method, device and electric vehicle | |
| CN118372853A (en) | A car automatic driving system | |
| CN108909833A (en) | Intelligent automobile rotating direction control method based on Policy iteration | |
| CN113298324B (en) | Track prediction model method, system and device based on deep reinforcement learning and neural network | |
| CN117198064A (en) | Vehicle speed prediction method based on clustering and neural network | |
| CN114445777A (en) | LSTM neural network pedestrian trajectory prediction method based on group behavior optimization | |
| CN119018181B (en) | An end-to-end autonomous driving control system and equipment based on human preference reinforcement learning | |
| CN116088321B (en) | An automatic driving decision-making control method, device and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |