CN112597597A - Road feel simulation method based on K-Medoids and BP neural network - Google Patents

Road feel simulation method based on K-Medoids and BP neural network Download PDF

Info

Publication number
CN112597597A
CN112597597A CN202011605208.8A CN202011605208A CN112597597A CN 112597597 A CN112597597 A CN 112597597A CN 202011605208 A CN202011605208 A CN 202011605208A CN 112597597 A CN112597597 A CN 112597597A
Authority
CN
China
Prior art keywords
medoids
test data
road feel
neural network
test
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.)
Withdrawn
Application number
CN202011605208.8A
Other languages
Chinese (zh)
Inventor
赵蕊
蔡锦康
邓伟文
丁娟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Tianxingjian Intelligent Technology Co ltd
Original Assignee
Zhejiang Tianxingjian Intelligent Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang Tianxingjian Intelligent Technology Co ltd filed Critical Zhejiang Tianxingjian Intelligent Technology Co ltd
Priority to CN202011605208.8A priority Critical patent/CN112597597A/en
Publication of CN112597597A publication Critical patent/CN112597597A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/22—Design optimisation, verification or simulation using Petri net models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Health & Medical Sciences (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Steering Control In Accordance With Driving Conditions (AREA)

Abstract

The invention discloses a road sense simulation method based on K-Medoids and BP neural networks, which comprises the following steps: carrying out a real vehicle test and collecting test data; preprocessing test data; clustering test data by using a K-Medoids clustering algorithm and dividing a training data set and a test data set; training a road feel model based on K-Medoids and BP neural networks by using a BP neural network algorithm; testing a road feel model based on a K-Medoids and BP neural network; and performing road feel simulation according to the obtained road feel simulation model based on the K-Medoids and BP neural network. The method uses real vehicles to collect test data, establishes the road feel simulation model based on the K-Medoids clustering algorithm and the BP neural network algorithm, is easy to implement in the modeling implementation process, short in modeling time, high in model calculation speed, high in precision and good in robustness, and has obvious advantages compared with the prior art.

Description

Road feel simulation method based on K-Medoids and BP neural network
Technical Field
The invention relates to the field of vehicles, in particular to a road feel simulation method based on a K-Medoids and BP neural network.
Background
The steering road feel, also called steering force feel and steering wheel feedback torque, refers to the reverse resistance torque felt by the driver through the steering wheel feedback torque. The road feel can enable a driver to know the road condition in real time, so that corresponding decisions can be made. Therefore, it is important for a vehicle or a driving simulator using a steer-by-wire system to generate a road feel similar to that of a conventional steering system using a motor or the like. However, the road sensing simulation system realized by the prior art has the defects of low precision, poor real-time performance and the like.
The utility model patent with the application number of CN201420478919.7 and the name of "force feeling analog system based on C-EPS structure" discloses a force feeling analog system based on C-EPS structure, and it mainly simulates the road feeling through the mechanism modeling method, and the parameter that needs to be adjusted is numerous, and the precision is difficult to guarantee.
Disclosure of Invention
In order to solve the above technical problems, the present invention aims to provide a road feel simulation method with easy implementation of modeling process, fast calculation speed and high precision, which establishes a road feel simulation model based on a K-media (K center point) clustering algorithm and a BP neural network algorithm with real vehicle test data.
In order to achieve the above object, the present invention provides a road feel simulation method based on K-Medoids and BP neural networks, comprising the following steps:
step one, carrying out a real vehicle test and collecting test data: the method comprises the following steps that a driver performs a real vehicle test, a vehicle runs in a test road, and collected test data comprise vehicle speed, vehicle lateral acceleration, vehicle yaw velocity, vehicle vertical load, a driver steering wheel corner, steering wheel angular velocity and steering wheel moment;
step two, test data preprocessing: carrying out normalization processing on the test data after removing abnormal points to obtain a normalized test data set;
step three, clustering the normalized test data: clustering the normalized test data by using a K-Medoids clustering algorithm, dividing the normalized test data into a plurality of data classes after clustering, and dividing the clustered test data into a clustered training data set and a clustered test data set;
step four, training a road feel model based on the K-Medoids and the BP neural network: when a road feel simulation model based on K-Medoids and BP neural networks is trained by using a clustered training data set and a BP neural network algorithm, input variables of the BP neural network model are vehicle speed, vehicle lateral acceleration, vehicle yaw velocity, vehicle vertical load, driver steering wheel turning angle and steering wheel angular velocity, output variables are steering wheel moments, and the same number of road feel simulation models based on the K-Medoids and the BP neural networks are obtained through training;
step five, testing a road feel model based on a K-Medoids and BP neural network; testing the obtained road feel simulation model based on the K-Medoids and BP neural network by using the clustered test data set;
step six, judging whether a supplementary test is needed;
and seventhly, carrying out road feel simulation according to the obtained road feel simulation model based on the K-Medoids and the BP neural network.
Further, in the real vehicle test of the step one: the test road types comprise urban roads, expressways, suburban roads and rural roads; the vehicle running working conditions comprise straight running, backing, turning, pivot steering, uphill and downhill working conditions.
Further, in step two, the removed abnormal points include data points beyond the normal value range, data points with severely deviated distribution, and data points with a variation range beyond the normal range.
The data points beyond the normal value range are defined as: and (3) acquiring a certain data point in a certain real vehicle test, wherein the numerical value of one or more variables exceeds the actual normal value range of the corresponding variable of the real vehicle test. If the maximum vehicle speed is only 50km/h in a certain test, the data points with the vehicle speed value more than 50km/h in the data collected in the test are all out of range points. For another example, if the steering wheel angle range is [ -300 °,300 ° ] in a certain test, the points of the steering wheel angle values exceeding [ -300 °,300 ° ] in the data collected in the test are all out of the normal range.
The heavily distributed data points are defined as: and calculating the standard deviation of each variable of the test data acquired in a certain real vehicle test, and if the numerical value of one or more variables of a certain data point is more than 2.5 times of the standard deviation of the corresponding variable or less than minus 2.5 times of the standard deviation of the corresponding variable, determining that the distribution of the data point is seriously deviated.
The data points with the variation amplitude exceeding the normal range are defined as follows: the maximum instantaneous change amplitude of each variable under the normal condition is preset, and if the absolute value of the difference value of one or more variable values of a certain data point relative to the corresponding variable value of the previous data point in the actual test data set is larger than the maximum instantaneous change amplitude of the related variable, the maximum instantaneous change amplitude of each variable exceeds the normal range. For example, when a high-speed driving test is performed using a small passenger car, the expert confirms that the maximum instantaneous change amplitude of the steering wheel torque is 0.3N, and when the absolute value of the difference between the steering wheel torque value and the previous data point in all the data points is greater than 0.3N, the change amplitude is regarded as a point beyond the normal range.
Further, in step two, the test data is normalized according to the following formula:
Figure BDA0002870276530000031
wherein i is a data number, j is a variable number, and xi,jDenotes the j variable, X, in the non-normalized i group of datajRepresents the set of variable data values corresponding to all j, min represents the minimum value, and max represents the maximum value.
Further, in the third step, the normalized test data set is divided into a training data set after clustering and a test data set after clustering by a random division method. Specifically, the method for randomly dividing includes: the normalized test data set is randomly divided into a clustered training data set and a clustered test data set according to a certain proportion of the number of data points. In a preferred embodiment, the ratio is 5: 1. The number of data points in the training data set is typically (but not necessarily) greater than the number of data points in the test data set.
Preferably, in step three, the clustering step of the K-Medoids algorithm is as follows:
(1) and determining the number k of the required communities. In a specific embodiment, the number k of the communities is 5;
(2) randomly selecting k data points in a data set to be clustered as central points of k clusters;
(3) calculating Euclidean distances from all non-central points to k central points determined in the previous step, wherein a community corresponding to the central point closest to the data point is a community to which the data point belongs;
(4) sequentially selecting one point in each community, calculating the sum of Euclidean distances between the point and all other points in the community where the point is located at present, and taking the point with the minimum sum of the Euclidean distances as a new central point of the community;
(5) repeating the steps (2) and (3) until the central point of each cluster is not changed;
preferably, in step four, when the road feel simulation model based on the K-Medoids and the BP neural network is trained, the BP neural network model has 2 hidden layers, each hidden layer has 10 nodes, 1 input layer and 1 output layer; all the activation functions of all the nodes are Sigmoid functions and are all fully connected; the learning function uses the learngdm function. The upper iteration limit is 1000 generations.
When a road feel simulation model based on a BP neural network is trained, the model obtained by training the same type of training data points is related to the type of the data points, namely the road feel simulation model corresponding to a certain type of training data points can only be used for predicting the steering wheel torque of the type of data points. After the training data points of multiple types are trained, multiple corresponding road feel simulation models are obtained.
The related parameter determining steps for training the BP neural network are as follows:
1) determining input parameters and output parameters of a BP neural network;
2) determining the number of hidden layers of the BP neural network;
3) determining the number of nodes of the hidden layer;
4) determining an activation function and an integral learning function of each node;
5) training a BP neural network model using the training data;
6) and judging whether to return to the step 2) according to the model test result. And if the training result meets the requirement, directly carrying out the next step without returning, otherwise, returning to the step 2).
Further, when testing the road sensing simulation model based on the K-Medoids and BP neural network, the mean square error, i.e. the MSE value, can be used, but is not limited to the use, as the criterion of the model quality. When the road feel simulation model based on the K-Medoids and BP neural network is tested by using the clustered test data set, the steps are as follows:
1) taking out the test data points in the test data set after clustering, and inputting the numerical values of the input variables corresponding to the test data points into the road feel simulation model corresponding to the class to which the test data points belong to obtain a predicted steering wheel torque value;
2) and calculating the MSE value between the predicted steering wheel moment value and the real steering wheel moment value through the model of the test data points of the whole clustered test data set after calculation.
And when judging whether a supplementary test is needed, if the MSE value is smaller than a preset threshold value alpha, the road feel simulation model based on the data drive obtained by training is considered to be acceptable, and the supplementary test is not needed. Otherwise it is not acceptable. The threshold value alpha is determined empirically by an expert. In a preferred embodiment, the threshold α is set to 0.15.
Due to the adoption of the technical scheme, the invention achieves the following technical effects: the method is based on data acquired by a real vehicle test, 5 central points are obtained after clustering through a K-Medoids clustering algorithm, modeling is carried out by adopting a BP neural network algorithm, 5 road feel simulation models for predicting steering wheel feedback torque are obtained, and the steering wheel feedback torque is simulated according to the obtained 5 road feel simulation models, so that the problems that the model precision of the traditional mechanism modeling is low, the real-time performance in the application process is difficult to guarantee and the like are solved; compared with the prior art, the road feel simulation model has the advantages of convenient data acquisition, easy implementation of the modeling process, short modeling time period, high model technology speed, high precision and good robustness.
Drawings
Fig. 1 is a flow chart of modeling steps in a road feel simulation method based on K-Medoids and BP neural networks according to the invention.
FIG. 2 is a graph of steering wheel torque collected for a vertical parking space in an embodiment of the present invention.
FIG. 3 is a partial graph of a model test curve for performing a model test according to an embodiment of the present invention.
Detailed Description
In order to make the technical solution of the embodiments of the present invention better understood, the technical solution of the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments, 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 obtained by equivalent changes and modifications by one skilled in the art based on the embodiments of the present invention, shall fall within the scope of the present invention.
Example one
Referring to fig. 1 to 3, the present embodiment provides a road feel simulation method based on K-media and BP neural networks, including modeling steps S1-S6, and a model application step. Steps S1-S6 of the modeling process are described in detail below in conjunction with FIG. 1.
S1, carrying out real vehicle test and collecting test data
And selecting a driver to carry out a real vehicle test, and driving the vehicle to run in the test road. The test road types include, but are not limited to, urban roads, expressways, suburban roads, rural roads, and the like; vehicle driving conditions include, but are not limited to, straight ahead, reverse, turning, pivot steering, uphill, downhill, and the like.
The selected drivers had a driving age of 3.5 years and were driven no less than 5.5 hours per week in the last year. The data acquisition frequency was 100 Hz.
The collected test data includes vehicle speed, vehicle lateral acceleration, vehicle yaw rate, vehicle vertical load, driver steering wheel angle, steering wheel angular velocity, and steering wheel moment. As shown in fig. 2, the steering wheel torque data of the vertical vehicle parking condition collected in the test of the present embodiment is represented by an actual steering wheel torque (Nm) -time(s) curve.
S2, preprocessing test data
The mode of processing the test data comprises removing abnormal points and normalizing the data. The removed abnormal points include data points outside the normal value range, data points with severely deviated distribution and data points with the variation amplitude exceeding the normal range. The embodiment further includes filtering the abnormal points of the test data by using a filter, and the method for deleting the abnormal points by using the filter includes: and inputting the test data into a low-pass filter, a high-pass filter or a band-pass filter at one time according to the original acquisition sequence to obtain the filtered test data.
After the abnormal points are removed, the test data are normalized according to the following formula, and normalized test data are obtained:
Figure BDA0002870276530000051
wherein i is a data number, j is a variable number, and xi,jDenotes the j variable, X, in the non-normalized i group of datajRepresenting a set consisting of variable data values corresponding to all j; min represents the minimum value of the relevant variable after the abnormal point is removed; max represents the maximum value of the correlation variable after the outlier is removed.
S3, clustering test data after normalization
And (5) clustering the normalized test data by using a K-Medoids clustering algorithm, wherein the number of communities in clustering is set to be 5.5 central points are obtained after clustering, and the normalized test data are divided into 5 data classes according to 5 central coordinates. After clustering, dividing according to a random division method according to the ratio of the number of data points of 5:1 to obtain a training data set after clustering and a test data set after clustering.
S4, training a road feel model based on K-Medoids and BP neural network
And training by using the clustered training data set and the BP neural network algorithm to obtain the road feel simulation model based on the K-Medoids and the BP neural network, wherein the quantity of the road feel simulation model is the same as that of the data set.
When the model is trained, the model obtained by training the same type of modeling data points is related to the related type, namely, the data points of a certain type of model can only be used for predicting the data points of the type. The number of the models obtained by training the training data points is the same as the number of the types of the data points, and the number of the road sense simulation models based on the K-Medoids and the BP neural network obtained in the embodiment is 5. The input variables of the BP neural network model comprise vehicle speed, vehicle lateral acceleration, vehicle yaw velocity, steering wheel turning angle, steering wheel angular velocity and vehicle vertical load; the output variable is the steering wheel torque.
1) Determining input parameters and output parameters of the BP neural network: the input variables comprise vehicle speed, vehicle lateral acceleration, vehicle yaw rate, steering wheel angle and steering wheel angular speed; the output variable is steering wheel torque;
2) determining the number of hidden layers of the BP neural network: selecting 2 hidden layers;
3) determining the number of nodes of the hidden layer: each hidden layer node is determined to be 10;
4) determining an activation function and a whole learning function of each node: the activation function is determined as a Sigmoid function, and the learning function uses a learngdm function;
5) training a BP neural network model using the training data;
6) and judging whether to return to the step 2) according to the model test result. And if the training result meets the requirement, directly carrying out the next step without returning, otherwise, returning to the step 2).
Specifically, the process of training the BP neural network is as follows:
note: i denotes the number of the neural network layer, j denotes the number of the neuron, and m denotes the output layer, i.e., the maximum value of i.
A forward process:
1. calculating input and output values for each neuron
1) Input layer neurons: input value X1,j＝Pj(input layer, i ═ 1); output value Y1,j＝X1,j
2) Hidden layer neurons: input value
Figure BDA0002870276530000061
And (3) outputting a value: y isi,j＝f(Xi,j)
3) Output layer neurons: input value
Figure BDA0002870276530000071
(output layer, i ═ m);
and (3) outputting a value: y ism,j＝Xm,j
And (3) reversing the process:
1) calculate each output neuron error: ej＝Dj-Yj
Calculating an objective function value:
Figure BDA0002870276530000072
2) updating the weight: the calculation of the activation function is only performed at the hidden layer, so the weights of the input layer need to be updated, and the weights of the output layer are not used and do not need to be updated.
First of all, the first step is to,
Figure BDA0002870276530000073
order to
Figure BDA0002870276530000074
Then:
Figure BDA0002870276530000075
since i +1>1, it will not be an input layer, for the hidden and output layers:
Figure BDA0002870276530000076
calculating deltai,k：
Figure BDA0002870276530000077
If i +1 is the output layer, i +1 ═ m, then:
Figure BDA0002870276530000078
Figure BDA0002870276530000079
δi,k＝f(X(i+1,k))'·(Y(i+1),k-Dk)＝(Y(i+1),k-Dk)
if i +1 is a hidden layer, i.e., i +1< m, then
Figure BDA0002870276530000081
Figure BDA0002870276530000082
Then:
Figure BDA0002870276530000083
therefore, the first and second electrodes are formed on the substrate,
ΔWi,j,k＝ηδi,k·Yi,j
1) updating the threshold value: the thresholds of the neurons of the input and output layers do not need to be updated, for the hidden layer:
if the (i + 1) th layer is an output layer, i.e. i +1 ═ m,
Figure BDA0002870276530000084
then
Δθi,j＝η·(Y(i+1),k-Dk)
If the (i + 1) th layer is an implied layer, i.e. i +1< m,
Figure BDA0002870276530000085
Figure BDA0002870276530000086
then:
Δθi,j＝η·δi-1,j
the above process is repeated, and only when the error of the output layer reaches the preset threshold value, or the iteration frequency reaches the maximum value.
In this embodiment, the BP neural network model has 2 hidden layers, each hidden layer has 10 nodes, 1 input layer and 1 output layer; all the activation functions of all the nodes are Sigmoid functions and are all fully connected; the learning function uses a learngdm function; the upper iteration limit is 1000 generations.
The Hewlett packard Z1G6 workstation was used for training in this example, and the total training time of 5 road sense simulation models was 2 hours and 36 minutes.
S5, testing the road feel model based on the K-Medoids and BP neural network
The method for testing the road feel simulation model based on the K-Medoids and BP neural network by using the clustered test data set comprises the following steps:
1) taking out the test data points in the test data set after clustering, and inputting the numerical values of the input variables corresponding to the test data points into the road feel simulation model corresponding to the class to which the test data points belong to obtain a predicted steering wheel torque value;
2) and calculating the MSE value between the predicted steering wheel moment value and the real steering wheel moment value through the model of the test data points of the whole clustered test data set after calculation.
As shown in fig. 3, which represents a local part of the model test curve, the steering wheel moment-time curve (sim, solid line) of the test data point predicted by the model almost coincides with the real steering wheel moment-time curve (real, dashed line), and the MSE value is calculated to be 0.12508.
S6, judging whether the modeling is successful or not
And when judging whether a supplementary test is needed, if the MSE value is smaller than a preset threshold value alpha, determining that the road feel simulation model based on the K-Medoids and BP neural network obtained by training is acceptable and not needing the supplementary test. Otherwise it is not acceptable. The MSE value obtained in this embodiment is 0.12508, which is smaller than the preset threshold α of 0.15, and the obtained model is acceptable.
After the modeling is finished, the road feel simulation method further comprises the steps of predicting the moment of a steering wheel by using the 5 obtained road feel simulation models, inputting the 5 obtained road feel simulation models into a driving simulator, acquiring running state parameters such as the speed, the lateral acceleration, the yaw angular velocity, the steering angle, the steering wheel angular velocity and the vertical load of the vehicle of the simulated vehicle in real time when a simulated driving test is carried out on the driving simulator, inputting the running state parameters into the road feel simulation models as input variables, determining the data types of the simulated vehicle according to the correlation degree of the running state parameters and the 5 central coordinates, calculating the moment value of the steering wheel through the road feel simulation models corresponding to the data types, and controlling the steering wheel in real time according to the moment value of the steering wheel, so that more vivid road feel is simulated.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention; also, the above description should be understood as being readily apparent to those skilled in the relevant art and can be implemented, and therefore, other equivalent changes and modifications without departing from the concept disclosed herein are intended to be included within the scope of the present invention.

Claims (10)

1. A road sense simulation method based on K-Medoids and BP neural networks is characterized by comprising the following steps:
step one, carrying out a real vehicle test and collecting test data: the method comprises the following steps that a driver performs a real vehicle test, a vehicle runs in a test road, and collected test data comprise vehicle speed, vehicle lateral acceleration, vehicle yaw velocity, vehicle vertical load, a driver steering wheel corner, steering wheel angular velocity and steering wheel moment;
step two, test data preprocessing: carrying out normalization processing on the test data after removing abnormal points to obtain a normalized test data set;
step three, clustering the normalized test data: clustering the normalized test data by using a K-Medoids clustering algorithm, dividing the normalized test data into a plurality of data classes after clustering, and dividing the clustered test data into a clustered training data set and a clustered test data set;
step four, training a road feel model based on the K-Medoids and the BP neural network: when a road feel simulation model based on K-Medoids and BP neural networks is trained by using a clustered training data set and a BP neural network algorithm, input variables of the BP neural network model are vehicle speed, vehicle lateral acceleration, vehicle yaw velocity, vehicle vertical load, driver steering wheel turning angle and steering wheel angular velocity, output variables are steering wheel moments, and the same number of road feel simulation models based on the K-Medoids and the BP neural networks are obtained through training;
step five, testing a road feel model based on a K-Medoids and BP neural network; testing the obtained road feel simulation model based on the K-Medoids and BP neural network by using the clustered test data set;
step six, judging whether a supplementary test is needed;
and seventhly, carrying out road feel simulation according to the obtained road feel simulation model based on the K-Medoids and the BP neural network.
2. The road feel simulation method based on K-Medoids and BP neural networks according to claim 1, characterized in that in the real vehicle test of step one:
the test road types comprise urban roads, expressways, suburban roads and rural roads;
the vehicle running working conditions comprise straight running, backing, turning, pivot steering, uphill and downhill working conditions.
3. The road feel simulation method based on K-Medoids and BP neural networks as claimed in claim 1, wherein in step two, the removed abnormal points include data points beyond the normal value range, data points with severely deviated distribution and data points with the variation amplitude beyond the normal range.
4. The road feel simulation method based on K-Medoids and BP neural network according to claim 1, characterized in that in step two, the test data is normalized according to the following formula:
Figure FDA0002870276520000011
wherein i is a data number, j is a variable number, and xi,jDenotes the j variable, X, in the non-normalized i group of datajRepresents the set of variable data values corresponding to all j, min represents the minimum value, and max represents the maximum value.
5. The road feel simulation method based on K-Medoids and BP neural networks according to claim 1, characterized in that in step three, the normalized test data set is randomly divided into a clustered training data set and a clustered test data set according to a certain proportion of the number of data points.
6. The road feel simulation method based on K-Medoids and BP neural network according to any one of claim 1, characterized in that in step three, when clustering is performed on the normalized test data by using a K-Medoids clustering algorithm, the number of clusters is set to 5, 5 central coordinates are obtained after clustering, and the normalized test data is divided into 5 data classes according to the 5 central coordinates.
7. The road feel simulation method based on K-Medoids and BP neural network as claimed in any one of claims 1-6, wherein in step four, when training the road feel simulation model based on K-Medoids and BP neural network, the BP neural network model has 2 hidden layers, each hidden layer has 10 nodes, 1 input layer and 1 output layer; all the activation functions of all the nodes are Sigmoid functions and are all fully connected; the learning function uses the learngdm function.
8. The road feel simulation method based on K-Medoids and BP neural network as claimed in claim 7, wherein when training the road feel simulation model based on K-Medoids and BP neural network, the iteration upper limit is 1000 generations.
9. The road feel simulation method based on the K-Medoids and BP neural network as claimed in claim 1, wherein when testing the road feel simulation model based on the K-Medoids and BP neural network, the steps are as follows:
1) taking out the test data points in the test data set after clustering, and inputting the numerical values of the input variables corresponding to the test data points into the road feel simulation model corresponding to the class to which the test data points belong to obtain a predicted steering wheel torque value;
2) calculating the MSE value between the predicted steering wheel moment value and the real steering wheel moment value through a model at the test data points of the whole clustered test data set;
and when judging whether a supplementary test is needed, if the MSE value is smaller than a preset threshold value alpha, the road feel simulation model based on the data drive obtained by training is considered to be acceptable, and the supplementary test is not needed.
10. The K-Medoids and BP neural network-based road feel simulation method according to claim 9, wherein the threshold α is 0.15.
CN202011605208.8A 2020-12-29 2020-12-29 Road feel simulation method based on K-Medoids and BP neural network Withdrawn CN112597597A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011605208.8A CN112597597A (en) 2020-12-29 2020-12-29 Road feel simulation method based on K-Medoids and BP neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011605208.8A CN112597597A (en) 2020-12-29 2020-12-29 Road feel simulation method based on K-Medoids and BP neural network

Publications (1)

Publication Number Publication Date
CN112597597A true CN112597597A (en) 2021-04-02

Family

ID=75204067

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011605208.8A Withdrawn CN112597597A (en) 2020-12-29 2020-12-29 Road feel simulation method based on K-Medoids and BP neural network

Country Status (1)

Country Link
CN (1) CN112597597A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113096828A (en) * 2021-04-19 2021-07-09 梅里医疗科技(洋浦)有限责任公司 Diagnosis, prediction and major health management platform based on cancer genome big data core algorithm

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113096828A (en) * 2021-04-19 2021-07-09 梅里医疗科技(洋浦)有限责任公司 Diagnosis, prediction and major health management platform based on cancer genome big data core algorithm

Similar Documents

Publication Publication Date Title
CN107688343B (en) Energy control method of hybrid power vehicle
CN105741637B (en) Four-wheel hub motor electric car automated steering control method
CN112801143B (en) Steering road feel simulation method based on K-Means and Gaussian process regression
CN108482481B (en) Four-wheel steering control method for four-wheel independent drive and steering electric automobile
CN105034986B (en) A kind of driver&#39;s steering characteristic on-line identification method and device
CN112149796B (en) Driving style identification method based on improved genetic algorithm optimization BP neural network
CN113239986B (en) Training method and device for vehicle track evaluation network model and storage medium
CN112749508B (en) Road feel simulation method based on GMM and BP neural network
CN112632706B (en) Road feel simulation method based on GMM and CART regression tree
CN112046489A (en) Driving style identification algorithm based on factor analysis and machine learning
CN112632705B (en) Road feel simulation method based on GMM and Gaussian process regression
CN112550450B (en) Road feel simulation method based on K-Means and CART regression tree
CN111198501B (en) Method for determining fuel equivalent factor by RBF neural network
CN112597597A (en) Road feel simulation method based on K-Medoids and BP neural network
CN112528568B (en) Road feel simulation method based on K-Means and BP neural network
CN111907523A (en) Vehicle following optimization control method based on fuzzy reasoning
CN113642114B (en) Personified random following driving behavior modeling method capable of making mistakes
CN110816531B (en) Control system and control method for safe distance between unmanned automobile vehicles
CN109910865B (en) Vehicle early warning braking method based on Internet of things
CN112937587A (en) Road feel simulation method based on K-Medoids and classification regression tree
CN112685956B (en) Road feel simulation method based on CART regression tree
CN112572606B (en) Working condition fusion road sense simulation method based on SVM (support vector machine) and BP (back propagation) neural network
CN113705865B (en) Automobile stability factor prediction method based on deep neural network
CN112651134A (en) Steering road feel simulation method based on Gaussian process regression
CN112836722B (en) Road feel simulation method based on data driving

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20210402