CN111009134A - Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle - Google Patents

Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle Download PDF

Info

Publication number
CN111009134A
CN111009134A CN201911168690.0A CN201911168690A CN111009134A CN 111009134 A CN111009134 A CN 111009134A CN 201911168690 A CN201911168690 A CN 201911168690A CN 111009134 A CN111009134 A CN 111009134A
Authority
CN
China
Prior art keywords
vehicle
speed
neural network
historical
vehicle speed
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
Application number
CN201911168690.0A
Other languages
Chinese (zh)
Inventor
孙超
李军求
孙逢春
郭婷婷
励夏
孙海迪
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.)
Beijing Shouke Energy Technology Co ltd
Beijing Institute of Technology BIT
Original Assignee
Beijing Shouke Energy Technology Co ltd
Beijing Institute of Technology BIT
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 Beijing Shouke Energy Technology Co ltd, Beijing Institute of Technology BIT filed Critical Beijing Shouke Energy Technology Co ltd
Priority to CN201911168690.0A priority Critical patent/CN111009134A/en
Publication of CN111009134A publication Critical patent/CN111009134A/en
Priority to PCT/CN2020/105463 priority patent/WO2021103625A1/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • 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
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)

Abstract

The invention discloses a short-term vehicle speed working condition real-time prediction method based on interaction between a front vehicle and a self vehicle, which comprises the following steps of: s1, obtaining historical speed and distance information of a current vehicle and a previous vehicle, and extracting effective data; s2, constructing a future vehicle speed prediction model based on an artificial neural network; s3, performing off-line training on the constructed future vehicle speed prediction model; s4, predicting the future speed of the vehicle on line; s5, realizing self-adaptive learning of a vehicle speed prediction neural network; s6, calculating a torque demand according to the predicted short-term speed; and S7, calculating the optimal torque distribution according to the torque demand and a dynamic programming algorithm. The method adopts an artificial neural network method to predict the short-term speed of the vehicle, so that the accuracy of speed prediction is improved; and the predicted short-term speed of the automobile is applied to an energy management control strategy, so that the fuel economy is improved.

Description

Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle
Technical Field
The invention relates to vehicle speed working condition prediction of a vehicle, in particular to a short-term vehicle speed working condition real-time prediction method based on interaction of a front vehicle and a self vehicle.
Background
In recent years, the rapidly increasing automobile demand in China brings about a rapid increase in petroleum consumption, and meanwhile, the energy safety problem in China is more prominent. The increasingly deteriorating environment also promotes the urgent need for energy conservation and emission reduction in countries around the world. How to improve and perfect the optimization performance and the theoretical system of the energy management algorithm of the hybrid power system based on the existing research level and further realize the real-time online application of the power drive control algorithm based on the optimization is an important problem to be solved urgently in optimizing fuel consumption in the current automobile development.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides a short-term vehicle speed working condition real-time prediction method based on the interaction of a front vehicle and a self vehicle, greatly improves the accuracy of the future short-term vehicle speed prediction, applies the short-term vehicle speed prediction method to an energy management strategy and improves the fuel economy.
The purpose of the invention is realized by the following technical scheme: a short-term vehicle speed working condition real-time prediction method based on interaction of a front vehicle and a self vehicle is characterized by comprising the following steps: the method comprises the following steps:
s1, obtaining historical speed and distance information of a current vehicle and a previous vehicle, and extracting effective data;
s2, constructing a future vehicle speed prediction model based on an artificial neural network;
s3, performing off-line training on the constructed future vehicle speed prediction model;
s4, predicting the future speed of the vehicle on line;
s5, realizing self-adaptive learning of a vehicle speed prediction neural network;
s6, calculating a torque demand according to the predicted short-term speed;
and S7, calculating the optimal torque distribution according to the torque demand and a dynamic programming algorithm.
Wherein the step S1 includes the following substeps:
s101, obtaining historical speed information of different drivers of a vehicle when the vehicle runs on different road working conditions based on a vehicle-mounted OBD or wireless data acquisition system, and storing the historical speed information in a speed database to form a sample working condition 1;
s102, acquiring historical speed information of different drivers of a front vehicle corresponding to the self vehicle when the front vehicle runs on different road working conditions by using a vehicle networking technology, and storing the historical speed information in a speed database to form a sample working condition 2;
continuously acquiring front and rear vehicle distance information by using a millimeter wave radar, and storing the front and rear vehicle distance information in a historical vehicle distance database to form a sample working condition 3;
s103, extracting effective actual measurement data vehicle instantaneous speed v of the vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 10(t)；
S104, extracting effective actual measurement data vehicle instantaneous speed v of the front vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 21(t)；
And extracting effective actual measurement data of the front and rear vehicles at the time point t when the front and rear vehicles run in each sample working condition from the historical vehicle distance database sample working condition 3 to obtain the instantaneous vehicle distance S (t) of the front and rear vehicles.
The step S2 is to construct a future vehicle speed prediction model of the artificial neural network based on the historical vehicle speed or the historical vehicle distance information, and specifically comprises the following steps:
s201, selecting a radial basis function artificial neural network as a nonlinear prediction function to predict the future vehicle speed of the vehicle, and constructing a future vehicle speed prediction model based on the radial basis function artificial neural network;
s202, the radial basis function artificial neural network is composed of an input layer, a hidden layer and an output layer;
a1, when a model is built based on historical vehicle speed information, selecting the behavior Q of a driver and a time period H on the vehiclehHistorical speed in seconds and previous time period HsThe historical vehicle speed in seconds is used as the input of the neural network vehicle speed prediction model and is defined as:
Figure BDA0002288139170000021
when a model is constructed based on historical vehicle distance information, the behavior Q of a driver and the last time period H of the self vehicle are selectedhHistorical speed in seconds and previous and current time period HsThe historical distance in seconds is used as the input of a neural network vehicle speed prediction model, and is defined as:
Figure BDA0002288139170000022
wherein N isinFor the input of the prediction model, VkAnd VnThe current speed, V, of the vehicle from the current time to the previous timek-1And Vn-1The vehicle speeds of the own vehicle and the previous vehicle at the same time are respectively; snIs the distance between the preceding vehicle and the current vehicle, Sn-1The distance between the front vehicle and the self vehicle at the moment; time period HhI.e. the historical speed vector length H of the bicyclesThe length of the vector of the historical speed/distance of the vehicle in front, namely the length of the input vector of the neural network model, HhAnd HsAre all positive integers;
a2, selecting the neuron number O according to the training precision requirement of the neural network, wherein the activation function of the hidden layer is as follows:
a1＝exp(-‖n-c‖2/2b2),n＝Wa0+b；
wherein O is a positive integer, a1And a0The neuron outputs of the current layer and the previous layer respectively, n is the accumulated output, c is the neuron node center, b is the neuron radial basis function diffusion width, and W is the weight value；
A3, next future time period HpSpeed in seconds as output, HpThe length of the future vehicle speed vector is predicted, namely the length of the output vector of the neural network model; suppose fnIs a nonlinear function of the neural network prediction;
when a model is constructed based on historical vehicle speed information, the following steps are provided:
Figure BDA0002288139170000023
when a model is constructed based on historical vehicle distance information, the following steps are provided:
Figure BDA0002288139170000024
Figure BDA0002288139170000031
further, the step S3 includes the following sub-steps:
s301, constructing input vector parameters and output vector parameters according to the step S2, inputting the input parameter vectors and the output parameter vectors into a radial basis function artificial neural network model to form training samples for off-line training, and establishing a stable radial basis function artificial neural network structure;
in the model constructed based on the historical vehicle speed information, the input vector parameters are as follows:
Figure BDA0002288139170000032
the output vector parameters are:
Figure BDA0002288139170000033
in the model constructed based on the historical vehicle distance information, the input vector parameters are as follows:
Figure BDA0002288139170000034
the output vector parameters are:
Figure BDA0002288139170000035
s302, determining a connection mode of the radial basis function artificial neural network as g-h-m, namely g inputs, h hidden layers and m outputs;
s303, selecting a RBF neural network learning method of selecting a central line in a self-organizing way, wherein the core is to solve the center of a hidden layer basic function, the variance of the basic function and the weight from a hidden layer unit to an output unit, and thus the jth output in the RBF neural network is represented as:
Figure BDA0002288139170000036
in the formula (I), the compound is shown in the specification,
Figure BDA0002288139170000037
p is the pth input sample, P ═ 1,2, …, P; p is the total number of samples, ciFor the center of the hidden layer node of the network, i is 1,2, …, h is the number of the hidden layer nodes, | xp-ci‖2Is a European norm, sigmaiIs the width of the basis function, ωijThe connection weight from hidden layer to output layer, j is 1,2, …, m is the node number of output layer, yjIs the actual output of the jth output node of the neural network corresponding to the input sample;
the off-line training of the radial basis function artificial neural network comprises the following steps:
for the weight value omegaijAssigning a random number with an initial value of 0 to 1, setting the number of hidden layer neurons as h, setting an initial network error E as 0, and setting a maximum error epsilon as a positive decimal;
determining center c of basis function based on fuzzy K-means clustering algorithmiAnd variance σi，i＝1,2,…,h；
Method for adjusting weight omega from hidden layer to output layer of network by adopting gradient descent methodijUp to a network error E<E, ending; the network error is expressed by mean square error, and the expression is as follows:
Figure BDA0002288139170000041
in the formula, E represents a network error,
Figure BDA0002288139170000042
to correspond to the input xpThe actual output of (a), y (x)p) The table is the desired output corresponding to the input and P is the total number of samples.
Further, the step S4 includes the following steps:
s401, embedding the neural network model obtained in the step S3 into a vehicle control system;
s402, the next time period H for the self-vehiclepPredicting the future vehicle speed in seconds:
if the model obtained in step S3 is a model based on the historical vehicle speed, a time period H on the vehicle is usedhHistorical speed in seconds and previous time period HsHistorical speed data in seconds for next time period H of the vehiclepPredicting the future vehicle speed in seconds;
if the model obtained in step S3 is a model based on the historical inter-vehicle distance, a time period H on the host vehicle is usedhHistorical speed in seconds and previous and current time period HsHistorical vehicle distance data in seconds for next time period H of the self vehiclepPredicting the future vehicle speed in seconds;
s403, in the running process of the real vehicle, the mode of predicting the future vehicle speed comprises the following steps:
firstly, continuously acquiring a time period H on the vehicle based on a vehicle speed acquisition systemhReal-time speed data in seconds is continuously acquired in a time period H on the front vehicle based on the Internet of vehicles technologysReal-time vehicle speed data in seconds are fused with the style of a driver to form a neural network input vector, so that future vehicle speed prediction of the vehicle is realized;
secondly, continuously acquiring time on the vehicle based on a vehicle speed acquisition systemSegment HhReal-time speed data in seconds is obtained continuously by using millimeter wave radar to obtain a time period H on a front vehicle and a self vehiclesReal-time vehicle distance data in seconds are fused with the style of a driver to form a neural network input vector, and the future vehicle speed prediction of the vehicle is realized.
Further, the step S5 includes the following sub-steps:
s501, collecting speed data of a current vehicle and a previous vehicle or distance data of the previous vehicle and the current vehicle in the previous period, and updating a sample speed or distance database; each cycle is one week, one month or one year;
s502, relearning the neural network prediction model by using the updated sample vehicle speed data or vehicle distance data under the condition that the vehicle does not run;
s503, vehicle speed prediction is carried out by using the neural network model obtained in the step S502;
s504, after collecting new sample data, returning to 501, and repeating the whole process of the step S5.
Further, the torque request in step S6 is calculated as follows:
Figure BDA0002288139170000043
wherein V (T) is the predicted real-time speed of the vehicle, m is the mass of the vehicle, and TwheelAs wheel torque, RwheelIs wheel radius, theta is road slope, CrIs the road surface resistance coefficient, CdIs the wind resistance coefficient, and A is the frontal area of the vehicle.
Further, the step S7 includes the following sub-steps:
s701, dividing a stage and selecting a stage variable k;
s702, selecting a state variable lambdak；
S703, selecting decision variables and determining allowable decision sets of all levels;
s704, writing a state transition equation as follows:
λk+1＝T(λk,vk)；
wherein, T: (λk,vk) Denotes the state variable as λkThe vehicle speed is vkA state transition function of time;
s705, determining the form of a stage target function, wherein the target function must have separability and satisfy a recursion relation;
s706, writing a basic equation, namely a recursion equation and an endpoint condition which are met by an optimal value function, and taking an r function as a cost function:
r*(λk)＝min[r(λk,vk)+r*(λk+1)]
wherein, r (λ)k,vk) Denotes the state variable as λkThe vehicle speed is vkA corresponding cost function; r is*(λk) Denotes the state variable as λkMinimizing a function in reverse order of time; r is*(λk+1) Denotes the state variable as λk+1Minimizing a function in reverse order of time;
terminal conditions:
r*(λk+1)＝0
namely, find r*(λk+1) K value corresponding to 0;
s707, calculating the optimal value function in the state space and the corresponding optimal solution in a reverse order: that is, after k is obtained, the corresponding λ is calculated according to the formulas in steps S704 to S706k,vk；
S708, sequentially calculating an optimal control strategy in a given initial state according to the optimal value function and the optimal solution:
the optimal value function is r (lambda)k,vk)，r*(λk+1) The optimal solution is corresponding k, lambdak,vk；
And sequentially calculating the optimal control strategy in a given initial state, namely: the obtained optimal solution lambdak,vkAnd substituting the formula given in the step S6 to calculate the wheel torque, further calculating the optimal engine torque and the motor torque control variable, and obtaining the most effective control strategy in the given initial state.
The invention has the beneficial effects that: the invention provides a vehicle future short-term vehicle speed prediction method comprehensively considering vehicle state parameters, driver driving style and front vehicle state parameters through online learning of a neural network based on a front vehicle and self vehicle interaction mode and a vehicle networking technology and from the perspective of a human-vehicle-environment system, and improves the accuracy of vehicle speed prediction. And the proposed short-term vehicle speed prediction method is applied to a fuel consumption control strategy, so that better fuel economy is obtained.
Drawings
FIG. 1 is a schematic block diagram of a system of the present invention
Fig. 2 is a schematic diagram of a hardware device in the embodiment.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
As shown in FIG. 1, a method for predicting short-term vehicle speed working condition in real time based on interaction between a preceding vehicle and a current vehicle is characterized in that: the method comprises the following steps:
s1, obtaining historical speed and distance information of a current vehicle and a previous vehicle, and extracting effective data:
in embodiments of the present application, devices that need to be employed include, but are not limited to, on-board automatic Diagnostic Systems (OBDs), on-board units (OBUs) in dedicated short-range communication (DSRC) systems, drive-test units (RSUs), and millimeter-wave radars. The vehicle-mounted device OBD can acquire and store historical speed data of a vehicle, a road test unit (RSU) can acquire historical speed data of a front vehicle, a millimeter wave radar can acquire distance information between the front vehicle and a rear vehicle, and hardware equipment is shown in FIG. 2; step S1 specifically includes:
s101, obtaining historical speed information of different drivers of a vehicle when the vehicle runs on different road working conditions based on a vehicle-mounted OBD or wireless data acquisition system, and storing the historical speed information in a speed database to form a sample working condition 1;
s102, acquiring historical speed information of different drivers of a front vehicle corresponding to the self vehicle when the front vehicle runs on different road working conditions by using a vehicle networking technology (optional special short-range communication equipment for V2V), and storing the historical speed information in a speed database to form a sample working condition 2;
continuously acquiring front and rear vehicle distance information by using a millimeter wave radar, and storing the front and rear vehicle distance information in a historical vehicle distance database to form a sample working condition 3;
s103, extracting effective actual measurement data vehicle instantaneous speed v of the vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 10(t)；
S104, extracting effective actual measurement data vehicle instantaneous speed v of the front vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 21(t)；
And extracting effective actual measurement data of the front and rear vehicles at the time point t when the front and rear vehicles run in each sample working condition from the historical vehicle distance database sample working condition 3 to obtain the instantaneous vehicle distance S (t) of the front and rear vehicles.
S2, constructing a future vehicle speed prediction model based on the artificial neural network:
the artificial neural network can be selected from various artificial neural network algorithms such as a back propagation neural network, a layer recursive neural network, a radial basis function neural network and the like, and in the embodiment of the application, the radial basis function artificial neural network is selected for model construction.
The method comprises the following steps of constructing a future vehicle speed prediction model of an artificial neural network based on historical vehicle speed or historical vehicle distance information, and specifically comprises the following steps:
s201, selecting a radial basis function artificial neural network as a nonlinear prediction function to predict the future vehicle speed of the vehicle, and constructing a future vehicle speed prediction model based on the radial basis function artificial neural network;
s202, the radial basis function artificial neural network is composed of an input layer, a hidden layer and an output layer;
a1, when a model is built based on historical vehicle speed information, selecting the behavior Q of a driver and a time period H on the vehiclehHistorical speed in seconds and previous time period HsThe historical vehicle speed in seconds is used as the input of the neural network vehicle speed prediction model and is defined as:
Figure BDA0002288139170000071
when a model is constructed based on historical vehicle distance information, the behavior Q of a driver and the last time period H of the self vehicle are selectedhHistorical speed in seconds and previous and current time period HsThe historical distance in seconds is used as the input of a neural network vehicle speed prediction model, and is defined as:
Figure BDA0002288139170000072
wherein N isinFor the input of the prediction model, VkAnd VnThe current speed, V, of the vehicle from the current time to the previous timek-1And Vn-1The vehicle speeds of the own vehicle and the previous vehicle at the same time are respectively; snIs the distance between the preceding vehicle and the current vehicle, Sn-1The distance between the front vehicle and the self vehicle at the moment; time period HhI.e. the historical speed vector length H of the bicyclesThe length of the vector of the historical speed/distance of the vehicle in front, namely the length of the input vector of the neural network model, HhAnd HsAre all positive integers;
a2, selecting the neuron number O according to the training precision requirement of the neural network, wherein the activation function of the hidden layer is as follows:
a1＝exp(-‖n-c‖2/2b2),n＝Wa0+b；
wherein O is a positive integer, a1And a0The neuron outputs of the current layer and the previous layer are respectively, n is cumulative output, c is a neuron node center, b is a neuron radial basis function diffusion width, and W is a weight value;
a3, next future time period HpSpeed in seconds as output, HpThe length of the future vehicle speed vector is predicted, namely the length of the output vector of the neural network model; suppose fnIs a nonlinear function of the neural network prediction;
when a model is constructed based on historical vehicle speed information, the following steps are provided:
Figure BDA0002288139170000073
when a model is constructed based on historical vehicle distance information, the following steps are provided:
Figure BDA0002288139170000074
s3, performing off-line training on the built future vehicle speed prediction model:
s301, constructing input vector parameters and output vector parameters according to the step S2, inputting the input parameter vectors and the output parameter vectors into a radial basis function artificial neural network model to form training samples for off-line training, and establishing a stable radial basis function artificial neural network structure;
in the model constructed based on the historical vehicle speed information, the input vector parameters are as follows:
Figure BDA0002288139170000075
the output vector parameters are:
Figure BDA0002288139170000081
in the model constructed based on the historical vehicle distance information, the input vector parameters are as follows:
Figure BDA0002288139170000082
the output vector parameters are:
Figure BDA0002288139170000083
s302, determining a connection mode of the radial basis function artificial neural network as g-h-m, namely g inputs, h hidden layers and m outputs;
s303, selecting a RBF neural network learning method of selecting a central line in a self-organizing way, wherein the core is to solve the center of a hidden layer basic function, the variance of the basic function and the weight from a hidden layer unit to an output unit, and thus the jth output in the RBF neural network is represented as:
Figure BDA0002288139170000084
in the formula (I), the compound is shown in the specification,
Figure BDA0002288139170000085
p is the pth input sample, P ═ 1,2, …, P; p is the total number of samples, ciFor the center of the hidden layer node of the network, i is 1,2, …, h is the number of the hidden layer nodes, | xp-ci‖2Is a European norm, sigmaiIs the width of the basis function, ωijThe connection weight from hidden layer to output layer, j is 1,2, …, m is the node number of output layer, yjIs the actual output of the jth output node of the neural network corresponding to the input sample;
the off-line training of the radial basis function artificial neural network comprises the following steps:
for the weight value omegaijAssigning a random number with an initial value of 0 to 1, setting the number of hidden layer neurons as h, setting an initial network error E as 0, and setting a maximum error epsilon as a positive decimal;
determining center c of basis function based on fuzzy K-means clustering algorithmiAnd variance σi，i＝1,2,…,h；
Method for adjusting weight omega from hidden layer to output layer of network by adopting gradient descent methodijUp to a network error E<E, ending; the network error is expressed by mean square error, and the expression is as follows:
Figure BDA0002288139170000086
in the formula, E represents a network error,
Figure BDA0002288139170000087
to correspond to the input xpThe actual output of (a), y (x)p) The table is the desired output corresponding to the input and P is the total number of samples.
S4, predicting the future speed of the vehicle on line:
s401, embedding the neural network model obtained in the step S3 into a vehicle control system;
s402, the next time period H for the self-vehiclepPredicting the future vehicle speed in seconds:
if the model obtained in step S3 is a model based on the historical vehicle speed, a time period H on the vehicle is usedhHistorical speed in seconds and previous time period HsHistorical speed data in seconds for next time period H of the vehiclepPredicting the future vehicle speed in seconds;
if the model obtained in step S3 is a model based on the historical inter-vehicle distance, a time period H on the host vehicle is usedhHistorical speed in seconds and previous and current time period HsHistorical vehicle distance data in seconds for next time period H of the self vehiclepPredicting the future vehicle speed in seconds;
s403, in the running process of the real vehicle, the mode of predicting the future vehicle speed comprises the following steps:
firstly, continuously acquiring a time period H on the vehicle based on a vehicle speed acquisition systemhReal-time speed data in seconds is obtained continuously in a time period H on the front vehicle based on the vehicle networking technology (optional special short-range communication equipment for V2V)sReal-time vehicle speed data in seconds are fused with the style of a driver to form a neural network input vector, so that future vehicle speed prediction of the vehicle is realized;
secondly, continuously acquiring a time period H on the vehicle based on a vehicle speed acquisition systemhReal-time speed data in seconds is obtained continuously by using millimeter wave radar to obtain a time period H on a front vehicle and a self vehiclesReal-time vehicle distance data in seconds are fused with the style of a driver to form a neural network input vector, and the future vehicle speed prediction of the vehicle is realized.
S5, realizing self-adaptive learning of the vehicle speed prediction neural network:
s501, collecting speed data of a current vehicle and a previous vehicle or distance data of the previous vehicle and the current vehicle in the previous period, and updating a sample speed or distance database; each cycle is one week, one month or one year;
s502, relearning the neural network prediction model by using the updated sample vehicle speed data or vehicle distance data under the condition that the vehicle does not run;
s503, vehicle speed prediction is carried out by using the neural network model obtained in the step S502;
s504, after collecting new sample data, returning to 501, and repeating the whole process of the step S5.
S6, calculating a torque demand according to the predicted short-term speed:
the torque demand is calculated as follows:
Figure BDA0002288139170000091
wherein V (T) is the predicted real-time speed of the vehicle, m is the mass of the vehicle, and TwheelAs wheel torque, RwheelIs wheel radius, theta is road slope, CrIs the road surface resistance coefficient, CdIs the wind resistance coefficient, and A is the frontal area of the vehicle.
And S7, calculating the optimal torque distribution according to the torque demand and a dynamic programming algorithm.
S701, dividing a stage and selecting a stage variable k;
s702, selecting a state variable lambdak；
S703, selecting decision variables and determining allowable decision sets of all levels;
s704, writing a state transition equation as follows:
λk+1＝T(λk,vk)；
wherein, T (λ) herek,vk) Denotes the state variable as λkThe vehicle speed is vkA state transition function of time;
s705, determining the form of a stage target function, wherein the target function must have separability and satisfy a recursion relation;
s706, writing a basic equation, namely a recursion equation and an endpoint condition which are met by an optimal value function, and taking an r function as a cost function:
r*(λk)＝min[r(λk,vk)+r*(λk+1)]
wherein, r (λ)k,vk) Denotes the state variable as λkThe vehicle speed is vkA corresponding cost function; r is*(λk) Denotes the state variable as λkMinimizing a function in reverse order of time; r is*(λk+1) Denotes the state variable as λk+1Minimizing a function in reverse order of time;
terminal conditions:
r*(λk+1)＝0
namely, find r*(λk+1) K value corresponding to 0;
s707, calculating the optimal value function in the state space and the corresponding optimal solution in a reverse order: that is, after k is obtained, the corresponding λ is calculated according to the formulas in steps S704 to S706k,vk；
S708, sequentially calculating an optimal control strategy in a given initial state according to the optimal value function and the optimal solution:
the optimal value function is r (lambda)k,vk)，r*(λk+1) The optimal solution is corresponding k, lambdak,vk；
And sequentially calculating the optimal control strategy in a given initial state, namely: the obtained optimal solution lambdak,vkAnd substituting the formula given in the step S6 to calculate the wheel torque, further calculating the optimal engine torque and the motor torque control variable, and obtaining the most effective control strategy in the given initial state.
In summary, the invention provides a vehicle future short-term vehicle speed prediction method comprehensively considering vehicle state parameters, driver driving style and front vehicle state parameters through online learning of a neural network based on a front vehicle and self vehicle interaction mode and a vehicle networking technology and from the perspective of a human-vehicle-environment system, and improves the accuracy of vehicle speed prediction. And the proposed short-term vehicle speed prediction method is applied to a fuel consumption control strategy, so that better fuel economy is obtained.
The foregoing is a preferred embodiment of the present invention, it is to be understood that the invention is not limited to the form disclosed herein, but is not to be construed as excluding other embodiments, and is capable of other combinations, modifications, and environments and is capable of changes within the scope of the inventive concept as expressed herein, commensurate with the above teachings, or the skill or knowledge of the relevant art. And that modifications and variations may be effected by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (8)

1. A short-term vehicle speed working condition real-time prediction method based on interaction of a front vehicle and a self vehicle is characterized by comprising the following steps: the method comprises the following steps:
s1, obtaining historical speed and distance information of a current vehicle and a previous vehicle, and extracting effective data;
s2, constructing a future vehicle speed prediction model based on an artificial neural network;
s3, performing off-line training on the constructed future vehicle speed prediction model;
s4, predicting the future speed of the vehicle on line;
s5, realizing self-adaptive learning of a vehicle speed prediction neural network;
s6, calculating a torque demand according to the predicted short-term speed;
and S7, calculating the optimal torque distribution according to the torque demand and a dynamic programming algorithm.
2. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S1 includes the following sub-steps:
s101, obtaining historical speed information of different drivers of a vehicle when the vehicle runs on different road working conditions based on a vehicle-mounted OBD or wireless data acquisition system, and storing the historical speed information in a speed database to form a sample working condition 1;
s102, acquiring historical speed information of different drivers of a front vehicle corresponding to the self vehicle when the front vehicle runs on different road working conditions by using a vehicle networking technology, and storing the historical speed information in a speed database to form a sample working condition 2;
continuously acquiring front and rear vehicle distance information by using a millimeter wave radar, and storing the front and rear vehicle distance information in a historical vehicle distance database to form a sample working condition 3;
s103, extracting effective actual measurement data vehicle instantaneous speed v of the vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 10(t)；
S104, extracting effective actual measurement data vehicle instantaneous speed v of the front vehicle running at each sample working condition time point t from the historical vehicle speed database sample working condition 21(t)；
And extracting effective actual measurement data of the front and rear vehicles at the time point t when the front and rear vehicles run in each sample working condition from the historical vehicle distance database sample working condition 3 to obtain the instantaneous vehicle distance S (t) of the front and rear vehicles.
3. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S2 is to construct a future vehicle speed prediction model of the artificial neural network based on the historical vehicle speed or the historical vehicle distance information, and specifically comprises the following steps:
s201, selecting a radial basis function artificial neural network as a nonlinear prediction function to predict the future vehicle speed of the vehicle, and constructing a future vehicle speed prediction model based on the radial basis function artificial neural network;
s202, the radial basis function artificial neural network is composed of an input layer, a hidden layer and an output layer;
a1, when a model is built based on historical vehicle speed information, selecting the behavior Q of a driver and a time period H on the vehiclehHistorical speed in seconds and previous time period HsThe historical vehicle speed in seconds is used as the input of the neural network vehicle speed prediction model and is defined as:
Figure FDA0002288139160000011
when a model is constructed based on historical vehicle distance information, the behavior Q of a driver and the last time period H of the self vehicle are selectedhHistorical speed in seconds and previous and current time period HsThe historical distance in seconds is used as the input of a neural network vehicle speed prediction model, and is defined as:
Figure FDA0002288139160000021
wherein N isinFor the input of the prediction model, VkAnd VnThe current speed, V, of the vehicle from the current time to the previous timek-1And Vn-1The vehicle speeds of the own vehicle and the previous vehicle at the same time are respectively; snIs the distance between the preceding vehicle and the current vehicle, Sn-1The distance between the front vehicle and the self vehicle at the moment; time period HhI.e. the historical speed vector length H of the bicyclesThe length of the vector of the historical speed/distance of the vehicle in front, namely the length of the input vector of the neural network model, HhAnd HsAre all positive integers;
a2, selecting the neuron number O according to the training precision requirement of the neural network, wherein the activation function of the hidden layer is as follows:
a1＝exp(-‖n-c‖2/2b2),n＝Wa0+b；
wherein O is a positive integer, a1And a0The neuron outputs of the current layer and the previous layer are respectively, n is cumulative output, c is a neuron node center, b is a neuron radial basis function diffusion width, and W is a weight value;
a3, next future time period HpSpeed in seconds as output, HpThe length of the future vehicle speed vector is predicted, namely the length of the output vector of the neural network model; suppose fnIs a nonlinear function of the neural network prediction;
when a model is constructed based on historical vehicle speed information, the following steps are provided:
Figure FDA0002288139160000022
when a model is constructed based on historical vehicle distance information, the following steps are provided:
Figure FDA0002288139160000023
4. the method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S3 includes the following sub-steps:
s301, constructing input vector parameters and output vector parameters according to the step S2, inputting the input parameter vectors and the output parameter vectors into a radial basis function artificial neural network model to form training samples for off-line training, and establishing a stable radial basis function artificial neural network structure;
s302, determining a connection mode of the radial basis function artificial neural network as g-h-m, namely g inputs, h hidden layers and m outputs;
s303, selecting a RBF neural network learning method of selecting a central line in a self-organizing way, wherein the core is to solve the center of a hidden layer basic function, the variance of the basic function and the weight from a hidden layer unit to an output unit, and thus the jth output in the RBF neural network is represented as:
Figure FDA0002288139160000024
in the formula (I), the compound is shown in the specification,
Figure FDA0002288139160000031
p is the pth input sample, P ═ 1,2, …, P; p is the total number of samples, ciFor the center of the hidden layer node of the network, i is 1,2, …, h is the number of the hidden layer nodes, | xp-ci‖2Is a European norm, sigmaiIs the width of the basis function, ωijThe connection weight from hidden layer to output layer, j is 1,2, …, m is the node number of output layer, yjIs the actual output of the jth output node of the neural network corresponding to the input sample;
the off-line training of the radial basis function artificial neural network comprises the following steps:
for the weight value omegaijAssigning a random number with an initial value of 0 to 1, setting the number of hidden layer neurons as h, setting an initial network error E as 0, and setting a maximum error epsilon as a positive decimal;
determining center c of basis function based on fuzzy K-means clustering algorithmiAnd variance σi，i＝1,2,…,h；
Method for adjusting weight omega from hidden layer to output layer of network by adopting gradient descent methodijUp to a network error E<E, ending; the network error is expressed by mean square error, and the expression is as follows:
Figure FDA0002288139160000032
in the formula, E represents a network error,
Figure FDA0002288139160000033
to correspond to the input xpThe actual output of (a), y (x)p) The table is the desired output corresponding to the input and P is the total number of samples.
5. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S4 includes the steps of:
s401, embedding the neural network model obtained in the step S3 into a vehicle control system;
s402, the next time period H for the self-vehiclepPredicting the future vehicle speed in seconds:
if the model obtained in step S3 is a model based on the historical vehicle speed, a time period H on the vehicle is usedhHistorical speed in seconds and previous time period HsHistorical speed data in seconds for next time period H of the vehiclepPredicting the future vehicle speed in seconds;
if the model obtained in step S3 is a model based on the historical inter-vehicle distance, a time period H on the host vehicle is usedhHistorical speed in seconds and previous and current time period HsHistorical vehicle distance data in seconds for next time period H of the self vehiclepPredicting the future vehicle speed in seconds;
s403, in the running process of the real vehicle, the mode of predicting the future vehicle speed comprises the following steps:
firstly, continuously acquiring a time period H on the vehicle based on a vehicle speed acquisition systemhSecond of fullThe time-speed data is continuously acquired in a previous time period H based on the Internet of vehicles technologysReal-time vehicle speed data in seconds are fused with the style of a driver to form a neural network input vector, so that future vehicle speed prediction of the vehicle is realized;
secondly, continuously acquiring a time period H on the vehicle based on a vehicle speed acquisition systemhReal-time speed data in seconds is obtained continuously by using millimeter wave radar to obtain a time period H on a front vehicle and a self vehiclesReal-time vehicle distance data in seconds are fused with the style of a driver to form a neural network input vector, and the future vehicle speed prediction of the vehicle is realized.
6. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S5 includes the following sub-steps:
s501, collecting speed data of a current vehicle and a previous vehicle or distance data of the previous vehicle and the current vehicle in the previous period, and updating a sample speed or distance database; each cycle is one week, one month or one year;
s502, relearning the neural network prediction model by using the updated sample vehicle speed data or vehicle distance data under the condition that the vehicle does not run;
s503, vehicle speed prediction is carried out by using the neural network model obtained in the step S502;
s504, after collecting new sample data, returning to 501, and repeating the whole process of the step S5.
7. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the torque demand in step S6 is calculated as follows:
Figure FDA0002288139160000041
wherein V (T) is the predicted real-time speed of the vehicle, m is the mass of the vehicle, and TwheelAs wheel torque, RwheelIs wheel radius, theta is road slope, CrFor the road surface resistance systemNumber, CdIs the wind resistance coefficient, and A is the frontal area of the vehicle.
8. The method for predicting the short-term vehicle speed working condition in real time based on the interaction between the front vehicle and the self vehicle as claimed in claim 1, wherein the method comprises the following steps: the step S7 includes the following sub-steps:
s701, dividing a stage and selecting a stage variable k;
s702, selecting a state variable lambdak；
S703, selecting decision variables and determining allowable decision sets of all levels;
s704, writing a state transition equation as follows:
λk+1＝T(λk,vk)；
wherein, T (λ) herek,vk) Denotes the state variable as λkThe vehicle speed is vkA state transition function of time;
s705, determining the form of a stage target function, wherein the target function must have separability and satisfy a recursion relation;
s706, writing a basic equation, namely a recursion equation and an endpoint condition which are met by an optimal value function, and taking an r function as a cost function:
r*(λk)＝min[r(λk,vk)+r*(λk+1)]
wherein, r (λ)k,vk) Denotes the state variable as λkThe vehicle speed is vkA corresponding cost function; r is*(λk) Denotes the state variable as λkMinimizing a function in reverse order of time; r is*(λk+1) Denotes the state variable as λk+1Minimizing a function in reverse order of time;
terminal conditions:
r*(λk+1)＝0
namely, find r*(λk+1) K value corresponding to 0;
s707, calculating the optimal value function in the state space and the corresponding optimal solution in a reverse order: that is, after k is obtained, the corresponding λ is calculated according to the formulas in steps S704 to S706k,vk；
S708, sequentially calculating an optimal control strategy in a given initial state according to the optimal value function and the optimal solution:
the optimal value function is r (lambda)k,vk)，r*(λk+1) The optimal solution is corresponding k, lambdak,vk；
And sequentially calculating the optimal control strategy in a given initial state, namely: the obtained optimal solution lambdak,vkAnd substituting the formula given in the step S6 to calculate the wheel torque, further calculating the optimal engine torque and the motor torque control variable, and obtaining the most effective control strategy in the given initial state.
CN201911168690.0A 2019-11-25 2019-11-25 Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle Pending CN111009134A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201911168690.0A CN111009134A (en) 2019-11-25 2019-11-25 Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle
PCT/CN2020/105463 WO2021103625A1 (en) 2019-11-25 2020-07-29 Short-term vehicle speed condition real-time prediction method based on interaction between vehicle ahead and current vehicle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911168690.0A CN111009134A (en) 2019-11-25 2019-11-25 Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle

Publications (1)

Publication Number Publication Date
CN111009134A true CN111009134A (en) 2020-04-14

Family

ID=70113808

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911168690.0A Pending CN111009134A (en) 2019-11-25 2019-11-25 Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle

Country Status (2)

Country Link
CN (1) CN111009134A (en)
WO (1) WO2021103625A1 (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111791887A (en) * 2020-07-03 2020-10-20 北京理工大学 Vehicle energy-saving driving method based on layered vehicle speed planning
CN112785845A (en) * 2020-12-30 2021-05-11 桂林电子科技大学 Vehicle speed prediction method based on K-means clustering and RBF neural network
WO2021103625A1 (en) * 2019-11-25 2021-06-03 北京理工大学 Short-term vehicle speed condition real-time prediction method based on interaction between vehicle ahead and current vehicle
CN112937547A (en) * 2021-01-28 2021-06-11 北京理工大学 Plug-in hybrid power bus energy management method based on global working conditions
CN113212414A (en) * 2021-06-16 2021-08-06 北京理工大学 Power reserve prediction control method of series-type electromechanical compound transmission system
CN113267345A (en) * 2021-04-23 2021-08-17 联合汽车电子有限公司 Method for predicting resistance of unknown road section in front of vehicle, storage medium, controller and system
CN116331184A (en) * 2023-03-31 2023-06-27 长安大学 Vehicle energy control method and system based on machine vision and radar information fusion
CN117962633A (en) * 2024-01-10 2024-05-03 南京航空航天大学 Electric automobile moment distribution energy-saving control method based on deep reinforcement learning

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113722837B (en) * 2021-09-16 2023-09-29 吉林大学 Method for establishing diesel vehicle fuel consumption and emission model according to data clusters of different working conditions
CN113989534B (en) * 2021-10-15 2024-07-02 厦门大学 Bus inter-station load identification method and system based on multi-source operation data
CN114154227A (en) * 2021-10-18 2022-03-08 瑞立集团瑞安汽车零部件有限公司 Self-adaptive learning method for braking process of heavy trailer
CN114004993A (en) * 2021-10-25 2022-02-01 厦门大学 IA-SVM running condition identification method and device based on LSTM speed prediction optimization
CN113987679A (en) * 2021-10-30 2022-01-28 福州大学 Train stability prediction method and system based on MOGRIFIER-GRU deep learning
CN114065402B (en) * 2021-11-25 2024-04-19 华中科技大学 Vehicle full-state sound quality forecasting method based on HSSA-BPNN model
CN114460936B (en) * 2022-01-13 2024-04-02 华中科技大学 Automatic driving automobile path planning method and system based on offline incremental learning
CN114228747B (en) * 2022-01-25 2023-04-25 同济大学 Expressway lane change decision method based on Bayesian network
CN115880905B (en) * 2022-11-29 2024-05-17 吉林大学 Device and method for predicting intersection time of vehicle passing traffic signal lamp in real time
CN116522779B (en) * 2023-04-29 2023-10-20 中国人民解放军63963部队 Armored vehicle dynamic performance and engine operation MAP graph matching system and method
CN116702096B (en) * 2023-08-04 2023-10-03 中汽研汽车检验中心(昆明)有限公司 Method and device for measuring and calculating road sliding resistance of vehicle plateau environment
CN116821775A (en) * 2023-08-29 2023-09-29 陕西重型汽车有限公司 Load estimation method based on machine learning
CN117111620B (en) * 2023-10-23 2024-03-29 山东省科学院海洋仪器仪表研究所 Autonomous decision-making method for task allocation of heterogeneous unmanned system
CN117131955B (en) * 2023-10-27 2024-01-16 北京航空航天大学 Short-time vehicle speed prediction method considering multiple constraint conditions
CN117746066B (en) * 2024-02-20 2024-05-07 贵州博睿科讯科技发展有限公司 Diffusion model guided high-speed vehicle detection integrated learning method and device

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4200863B2 (en) * 2003-09-16 2008-12-24 株式会社エクォス・リサーチ Traveling speed pattern estimation device and hybrid vehicle drive control device
JP4786553B2 (en) * 2007-01-04 2011-10-05 トヨタ自動車株式会社 Rotational speed prediction apparatus, prediction method, program for realizing the method, and recording medium recording the program
CN103914985A (en) * 2014-04-25 2014-07-09 大连理工大学 Method for predicting future speed trajectory of hybrid power bus
CN105247192A (en) * 2013-05-31 2016-01-13 日立汽车系统株式会社 Vehicle control device and vehicle control method
CN108016427A (en) * 2016-11-02 2018-05-11 现代自动车株式会社 The method of motor vehicle driven by mixed power and control speed changer
CN109131350A (en) * 2018-08-23 2019-01-04 北京理工大学 A kind of hybrid vehicle energy management method and system
CN109649389A (en) * 2018-11-05 2019-04-19 江苏大学 A kind of control system and control method of the adjustable adaptive cruise of acceleration

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111009134A (en) * 2019-11-25 2020-04-14 北京理工大学 Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4200863B2 (en) * 2003-09-16 2008-12-24 株式会社エクォス・リサーチ Traveling speed pattern estimation device and hybrid vehicle drive control device
JP4786553B2 (en) * 2007-01-04 2011-10-05 トヨタ自動車株式会社 Rotational speed prediction apparatus, prediction method, program for realizing the method, and recording medium recording the program
CN105247192A (en) * 2013-05-31 2016-01-13 日立汽车系统株式会社 Vehicle control device and vehicle control method
CN103914985A (en) * 2014-04-25 2014-07-09 大连理工大学 Method for predicting future speed trajectory of hybrid power bus
CN108016427A (en) * 2016-11-02 2018-05-11 现代自动车株式会社 The method of motor vehicle driven by mixed power and control speed changer
CN109131350A (en) * 2018-08-23 2019-01-04 北京理工大学 A kind of hybrid vehicle energy management method and system
CN109649389A (en) * 2018-11-05 2019-04-19 江苏大学 A kind of control system and control method of the adjustable adaptive cruise of acceleration

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
孙超 等: "混合动力汽车预测能量管理研究", 《中国博士学位论文全文数据库·工程科技Ⅱ辑》 *
王旭 等: "考虑交通信息的PHEV电量消耗轨迹预测与跟随算法研究", 《重庆理工大学学报(自然科学)》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021103625A1 (en) * 2019-11-25 2021-06-03 北京理工大学 Short-term vehicle speed condition real-time prediction method based on interaction between vehicle ahead and current vehicle
CN111791887A (en) * 2020-07-03 2020-10-20 北京理工大学 Vehicle energy-saving driving method based on layered vehicle speed planning
CN112785845A (en) * 2020-12-30 2021-05-11 桂林电子科技大学 Vehicle speed prediction method based on K-means clustering and RBF neural network
CN112937547A (en) * 2021-01-28 2021-06-11 北京理工大学 Plug-in hybrid power bus energy management method based on global working conditions
CN113267345A (en) * 2021-04-23 2021-08-17 联合汽车电子有限公司 Method for predicting resistance of unknown road section in front of vehicle, storage medium, controller and system
CN113212414A (en) * 2021-06-16 2021-08-06 北京理工大学 Power reserve prediction control method of series-type electromechanical compound transmission system
CN116331184A (en) * 2023-03-31 2023-06-27 长安大学 Vehicle energy control method and system based on machine vision and radar information fusion
CN116331184B (en) * 2023-03-31 2024-04-02 长安大学 Vehicle energy control method and system based on machine vision and radar information fusion
CN117962633A (en) * 2024-01-10 2024-05-03 南京航空航天大学 Electric automobile moment distribution energy-saving control method based on deep reinforcement learning
CN117962633B (en) * 2024-01-10 2024-07-02 南京航空航天大学 Electric automobile moment distribution energy-saving control method based on deep reinforcement learning

Also Published As

Publication number Publication date
WO2021103625A1 (en) 2021-06-03

Similar Documents

Publication Publication Date Title
CN111009134A (en) Short-term vehicle speed working condition real-time prediction method based on interaction between front vehicle and self vehicle
CN111267831B (en) Intelligent time-domain-variable model prediction energy management method for hybrid electric vehicle
CN110991757B (en) Comprehensive prediction energy management method for hybrid electric vehicle
Huang et al. Model predictive control power management strategies for HEVs: A review
CN110341690B (en) PHEV energy management method based on deterministic strategy gradient learning
Yuan et al. Intelligent energy management strategy based on hierarchical approximate global optimization for plug-in fuel cell hybrid electric vehicles
CN107688343B (en) Energy control method of hybrid power vehicle
Lin et al. An ensemble learning velocity prediction-based energy management strategy for a plug-in hybrid electric vehicle considering driving pattern adaptive reference SOC
Sun et al. Velocity predictors for predictive energy management in hybrid electric vehicles
Yufang et al. Investigating long‐term vehicle speed prediction based on BP‐LSTM algorithms
US8965597B2 (en) Road grade auto-mapping
CN109131350B (en) Energy management method and system for hybrid electric vehicle
CN112668799A (en) Intelligent energy management method and storage medium for PHEV (Power electric vehicle) based on big driving data
CN113525396B (en) Hybrid electric vehicle layered prediction energy management method integrating deep reinforcement learning
CN112286212B (en) Vehicle network cooperative energy-saving control method
CN113135113B (en) Global SOC (System on chip) planning method and device
CN111547039B (en) Hybrid vehicle accelerator control method and system based on deep reinforcement learning
CN113095558A (en) Iterative optimization multi-scale fusion vehicle speed prediction algorithm for intelligent networked automobile
EP3964383A1 (en) Vehicle control unit (vcu) and operating method thereof
CN115257695A (en) Energy management control system and method for hybrid electric vehicle
Han et al. Ensemblefollower: A hybrid car-following framework based on reinforcement learning and hierarchical planning
CN113246958B (en) TD 3-based multi-target HEV energy management method and system
Gan et al. Intelligent learning algorithm and intelligent transportation-based energy management strategies for hybrid electric vehicles: A review
Xia et al. A predictive energy management strategy for multi-mode plug-in hybrid electric vehicle based on long short-term memory neural network
Chen et al. Traffic signal optimization control method based on adaptive weighted averaged double deep Q network

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: 20200414