CN115018353A - Intelligent network-connected automobile decision planning method under heterogeneous traffic flow - Google Patents

Intelligent network-connected automobile decision planning method under heterogeneous traffic flow Download PDF

Info

Publication number
CN115018353A
CN115018353A CN202210712562.3A CN202210712562A CN115018353A CN 115018353 A CN115018353 A CN 115018353A CN 202210712562 A CN202210712562 A CN 202210712562A CN 115018353 A CN115018353 A CN 115018353A
Authority
CN
China
Prior art keywords
vehicle
lane
vehicles
intelligent
information
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
CN202210712562.3A
Other languages
Chinese (zh)
Inventor
盛小飞
李金�
蒋亮
李书建
许玮
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.)
Wuhu Simba Network Technology Co ltd
Original Assignee
Wuhu Simba Network 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 Wuhu Simba Network Technology Co ltd filed Critical Wuhu Simba Network Technology Co ltd
Priority to CN202210712562.3A priority Critical patent/CN115018353A/en
Publication of CN115018353A publication Critical patent/CN115018353A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Primary Health Care (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses an intelligent network connection automobile decision planning method under heterogeneous traffic flow, which comprises the following steps: recognizing the intelligent networking state of other vehicles on the local road section, and calculating the traffic capacity of each lane; acquiring the motion state of other vehicles on the road within a period of time in the future according to the intelligent networking state identification result of other vehicles; establishing an omnidirectional collision risk assessment model, and calculating the driving risk of the intelligent networked vehicle in real time; according to the traffic capacity, the driving risk and the driving risk change rate of a lane where the intelligent networked vehicle is located and an adjacent lane, whether the vehicle needs to change the lane or not is determined, and then a lane change track is determined according to the traffic efficiency of the adjacent lane and the driving risk change condition from the lane change to the adjacent lane. According to the method, the road side base station can be used for accurately acquiring the future motion state of a part of vehicles with networking capability, a prediction method is not needed for indirect acquisition, the calculation efficiency is improved, and the precision of a decision planning stage is also improved.

Description

Intelligent network-connected automobile decision planning method under heterogeneous traffic flow
Technical Field
The invention belongs to the technical field of intelligent traffic systems, and particularly relates to an intelligent network-connected automobile decision planning method under heterogeneous traffic flow.
Background
With the development of the automobile industry, the mileage of highways and the automobile keeping amount in China also increase year by year. While the traffic industry develops rapidly, new problems are brought to the traffic industry: environmental pollution and traffic accidents, which seriously affect the production and life of people. For the current existing traffic problem, the american intelligent transportation association has first proposed the concept of an intelligent traffic system, which considers that the driver is the ring that is the most unstable and the most random in the whole driving process, and particularly points out the important role of the intelligent vehicle in the traffic problem. The Society of Automotive Engineers (SAE) divides automated driving into 6 grades, among them, L0 (manual driving), L1 (assisted driving), L2 (partial automated driving), L3 (conditional automated driving), L4 (advanced automated driving), L5 (full automated driving). The intelligent vehicles of the L4 and L5 grades can highly or completely realize automatic driving, and researches find that the intelligent vehicles are superior to common drivers in emergency condition handling performance. On the basis, the intelligent networking automobile is provided with advanced vehicle-mounted sensing equipment, a controller and an actuator, and is combined with modern communication technology, so that information interaction (V2X) between the automobile and platforms such as the automobile, the road and the cloud can be realized, the intelligent networking automobile has the functions of environmental perception, decision planning, cooperative control and the like, and has become a future development target of the traditional automobile industry as a strategic high point of automobile technology, and under the background of a new round of scientific and technological revolution represented by 5G, big data and cloud computing, China successively releases multiple development plans to guide the development of the intelligent networking automobile. Therefore, the development of the intelligent internet automobile completely accords with the industrial reform trend of intelligent manufacturing in China, and has very strong practical significance.
In recent years, the permeability of vehicles with automatic driving functions in the domestic market is rapidly improved, automatic driving vehicles with different levels of automatic driving functions at the present stage will gradually infiltrate into the traditional road traffic environment in a certain proportion, the traffic flow is changed from pure manual driving to pure automatic driving, a long infiltration process is needed, and a heterogeneous traffic flow formed by mixing automatic driving vehicles and manual driving vehicles inevitably exists. Meanwhile, with the development of the intelligent networking technology, vehicles with different intelligent networking degrees will appear in road traffic successively, and a phenomenon that highly complex heterogeneous traffic flows with different automatic driving grades and different intelligent networking degrees are mixed on roads will exist for a long time in the future.
However, the decision planning layer is the most core link of the intelligent networked automobile, most of the decision planning methods related to the intelligent networked automobile at present are concentrated in homogeneous traffic flows with the same intelligent and networked degrees, most researchers can default that the environment where the automobile is located is the homogeneous traffic flow when designing a decision planning algorithm, and the situation that heterogeneous traffic flows such as the intelligent automobile, the networked automobile and a common automobile are mixed in the future is not considered.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide an intelligent network-connected automobile decision planning method under heterogeneous traffic flow, so as to solve the problems of low efficiency, low precision and poor reliability in a decision planning stage caused by the fact that the existing decision planning method does not fully consider heterogeneous traffic flow mixed operation working conditions. The method fully utilizes the vehicle-road-cloud network connection facility, identifies the intelligent network connection state and the traffic capacity of the vehicles running on the local road section through the road side base station, adopts a diversified method for heterogeneous vehicles to obtain the future motion state information of the heterogeneous vehicles, combines an omnidirectional collision risk assessment model to monitor the running risk of the intelligent network connection vehicle in real time, finally carries out decision planning according to the real-time risk and the road traffic capacity, fully utilizes the intelligent network connection technology, and can greatly improve the calculation efficiency, the precision and the reliability in the decision planning stage.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the invention discloses an intelligent network connection automobile decision planning method under heterogeneous traffic flow, which comprises the following steps:
(1) the intelligent internet vehicle receives the check information stream sent by the road side base station, identifies the intelligent internet state of other vehicles on the local road section, and calculates the traffic capacity of each lane;
(2) acquiring the motion state of other vehicles on the road within a period of time in the future according to the intelligent networking state identification result of other vehicles in the step (1);
(3) establishing an omnidirectional collision risk assessment model, inputting the motion state information acquired in the step (2) into the omnidirectional collision risk assessment model, and calculating the driving risk of the intelligent networked vehicle in real time;
(4) according to the traffic capacity, the driving risk and the driving risk change rate of a lane where the intelligent networked vehicle is located and an adjacent lane, whether the vehicle needs to change the lane or not is determined, and then a lane change track is determined according to the traffic efficiency of the adjacent lane and the driving risk change condition from the lane change to the adjacent lane.
Further, the step of identifying the intelligent networking state of other vehicles on the local road section in the step (1) comprises the following steps: the road side base station sends a check information stream to a vehicle running on the road section, the vehicle returns a corresponding check information stream after receiving the check information stream to reflect the receiving state of the vehicle, if the road side base station receives the check information stream returned by the vehicle, the road side base station divides the vehicle into intelligent network connection vehicles, and the vehicle is judged to be in an automatic or manual driving state according to the returned information; and if the road side base station does not receive the check information flow returned by the vehicle, dividing the vehicle into non-intelligent network connection vehicles.
Further, the calculation method of the traffic capacity of each lane in the step (1) is as follows:
Q i ＝VR i ×D i (1)
in the formula, Q i The traffic average flow of the ith lane represents the traffic capacity of the lane; VR (virtual reality) i The average vehicle speed of the i-th lane is taken as the average vehicle speed of the i-th lane; d i Is the average traffic density of the ith lane.
Further, the step (2) of obtaining the motion state of other vehicles on the road in a future period of time by adopting a diversified method comprises the following steps:
(21) if the other vehicles are intelligent network connection vehicles and are in an automatic driving state according to the identification result in the step (1), directly utilizing network connection communication to obtain the future motion states of the other vehicles and calibrating the obtained information by combining the roadside base station information;
(22) if the other vehicles are intelligent network communication vehicles and are in a manual driving state in the identification result in the step (1), acquiring the state and vehicle state information of a driver of the vehicle by utilizing network communication, identifying the driving intention by combining the state of the driver and the state of the vehicle, and predicting the future motion state of the vehicle;
(23) and (3) if the other vehicle is a non-intelligent networked vehicle as a result of the identification in the step (1), acquiring the state information of the vehicle by using a vehicle-mounted sensing sensor, and performing driving intention identification and future motion state prediction of the vehicle by using the acquired vehicle state information.
Further, the driving intention identifying step in the step (22) is as follows:
(221) the method for collecting the driver state information and the vehicle state information offline comprises the following steps: establishing a driver intention recognition data set by information of a driver sight line focus, a heart rate, a breathing rate, a head rotation angle, a vehicle speed, a vehicle acceleration, a yaw rate, a steering wheel angle, a steering wheel angular speed, a vehicle offset lane center line position and a vehicle lateral position;
(222) calibrating the lane change intention of the driver corresponding to each group of data in the data set established in the step (221), updating the weights of all parameters contained in the data set by adopting a Relieff algorithm, and sequencing the weights of all parameters to screen out characteristic parameters which can best reflect the lane change intention of the driver, wherein the calculation method of the weight W (A) of any parameter A comprises the following steps:
Figure BDA0003707394170000031
Figure BDA0003707394170000032
in the formula, diff (A, R, H) j ) Denotes sample R and sample H j The difference in parameter A, p (C) is the proportion of C ≠ class (R), p (classes (R)) is the proportion of samples of the same class as samples R, M j (C) Indicating that the jth nearest neighbor sample in the class C ≠ class (R); k is the number of samples selected and classified as the same as the parameter A; m is a sample class; max and min are functions for solving the maximum value and the minimum value respectively;
(223) and (3) taking the characteristic parameters screened in the step (222) as an input layer of the LSTM neural network, taking the lane changing intention of the driver corresponding to each group of parameters as an output layer of the LSTM neural network, and training the LSTM neural network to identify the driving intention of the driver, wherein the specific steps are as follows:
(2231) calculating a forgetting gate:
f t ＝σ(W f ·[h t-1 ,X t ])+b f ) (4)
in the formula (f) t The value range is 0 to 1 for the forgetting gate at the current moment; w f A forgetting gate weight value; x t Is the input value at the current moment; h is t-1 Is the output value of the last moment; b f Biasing for a forget gate; sigma is sigmoid function;
(2232) calculation input gate:
i t ＝σ(W i ·[h t-1 ,X t ])+b i ) (5)
in the formula i t The input gate at the current moment has a value range of 0 to 1; w i The weight value of the input gate; b i Biasing the input gate;
(2233) calculating candidate memory cell information:
Figure BDA0003707394170000047
in the formula (I), the compound is shown in the specification,
Figure BDA0003707394170000048
candidate information to be updated to the memory unit at the current time; w C The candidate information weight value is obtained; b is a mixture of C Biasing the candidate information; tan h is a hyperbolic tangent function;
(2234) calculating new memory cell information:
Figure BDA0003707394170000041
in the formula, C t New memory cell information at the current moment; c t-1 Memory cell information of the previous moment;
(2235) calculating the LSTM neural network output:
o t ＝σ(W o ·[h t-1 ,X t ])+b o ) (8)
h t ＝o t ·tanh(C t ) (9)
in the formula o t Is the initial output of the current moment; w o Is the initial output weight value; b o Is an initial bias; h is t The output of the current moment is the driving intention of the driver.
Further, the step (22) of predicting the future motion state of the vehicle comprises the following steps:
(224) establishing a short time domain low-speed kinematic prediction model, predicting the future motion state of the vehicle when the vehicle runs at low speed, and recording the future motion state as
Figure BDA0003707394170000042
Figure BDA0003707394170000043
Wherein X is the longitudinal position of the vehicle; y is the lateral position of the vehicle; v is the vehicle speed;
Figure BDA0003707394170000044
is the vehicle yaw angle; beta is the centroid slip angle; l f Is the distance from the vehicle center of mass to the front axle; l r Is the distance from the center of mass of the vehicle to the rear axle; a is vehicle acceleration; delta f Is the front wheel corner of the vehicle; sin, cos and tan are sine, cosine and tangent functions, respectively, X t Longitudinal position of vehicle at time t, X t+1 The longitudinal position of the vehicle at time t + 1; y is t For the transverse position of the vehicle at time t, Y t+1 The lateral position of the vehicle at time t + 1;
Figure BDA0003707394170000045
for the yaw angle of the vehicle at time t,
Figure BDA0003707394170000046
the vehicle yaw angle at time t + 1; v. of t Vehicle speed at time t, v t+1 The vehicle speed at the moment t + 1;
(225) establishing a short time domain high speed dynamics prediction model, predicting the future motion state of the vehicle when the vehicle runs at high speed, and recording the future motion state as
Figure BDA0003707394170000051
Figure BDA0003707394170000052
In the formula, m is the total vehicle mass of the vehicle;
Figure BDA0003707394170000053
and
Figure BDA0003707394170000054
respectively the longitudinal speed and the acceleration of the vehicle under a vehicle coordinate system;
Figure BDA0003707394170000055
and
Figure BDA0003707394170000056
respectively vehicle coordinatesTie-down vehicle lateral velocity and acceleration;
Figure BDA0003707394170000057
and
Figure BDA0003707394170000058
respectively the yaw angular velocity and the angular acceleration of the vehicle; c f 、C r The cornering stiffness of the front and rear wheels, respectively; i is z The moment of inertia of the whole vehicle mass around the z axis;
(226) combining the motion state SS in the short time domain acquired in steps (224) and (225) K Or SS D And adopting a fifth-order polynomial fitting to generate vehicle motion state information in a future long-term domain, and recording the information as
Figure BDA0003707394170000059
Figure BDA00037073941700000510
Wherein t is time; a is i And b i Are all polynomial coefficients, i is 0,1,2,3,4, 5; v. of X And v Y The components of the vehicle speed v on the X-axis and Y-axis, respectively.
Further, the driving intention recognition of the vehicle in the step (23) adopts an LSTM neural network method, and the input layer parameters of the LSTM neural network are selected as follows: target vehicle speed, lateral acceleration, vehicle offset lane centerline position, and vehicle lateral position information.
Further, the method for predicting the future motion state of the vehicle in the step (23) adopts a method of combining a fifth-order polynomial with multi-objective optimization, and after the driving intention is obtained, a series of candidate trajectories tra ═ L are generated by using the fifth-order polynomial 1 ,L 2 ,…,L n ]Then, designing an objective function, screening out a track which accords with the reality as a predicted track by adopting a multi-objective optimization method, and outputting the future motion state information of the vehicle, wherein the objective function is as follows:
Figure BDA0003707394170000061
in the formula, a yL (t) is the lateral acceleration of the trajectory L; a is ymax Maximum lateral acceleration allowed; LO L Is the length of the trajectory L; LO max Is the maximum allowable track length; Δ L is a trajectory error between the candidate trajectory and the previous time; delta max Is the maximum allowable trajectory error; w is a 1 、w 2 And w 3 The sum of the three is 1.
Further, the omni-directional collision risk assessment model in step (3) is established as follows:
(31) establishing an omnidirectional collision time OTTC model:
Figure BDA0003707394170000062
in the formula, OR is the linear distance between the vehicle and the mass center of other vehicles; Δ OV is the relative speed of the own vehicle and other vehicles; (X) 0 ,Y 0 ) Is the centroid position of the bicycle; (X) i ,Y i ) Is the centroid position of the target vehicle; v 0 And V i The speeds of the own vehicle and the other vehicles respectively; theta is an included angle between the direction of the head of the bicycle and a connecting line of the mass centers of the two bicycles; sgn is a sign function;
Figure BDA0003707394170000063
and
Figure BDA0003707394170000064
the yaw angles of the self vehicle and other vehicles respectively; pi is the circumference ratio;
(32) establishing an OTHW model of the omnidirectional collision headway:
Figure BDA0003707394170000065
(33) establishing an omnidirectional safe distance OR safe Model:
Figure BDA0003707394170000066
in the formula, D 0safe And D isafe The emergency braking distances of the self vehicle and other vehicles are respectively;
(34) combining the OTTC model, the OTHW model and the OR established in the steps (31) to (33) safe And (3) establishing an omnidirectional collision risk assessment model:
Figure BDA0003707394170000067
in the formula, xi is the risk of omnidirectional collision; g is the acceleration of gravity.
Further, the specific steps of the step (4) are as follows:
(41) the intelligent networked vehicle calculates the driving risk of the current position in real time and receives the average flow of each lane sent by the road side base station;
(42) if the driving risk xi of the current position of the intelligent networked vehicle is greater than the risk threshold xi max And is
Figure BDA0003707394170000068
Greater than a risk change rate threshold
Figure BDA0003707394170000069
I.e. xi>ξ max And is
Figure BDA00037073941700000610
Or the average flow rate Q of the current driving lane c Less than average flow Q of adjacent lanes n I.e. Q c <Q n (ii) a The intelligent networked vehicle sends a lane change decision command, and generates a candidate track for changing lanes to an adjacent lane by adopting a quintic polynomial method;
(43) inputting the candidate tracks generated in the step (42) into an omnidirectional collision risk evaluation model, evaluating the omnidirectional collision risk of each lane changing track, screening out candidate tracks meeting safe lane changing conditions, further screening the candidate tracks by using the target function designed in the step (23), and selecting an optimal lane changing track;
(44) if the optimal lane change track which cannot be solved in the step (43) finally is found, the optimal lane change track is xi>ξ max And is
Figure BDA0003707394170000071
While performing an emergency braking operation, and at Q c <Q n The operation of braking, decelerating and following the front vehicle is carried out.
The invention has the beneficial effects that:
1. the method considers the influence of heterogeneous (different intelligent levels) characteristics of vehicles running on a future road on the decision planning of the intelligent networked automobile, and greatly improves the decision planning efficiency of the intelligent networked automobile under heterogeneous traffic flows by fully utilizing the intelligent networking technology and the automatic driving technology.
2. According to the method, the road side base station can be used for accurately acquiring the future motion state of a part of vehicles with networking capability, a prediction method is not needed for indirect acquisition, the calculation efficiency is improved, and the precision of a decision planning stage is also improved.
Drawings
FIG. 1 is a schematic diagram of a method of the present invention;
fig. 2 is a heterogeneous traffic flow diagram.
Detailed Description
In order to facilitate understanding of those skilled in the art, the present invention is further described below with reference to the following examples and the accompanying drawings, which are not intended to limit the present invention.
Referring to fig. 1, the intelligent internet automobile decision planning method under heterogeneous traffic flow of the invention comprises the following steps:
(1) the intelligent network connection vehicle receives a check information stream sent by a side base station on the road, identifies the intelligent network connection state of other vehicles on the local road section, and calculates the traffic capacity of each lane; as shown with reference to FIG. 2;
the step of identifying the intelligent networking state of other vehicles on the local road section comprises the following steps: the road side base station sends a check information stream to a vehicle running on the road section, the vehicle returns a corresponding check information stream after receiving the check information stream to reflect the receiving state of the vehicle, if the road side base station receives the check information stream returned by the vehicle, the road side base station divides the vehicle into intelligent network connection vehicles, and the vehicle is judged to be in an automatic or manual driving state according to the returned information; and if the road side base station does not receive the check information flow returned by the vehicle, dividing the vehicle into non-intelligent network connection vehicles.
The calculation mode of the traffic capacity of each lane is as follows:
Q i ＝VR i ×D i (1)
in the formula, Q i The average traffic flow (vehicle/h) of the ith lane represents the traffic capacity of the lane; VR (virtual reality) i A lane average vehicle speed (m/s) for an ith lane; d i Is the average traffic density (vehicle/km) of the ith lane.
(2) Acquiring the motion state of other vehicles on the road within a period of time in the future according to the intelligent networking state identification result of other vehicles in the step (1);
the method for acquiring the motion state of other vehicles on the road in a future period of time by adopting a diversified method comprises the following steps:
(21) if the other vehicles are intelligent network connection vehicles and are in an automatic driving state according to the identification result in the step (1), directly utilizing network connection communication to obtain the future motion states of the other vehicles and calibrating the obtained information by combining the roadside base station information;
(22) if the other vehicles are intelligent network communication vehicles and are in a manual driving state in the identification result in the step (1), acquiring the state and vehicle state information of a driver of the vehicle by utilizing network communication, identifying the driving intention by combining the state of the driver and the state of the vehicle, and predicting the future motion state of the vehicle;
(23) and (3) if the other vehicle is a non-intelligent networked vehicle as a result of the identification in the step (1), acquiring the state information of the vehicle by using a vehicle-mounted sensing sensor, and performing driving intention identification and future motion state prediction of the vehicle by using the acquired vehicle state information.
Specifically, the driving intention identifying step in the step (22) is as follows:
(221) the off-line collection of driver state information and vehicle state information includes: establishing a driver intention recognition data set by information of a driver sight line focus, a heart rate, a breathing rate, a head rotation angle, a vehicle speed, a vehicle acceleration, a yaw rate, a steering wheel angle, a steering wheel angular speed, a vehicle offset lane center line position and a vehicle lateral position;
(222) calibrating the lane change intention of the driver corresponding to each group of data in the data set established in the step (221), updating the weights of all parameters contained in the data set by adopting a Relieff algorithm, and sequencing the weights of all parameters to screen out characteristic parameters which can best reflect the lane change intention of the driver, wherein the calculation method of the weight W (A) of any parameter A comprises the following steps:
Figure BDA0003707394170000081
Figure BDA0003707394170000082
in the formula, diff (A, R, H) j ) Denotes sample R and sample H j The difference in parameter A, p (C) is the proportion of C ≠ class (R), p (classes (R)) is the proportion of samples of the same class as samples R, M j (C) Indicating that the jth nearest neighbor sample in the class C ≠ class (R); k is the number of samples selected and classified as the same as the parameter A; m is a sample category; max and min are functions for solving the maximum value and the minimum value respectively;
(223) and (3) taking the characteristic parameters screened in the step (222) as an input layer of the LSTM neural network, taking the lane changing intention of the driver corresponding to each group of parameters as an output layer of the LSTM neural network, and training the LSTM neural network to identify the driving intention of the driver, wherein the specific steps are as follows:
(2231) calculating a forgetting gate:
f t ＝σ(W f ·[h t-1 ,X t ])+b f ) (4)
in the formula (f) t The value range of the forgotten gate at the current moment is 0 to 1; w f A forgetting gate weight value; x t Is the input value at the current moment; h is t-1 Is the output value of the last moment; b f Biasing for a forget gate; sigma is sigmoid function;
(2232) calculation input gate:
i t ＝σ(W i ·[h t-1 ,X t ])+b i ) (5)
in the formula i t The value range of the input gate at the current moment is 0 to 1; w i The weight value of the input gate; b i Biasing the input gate;
(2233) calculating candidate memory cell information:
Figure BDA0003707394170000091
in the formula (I), the compound is shown in the specification,
Figure BDA0003707394170000092
candidate information to be updated to the memory unit at the current time; w C Is a candidate information weight value; b C Biasing the candidate information; tan h is a hyperbolic tangent function;
(2234) calculating new memory cell information:
Figure BDA0003707394170000093
in the formula, C t New memory cell information at the current moment; c t-1 Memory cell information of the previous moment;
(2235) calculating the LSTM neural network output:
o t ＝σ(W o ·[h t-1 ,X t ])+b o ) (8)
h t ＝o t ·tanh(C t ) (9)
in the formula o t Is the initial output of the current moment; w is a group of o Is the initial output weight value; b o Is an initial bias; h is t The output of the current moment is the driving intention of the driver.
Specifically, the step (22) of predicting the future motion state of the vehicle comprises the following steps:
(224) establishing a short time domain low-speed kinematic prediction model, predicting the future motion state of the vehicle when the vehicle runs at low speed, and recording the future motion state as
Figure BDA0003707394170000094
Figure BDA0003707394170000101
Wherein X is the longitudinal position of the vehicle; y is the lateral position of the vehicle; v is the vehicle speed;
Figure BDA0003707394170000102
is the vehicle yaw angle; beta is the centroid slip angle; l f Is the distance from the center of mass of the vehicle to the front axle; l r Is the distance from the vehicle center of mass to the rear axle; a is vehicle acceleration; delta f Is the front wheel corner of the vehicle; sin, cos and tan are sine, cosine and tangent functions, respectively, X t Longitudinal position of vehicle at time t, X t+1 The longitudinal position of the vehicle at time t + 1; y is t For the transverse position of the vehicle at time t, Y t+1 The lateral position of the vehicle at the time t + 1;
Figure BDA0003707394170000103
for the yaw angle of the vehicle at time t,
Figure BDA0003707394170000104
the vehicle yaw angle at time t + 1; v. of t Vehicle speed at time t, v t+1 The vehicle speed at the moment t + 1;
(225) establishing a short time domain high-speed dynamics prediction modelType, predicting the future motion state of the vehicle at high speed, and recording
Figure BDA0003707394170000105
Figure BDA0003707394170000106
In the formula, m is the total vehicle mass of the vehicle;
Figure BDA0003707394170000107
and
Figure BDA0003707394170000108
respectively the longitudinal speed and the acceleration of the vehicle under a vehicle coordinate system;
Figure BDA0003707394170000109
and
Figure BDA00037073941700001010
respectively the lateral speed and the acceleration of the vehicle under a vehicle coordinate system;
Figure BDA00037073941700001011
and
Figure BDA00037073941700001012
yaw angular velocity and angular acceleration of the vehicle, respectively; c f 、C r The cornering stiffness of the front and rear wheels, respectively; i is z The moment of inertia of the whole vehicle mass around the z axis;
(226) combining the motion state SS in the short time domain acquired in steps (224) and (225) K Or SS D And adopting a fifth-order polynomial fitting to generate vehicle motion state information in a future long-term domain, and recording the information as
Figure BDA00037073941700001013
Figure BDA0003707394170000111
Wherein t is time; a is i And b i Are all polynomial coefficients, i is 0,1,2,3,4, 5; v. of X And v Y The components of the vehicle speed v on the X-axis and Y-axis, respectively.
In the step (23), the driving intention of the vehicle is identified by adopting an LSTM neural network method, and the input layer parameters of the LSTM neural network are selected as follows: target vehicle speed, lateral acceleration, vehicle offset lane centerline position, and vehicle lateral position information.
In the step (23), a method of combining a quintic polynomial with multi-target optimization is adopted for predicting the future motion state of the vehicle, and after the driving intention is obtained, a series of candidate trajectories tra ═ L are generated by adopting the quintic polynomial 1 ,L 2 ,…,L n ]Then, designing an objective function, screening out a track which accords with the reality as a predicted track by adopting a multi-objective optimization method, and outputting the future motion state information of the vehicle, wherein the objective function is as follows:
Figure BDA0003707394170000112
in the formula, a yL (t) is the lateral acceleration of the trajectory L; a is ymax Maximum lateral acceleration allowed; LO L Is the length of the trajectory L; LO max Is the maximum allowable track length; Δ L is a trajectory error between the candidate trajectory and the previous time; delta max Is the maximum allowable trajectory error; w is a 1 、w 2 And w 3 The sum of the three is 1.
(3) Establishing an omnidirectional collision risk evaluation model, inputting the motion state information acquired in the step (2) into the omnidirectional collision risk evaluation model, and calculating the driving risk of the intelligent networked vehicle in real time;
the omnidirectional collision risk assessment model is established by the following steps:
(31) establishing an OTTC model of the omnidirectional collision time:
Figure BDA0003707394170000113
in the formula, OR is the linear distance between the vehicle and the mass center of other vehicles; Δ OV is the relative speed of the own vehicle and other vehicles; (X) 0 ,Y 0 ) Is the centroid position of the bicycle; (X) i ,Y i ) Is the centroid position of the target vehicle; v 0 And V i The speeds of the own vehicle and other vehicles respectively; theta is an included angle between the direction of the head of the bicycle and a connecting line of the mass centers of the two bicycles (the anticlockwise direction is positive); sgn is a sign function;
Figure BDA0003707394170000114
and
Figure BDA0003707394170000115
the yaw angles of the self vehicle and other vehicles respectively; pi is the circumference ratio;
(32) establishing an OTHW model of the omnidirectional collision headway:
Figure BDA0003707394170000116
(33) establishing an omnidirectional safe distance OR safe Model:
Figure BDA0003707394170000121
in the formula, D 0safe And D isafe The emergency braking distances of the self vehicle and other vehicles are respectively;
(34) combining the OTTC model, the OTHW model and the OR established in the steps (31) to (33) safe And (3) establishing an omnidirectional collision risk assessment model:
Figure BDA0003707394170000122
in the formula, xi is the omnidirectional collision risk; g is the acceleration of gravity.
(4) According to the traffic capacity, the driving risk and the driving risk change rate of a lane where the intelligent networked vehicle is located and an adjacent lane, whether the vehicle needs to change the lane or not is determined, and then a lane change track is determined according to the traffic efficiency of the adjacent lane and the driving risk change condition from the lane change to the adjacent lane;
(41) the intelligent networked vehicle calculates the driving risk of the current position in real time and receives the average flow of each lane sent by the road side base station;
(42) if the driving risk xi of the current position of the intelligent networked vehicle is greater than the risk threshold xi max And is
Figure BDA0003707394170000123
Greater than a risk change rate threshold
Figure BDA0003707394170000124
I.e. xi>ξ max And is
Figure BDA0003707394170000125
Or the average flow rate Q of the current driving lane c Less than average flow Q of adjacent lanes n I.e. Q c <Q n (ii) a The intelligent networked vehicle sends a lane change decision command, and generates a candidate track for changing lanes to an adjacent lane by adopting a quintic polynomial method;
(43) inputting the candidate tracks generated in the step (42) into an omnidirectional collision risk evaluation model, evaluating the omnidirectional collision risk of each lane changing track, screening out candidate tracks meeting safe lane changing conditions, further screening the candidate tracks by using the target function designed in the step (23), and selecting an optimal lane changing track;
(44) if the optimal lane change track which cannot be solved in the step (43) finally is found, the optimal lane change track is xi>ξ max And is
Figure BDA0003707394170000126
While performing an emergency braking operation, and at Q c <Q n The operation of braking, decelerating and following the front vehicle is carried out.
While the invention has been described in terms of its preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention.

Claims (10)

1. An intelligent network connection automobile decision planning method under heterogeneous traffic flow is characterized by comprising the following steps:
(1) the intelligent internet vehicle receives the check information stream sent by the road side base station, identifies the intelligent internet state of other vehicles on the local road section, and calculates the traffic capacity of each lane;
(2) acquiring the motion state of other vehicles on the road within a period of time in the future according to the intelligent networking state identification result of other vehicles in the step (1);
(3) establishing an omnidirectional collision risk assessment model, inputting the motion state information acquired in the step (2) into the omnidirectional collision risk assessment model, and calculating the driving risk of the intelligent networked vehicle in real time;
(4) according to the traffic capacity, the driving risk and the driving risk change rate of a lane where the intelligent networked vehicle is located and an adjacent lane, whether the vehicle needs to change the lane or not is determined, and then a lane change track is determined according to the traffic efficiency of the adjacent lane and the driving risk change condition from the lane change to the adjacent lane.
2. The intelligent internet-connected automobile decision planning method under the heterogeneous traffic flow according to claim 1, wherein the step of identifying the intelligent internet-connected state of other vehicles on the local road section in the step (1) comprises the steps of: the road side base station sends a check information stream to a vehicle running on the road section, the vehicle returns a corresponding check information stream after receiving the check information stream to reflect the receiving state of the vehicle, if the road side base station receives the check information stream returned by the vehicle, the road side base station divides the vehicle into intelligent network connection vehicles, and the vehicle is judged to be in an automatic or manual driving state according to the returned information; and if the road side base station does not receive the check information flow returned by the vehicle, dividing the vehicle into non-intelligent network connection vehicles.
3. The intelligent internet automobile decision planning method under the heterogeneous traffic flow according to claim 1, wherein the calculation mode of the traffic capacity of each lane in the step (1) is as follows:
Q i ＝VR i ×D i (1)
in the formula, Q i The traffic average flow of the ith lane represents the traffic capacity of the lane; VR (virtual reality) i The average vehicle speed of the i-th lane is taken as the average vehicle speed of the i-th lane; d i Is the average traffic density of the ith lane.
4. The intelligent internet automobile decision planning method under the heterogeneous traffic flow according to claim 1, wherein the step (2) of obtaining the motion state of other vehicles on the road in a future period of time by adopting a diversified method comprises the following steps:
(21) if the other vehicles are intelligent network connection vehicles and are in an automatic driving state according to the identification result in the step (1), directly utilizing network connection communication to obtain the future motion states of the other vehicles and calibrating the obtained information by combining the roadside base station information;
(22) if the other vehicles are intelligent network communication vehicles and are in a manual driving state in the identification result in the step (1), acquiring the state and vehicle state information of a driver of the vehicle by utilizing network communication, identifying the driving intention by combining the state of the driver and the state of the vehicle, and predicting the future motion state of the vehicle;
(23) and (3) if the other vehicle is a non-intelligent networked vehicle as a result of the identification in the step (1), acquiring the state information of the vehicle by using a vehicle-mounted sensing sensor, and performing driving intention identification and future motion state prediction of the vehicle by using the acquired vehicle state information.
5. The intelligent internet automobile decision planning method under the heterogeneous traffic flow according to claim 4, wherein the driving intention identification step in the step (22) is as follows:
(221) the off-line collection of driver state information and vehicle state information includes: establishing a driver intention recognition data set by information of a driver sight line focus, a heart rate, a breathing rate, a head rotation angle, a vehicle speed, a vehicle acceleration, a yaw rate, a steering wheel angle, a steering wheel angular speed, a vehicle offset lane center line position and a vehicle lateral position;
(222) calibrating the lane changing intention of the driver corresponding to each group of data in the data set established in the step (221), updating the weights of all parameters contained in the data set by adopting a Relieff algorithm, sequencing the weights of all parameters to screen out characteristic parameters capable of reflecting the lane changing intention of the driver most, wherein the weight W (A) of any parameter A is calculated by the following steps:
Figure FDA0003707394160000021
Figure FDA0003707394160000022
in the formula, diff (A, R, H) j ) Denotes sample R and sample H j The difference in parameter A, p (C) is the proportion of C ≠ class (R), p (classes (R)) is the proportion of samples of the same class as samples R, M j (C) Indicating that the jth nearest neighbor sample in the class C ≠ class (R); k is the number of samples selected and classified as the same as the parameter A; m is a sample category; max and min are functions for solving the maximum value and the minimum value respectively;
(223) and (3) taking the characteristic parameters screened in the step (222) as an input layer of the LSTM neural network, taking the lane changing intention of the driver corresponding to each group of parameters as an output layer of the LSTM neural network, and training the LSTM neural network to identify the driving intention of the driver, wherein the specific steps are as follows:
(2231) calculating a forgetting gate:
f t ＝σ(W f ·[h t-1 ,X t ])+b f ) (4)
in the formula (f) t The value range is 0 to 1 for the forgetting gate at the current moment; w f To forget the doorA weight value; x t Is the input value at the current moment; h is t-1 Is the output value of the last moment; b is a mixture of f Biasing for a forget gate; sigma is sigmoid function;
(2232) calculation input gate:
i t ＝σ(W i ·[h t-1 ,X t ])+b i ) (5)
in the formula i t The input gate at the current moment has a value range of 0 to 1; w i The weight value of the input gate; b i Biasing the input gate;
(2233) calculating candidate memory cell information:
Figure FDA0003707394160000031
in the formula (I), the compound is shown in the specification,
Figure FDA0003707394160000032
candidate information to be updated to the memory unit at the current time; w C The candidate information weight value is obtained; b C Biasing the candidate information; tan h is a hyperbolic tangent function;
(2234) calculating new memory cell information:
Figure FDA0003707394160000033
in the formula, C t New memory cell information at the current moment; c t-1 Memory cell information of the previous moment;
(2235) calculating the LSTM neural network output:
o t ＝σ(W o ·[h t-1 ,X t ])+b o ) (8)
h t ＝o t ·tanh(C t ) (9)
in the formula o t Is the initial output of the current moment; w o Is the initial output weight value; b o Is an initial bias; h is t Is that whenThe output of the previous moment is the driving intention of the driver.
6. The intelligent networked automobile decision planning method under the heterogeneous traffic flow according to claim 4, wherein the step (22) of predicting the future motion state of the vehicle comprises the following steps:
(224) establishing a short time domain low-speed kinematic prediction model, predicting the future motion state of the vehicle when the vehicle runs at low speed, and recording the future motion state as
Figure FDA0003707394160000034
Figure FDA0003707394160000035
Wherein X is the longitudinal position of the vehicle; y is the lateral position of the vehicle; v is the vehicle speed;
Figure FDA0003707394160000036
is the vehicle yaw angle; beta is the centroid slip angle; l f Is the distance from the center of mass of the vehicle to the front axle; l r Is the distance from the center of mass of the vehicle to the rear axle; a is the vehicle acceleration; delta f Is the front wheel corner of the vehicle; sin, cos and tan are sine, cosine and tangent functions, respectively, X t Longitudinal position of vehicle at time t, X t+1 The longitudinal position of the vehicle at time t + 1; y is t For the transverse position of the vehicle at time t, Y t+1 The lateral position of the vehicle at time t + 1;
Figure FDA0003707394160000037
for the yaw angle of the vehicle at time t,
Figure FDA0003707394160000038
the vehicle yaw angle at time t + 1; v. of t Vehicle speed at time t, v t+1 The vehicle speed at the moment t + 1;
(225) establishing short time domain high speed dynamics prediction model for predicting vehicle high speed runningFuture movement states, note
Figure FDA0003707394160000041
Figure FDA0003707394160000042
In the formula, m is the total vehicle mass of the vehicle;
Figure FDA0003707394160000043
and
Figure FDA0003707394160000044
respectively the longitudinal speed and the acceleration of the vehicle under a vehicle coordinate system;
Figure FDA0003707394160000045
and
Figure FDA0003707394160000046
respectively the lateral speed and the acceleration of the vehicle under a vehicle coordinate system;
Figure FDA0003707394160000047
and
Figure FDA0003707394160000048
yaw angular velocity and angular acceleration of the vehicle, respectively; c f 、C r The cornering stiffness of the front and rear wheels, respectively; i is z The moment of inertia of the whole vehicle mass around the z axis;
(226) combining the motion state SS in the short time domain acquired in steps (224) and (225) K Or SS D And adopting a fifth-order polynomial fitting to generate vehicle motion state information in a future long-term domain, and recording the information as
Figure FDA0003707394160000049
Figure FDA00037073941600000410
Wherein t is time; a is i And b i Are all polynomial coefficients, i is 0,1,2,3,4, 5; v. of X And v Y The components of the vehicle speed v on the X-axis and Y-axis, respectively.
7. The intelligent networked automobile decision-making planning method under the heterogeneous traffic flow according to claim 4, wherein the driving intention recognition of the vehicles in the step (23) adopts an LSTM neural network method, and the input layer parameters of the LSTM neural network are selected as follows: target vehicle speed, lateral acceleration, vehicle offset lane centerline position, and vehicle lateral position information.
8. The intelligent internet automobile decision planning method under the heterogeneous traffic flow according to claim 4, wherein the future motion state prediction method of the vehicles in the step (23) adopts a method combining a quintic polynomial and multi-objective optimization, and after the driving intention is obtained, a series of candidate trajectories tra ═ L are generated by adopting the quintic polynomial 1 ,L 2 ,…,L n ]Then, designing an objective function, screening out a track which accords with the reality as a predicted track by adopting a multi-objective optimization method, and outputting the future motion state information of the vehicle, wherein the objective function is as follows:
Figure FDA0003707394160000051
in the formula, a yL (t) is the lateral acceleration of the trajectory L; a is ymax Maximum lateral acceleration allowed; LO L Is the length of the trajectory L; LO max Is the maximum allowable track length; Δ L is a trajectory error between the candidate trajectory and the previous time; delta max Is the maximum allowable trajectory error; w is a 1 、w 2 And w 3 The sum of the three is 1.
9. The intelligent internet automobile decision planning method under the heterogeneous traffic flow according to claim 4, wherein the omnidirectional collision risk assessment model in the step (3) is established by the following steps:
(31) establishing an omnidirectional collision time OTTC model:
Figure FDA0003707394160000052
in the formula, OR is the straight-line distance between the vehicle and the mass center of other vehicles; Δ OV is the relative speed of the own vehicle and other vehicles; (X) 0 ,Y 0 ) Is the centroid position of the bicycle; (X) i ,Y i ) Is the centroid position of the target vehicle; v 0 And V i The speeds of the own vehicle and other vehicles respectively; theta is an included angle between the direction of the head of the bicycle and a connecting line of the mass centers of the two bicycles; sgn is a sign function;
Figure FDA0003707394160000053
and
Figure FDA0003707394160000054
the yaw angles of the self vehicle and other vehicles respectively; pi is the circumference ratio;
(32) establishing an OTHW model of the omnidirectional collision headway:
Figure FDA0003707394160000055
(33) establishing an omnidirectional safe distance OR safe Model:
Figure FDA0003707394160000056
in the formula, D 0safe And D isafe The emergency braking distances of the self vehicle and other vehicles are respectively;
(34) establishing in combination with said steps (31) - (33)OTTC model, OTHW model and OR safe And (3) establishing an omnidirectional collision risk assessment model:
Figure FDA0003707394160000057
in the formula, xi is the omnidirectional collision risk; g is the gravitational acceleration.
10. The heterogeneous traffic flow intelligent internet automobile decision-making planning method according to claim 4, wherein the specific steps of the step (4) are as follows:
(41) the intelligent networked vehicle calculates the driving risk of the current position in real time and receives the average flow of each lane sent by the road side base station;
(42) if the driving risk xi of the current position of the intelligent networked vehicle is greater than the risk threshold xi max And is
Figure FDA0003707394160000061
Greater than a risk change rate threshold
Figure FDA0003707394160000062
I.e. xi>ξ max And is
Figure FDA0003707394160000063
Or the average flow rate Q of the current driving lane c Less than average flow Q of adjacent lanes n I.e. Q c ＜Q n (ii) a The intelligent networked vehicle sends a lane change decision command, and generates a candidate track for changing lanes to an adjacent lane by adopting a quintic polynomial method;
(43) inputting the candidate tracks generated in the step (42) into an omnidirectional collision risk evaluation model, evaluating the omnidirectional collision risk of each lane changing track, screening out candidate tracks meeting safe lane changing conditions, further screening the candidate tracks by using the target function designed in the step (23), and selecting an optimal lane changing track;
(44) if step (43) is not completedThe optimal track changing track is obtained through solving, and then the optimal track changing track is in xi>ξ max And is
Figure FDA0003707394160000064
While performing an emergency braking operation, and at Q c ＜Q n The operation of braking, decelerating and following the front vehicle is carried out.
CN202210712562.3A 2022-06-22 2022-06-22 Intelligent network-connected automobile decision planning method under heterogeneous traffic flow Pending CN115018353A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210712562.3A CN115018353A (en) 2022-06-22 2022-06-22 Intelligent network-connected automobile decision planning method under heterogeneous traffic flow

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210712562.3A CN115018353A (en) 2022-06-22 2022-06-22 Intelligent network-connected automobile decision planning method under heterogeneous traffic flow

Publications (1)

Publication Number Publication Date
CN115018353A true CN115018353A (en) 2022-09-06

Family

ID=83077877

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210712562.3A Pending CN115018353A (en) 2022-06-22 2022-06-22 Intelligent network-connected automobile decision planning method under heterogeneous traffic flow

Country Status (1)

Country Link
CN (1) CN115018353A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117196153A (en) * 2023-09-25 2023-12-08 广东汇通信息科技股份有限公司 Intelligent traffic comprehensive management system based on traffic big data

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117196153A (en) * 2023-09-25 2023-12-08 广东汇通信息科技股份有限公司 Intelligent traffic comprehensive management system based on traffic big data

Similar Documents

Publication Publication Date Title
CN113386795B (en) Intelligent decision-making and local track planning method for automatic driving vehicle and decision-making system thereof
CN110362096B (en) Unmanned vehicle dynamic trajectory planning method based on local optimality
CN111780981B (en) Intelligent vehicle formation lane change performance evaluation method
Zhang et al. Chassis coordinated control for full X-by-wire vehicles-A review
CN108848462B (en) Real-time vehicle track prediction method suitable for signal control intersection
CN109727469B (en) Comprehensive risk degree evaluation method for automatically driven vehicles under multiple lanes
WO2019213981A1 (en) Real-time driving risk assessment method employing equivalent force and device thereof
CN111717192A (en) Control method and system for automatically driving vehicle
CN112249008B (en) Unmanned automobile early warning method aiming at complex dynamic environment
CN111599179B (en) No-signal intersection vehicle motion planning method based on risk dynamic balance
CN112577506B (en) Automatic driving local path planning method and system
CN109643118A (en) The function of vehicle is influenced based on the information relevant to function of the environment about vehicle
CN113721606B (en) Following type automatic driving logistics vehicle control system and method
CN116499486B (en) Complex off-road environment path planning method and system and electronic equipment
CN115662131B (en) Multi-lane collaborative lane changing method for road accident section in network environment
CN116605188B (en) Automatic emergency braking control system for electric vehicle-two-wheel vehicle
CN115018353A (en) Intelligent network-connected automobile decision planning method under heterogeneous traffic flow
CN114580302A (en) Decision planning method for automatic driving automobile based on maximum entropy reinforcement learning
CN115523934A (en) Vehicle track prediction method and system based on deep learning
Tian et al. Vehicle state estimation based on multidimensional information fusion
CN108711285A (en) A kind of mixed traffic analogy method based on road cross
CN114987539A (en) Individual collision grading early warning method and system for automatic driving automobile based on risk field model
Lin et al. Adaptive prediction-based control for an ecological cruise control system on curved and hilly roads
Skrickij et al. Review of integrated chassis control techniques for automated ground vehicles
CN116039672A (en) Data-model back-end fusion method for automatic driving decision control function

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