CN113450564B - Intersection passing method based on NARX neural network and C-V2X technology - Google Patents
Intersection passing method based on NARX neural network and C-V2X technology Download PDFInfo
- Publication number
- CN113450564B CN113450564B CN202110559099.9A CN202110559099A CN113450564B CN 113450564 B CN113450564 B CN 113450564B CN 202110559099 A CN202110559099 A CN 202110559099A CN 113450564 B CN113450564 B CN 113450564B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- data
- information
- speed
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
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
- G08G1/0104—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
-
- 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
-
- 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
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
- G08G1/0141—Measuring and analyzing of parameters relative to traffic conditions for specific applications for traffic information dissemination
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/091—Traffic information broadcasting
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Analytical Chemistry (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Chemical & Material Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Atmospheric Sciences (AREA)
- Traffic Control Systems (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The invention discloses a method for passing through an intersection based on an NARX neural network and a C-V2X technology, and relates to the field of vehicle-road cooperation and 5G communication. The method comprises the following steps: collecting historical data of relevant parameters required by vehicle speed prediction, processing the data, dividing the data into training data and test data in proportion, and inputting the training data into an NARX neural network for training; inputting the test data into a trained NARX neural network, and processing the output value to obtain an actual predicted value; and judging whether the vehicle can smoothly pass through the intersection by reducing the starting and stopping times or not according to the actual predicted vehicle speed of the vehicle, the actual vehicle speed of the front vehicle, the distance between the vehicle and the intersection and the remaining time of a signal lamp, and broadcasting the recommended vehicle speed for economic driving to a driver. The model has high prediction precision and high transmission efficiency, and can be used for urban road vehicles. The invention can effectively improve the driving safety and improve the traffic jam and the environmental pollution.
Description
Technical Field
The invention relates to the field of vehicle-road cooperation and 5G communication, in particular to a scheme for realizing intelligent travel through information fusion of people, vehicles and roads.
Background
In 3 months of 2020, the U.S. department of transportation issues Intelligent Transportation System (ITS) strategic plan 2020 + 2025, which specifies the vision of "accelerating the application of an Intelligent transportation System and transforming social operation modes", and the vision of "leading the cooperation and innovation research, development and implementation of an Intelligent transportation System to provide the mission of safety and mobility of personnel's commuting and cargo transportation", and describes the key tasks and safeguards of the future five-year Intelligent transportation development in the U.S. In recent years, China gives full play to technical advantages and system advantages, establishes an Internet of vehicles industry development committee, promotes industrial development jointly in the fields of inter-industry coordinated communication, automobiles, traffic, public security and the like, supports the construction of national Internet of vehicles leaders in important areas such as Wuxi, Tianjin Xiqing and the like in Jiangsu, actively promotes the deployment of Internet of vehicles wireless communication technology based on the evolution of mobile communication technology, and achieves better effect.
The development and application of the internet traffic and intelligent driving technology bring new opportunities to the problems of improving driving safety, traffic congestion, environmental pollution and the like, and simultaneously provide new challenges for traffic control theories and technologies in new traffic forms. With the continuous progress of information technology, the demand of users is increasing, and the demand is more and more satisfied in order to better satisfy the multifunctional and diversified experience demand of users. The 5G network technology has the characteristics of high stability, high transmission rate, low time delay and flexible network architecture. The car networking communication architecture based on the 5G network technology is continuously changing to meet various requirements. The application of 5G network technology will make traditional car networking communication mode change, and car networking communication architecture is more intelligent, more nimble, and system element is also more diversified. The C-V2X communication technology is an important component of future intelligent traffic system development, allows vehicles, other vehicles, infrastructure and vulnerable traffic participants to be in deep communication interconnection, constructs a real comprehensive human, vehicle, road and cloud network, provides technical support for future traffic intellectualization and networking, and provides guarantee for vehicle users to realize driving safety.
In recent years, artificial neural networks have proven to be a promising technology in terms of time series prediction, energy assessment, and pattern reorganization. In the application of time series prediction, several neural network types are employed, such as non-linear autoregressive exogenous neural Network (NARX), non-linear autoregressive neural Network (NAR), and Recurrent Neural Network (RNN). The event occurring due to the vehicle speed is dynamic and is related to and influenced by the historical state of the system where the event occurs, including the driving habits of the driver, the running performance of the engine, the execution mechanism, the driving road condition and the like. In such dynamic and non-linear cases, it is highly advantageous to use neural network structures such as dynamic Recurrent Neural Networks (RNN), non-linear autoregressive neural Networks (NAR) and non-linear autoregressive exogenous neural Networks (NARX) as processing tools that can receive dynamic inputs represented by sets of time series.
The vehicle speed prediction refers to estimation and reasoning of a vehicle speed track sequence of a vehicle in a future period of time, and has wide application space in the fields of vehicle intellectualization, new energy vehicles and the like.
The method can effectively help solve the problem of urban traffic congestion by accurately predicting the driving speed on the urban road, and also helps a driver to stably drive through a complex intersection, thereby reducing the situations such as safety problems caused by frequent and urgent braking and urgent acceleration. In the field of automobile intellectualization, vehicle safety driving assistance and intelligent vehicle behavior decision analysis need to analyze vehicle driving data to make a corresponding strategy, and the more advanced the data prediction is, the more accurate the decision accuracy is. The vehicle speed running data of a period of time in the future can be predicted according to the data of a period of time before the current moment through a vehicle speed prediction algorithm, and transmitted to a decision mechanism, and the decision efficiency and precision can be improved through analysis and processing.
By predicting the running speed of the urban road and analyzing and processing the speed, a driver can stably pass through a signal lamp intersection under the optimal running condition by means of a data processing result, the number of starting and stopping times in the running process can be effectively reduced, and the fuel economy of the vehicle is improved. For a hybrid electric vehicle, the vehicle can be driven to achieve the optimal power distribution of the engine and the motor; for a pure electric new energy automobile, the energy consumption of a battery can be reasonably arranged through vehicle speed prediction, and the optimal driving range of the automobile is calculated.
The speed of the automobile can reflect the influence of the driver behavior, road environment and road facilities on actual driving, so that the automobile is influenced by various factors such as people, vehicles, roads, environment and the like, accurate prediction is very difficult, and the automobile speed prediction method has important significance on the research on the length of prediction time and the height of prediction accuracy.
Disclosure of Invention
In the system, under the 5G network environment, the most economical driving scheme is provided for a driver to stably pass through a crossing Road through information exchange and processing among a Road Side Unit (RSU), an On Board Unit (OBU) and an Edge server (Edge computing). The technical scheme comprises the following specific contents:
a method for passing through an intersection based on a NARX neural network and a C-V2X technology comprises the following steps:
the method comprises the following steps: the road side unit RSU acquires signal lamp information and sends the signal lamp information to the Edge cloud server Edge and the vehicle-mounted unit OBU, and meanwhile position information of a road junction stop line in the direction is acquired;
step two: the method comprises the steps that a vehicle-mounted unit obtains vehicle running information including position, vehicle speed, acceleration, engine rotating speed, throttle position and accelerator pedal position and sends the vehicle running information to an edge server;
step three: at the edge server side, training a suitable network structure by using a NARX neural network, and predicting the possible driving speed within 60s in the future;
step four: judging whether the starting and stopping times can be reduced or not according to the distance between the vehicle and the stop line of the intersection and the phase information of the signal lamp, and smoothly passing through the intersection;
step five: if the vehicle can smoothly pass through the intersection under the condition of single vehicle passing or vehicle following, a suggested vehicle speed area is given, and if the vehicle needs to stop for waiting, warning information is given;
step six: and sending prompt information or warning information to the vehicle-mounted unit, and displaying the vehicle-mounted unit in different forms of a vehicle central control screen, HUD head-up display and voice broadcast.
Further, the road side unit comprises a positioning module, a road side processing module and a 5G communication module which are sequentially connected, wherein the positioning module acquires signal lamp position information and stop line position information; the road side processing module is used for acquiring the phase and the remaining time of the signal lamp and numbering a running vehicle; the 5G communication module is responsible for establishing real-time communication between the vehicle-mounted unit and the edge server, and information interaction between the road side end and the vehicle-mounted end and between the road side end and the edge server is achieved.
Furthermore, the vehicle-mounted unit comprises a positioning module, a vehicle-mounted processing module, a 5G communication module, a display module and a voice unit, wherein the positioning module, the vehicle-mounted processing module, the display module and the voice unit are sequentially connected, and meanwhile, the vehicle-mounted processing module is also connected with the 5G communication module; the positioning module is responsible for positioning the vehicle, providing the position information of the vehicle and used for judging which intersection the vehicle is at and driving to which direction; the vehicle-mounted processing module is used for acquiring vehicle running data and historical running data; the 5G communication module is responsible for real-time communication between the vehicle end and the road side end, and between the vehicle end and the server end, and exchanges required information; the display module is responsible for displaying the phase and the residual time of the signal lamp and the comprehensive processing suggestion or warning of the server side in the vehicle so that a driver can make reasonable driving decision; the voice unit transmits the obtained suggestion or warning to the driver in a voice broadcasting mode.
Further, the edge server comprises a database, a data processing unit and a 5G communication module, wherein the database is connected with the data processing unit and then is connected with the base station and the 5G communication module; the database stores vehicle running data and vehicle numbers sent by a vehicle end and a neural network structure trained by a NARX neural network; the data processing unit is responsible for training the NARX neural network, predicting the vehicle speed and making a reasonable driving suggestion according to the vehicle speed suggestion mathematical model and the vehicle speed prediction result; the 5G communication module is responsible for real-time communication between the server side and the vehicle side and the road side to transmit data.
Further, in the neural network model, the actual numerical value of the sensor in the automobile CAN be analyzed by accessing the CAN bus analyzer into the OBD interface of the automobile, and parameters required by the training network are extracted, including: the method comprises the steps of generating an Excel file in real time according to vehicle speed, vehicle acceleration, engine speed, throttle position and accelerator pedal position, and sending the Excel file to an edge server.
Further, the specific process of the step 3-5 is as follows:
and data acquisition, namely receiving data of relevant parameters sent by the vehicle-mounted unit, storing the vehicle speed data in a row vector y (t) as shown in a formula (3-1), and storing other relevant parameter data in a matrix X (t) as shown in a formula (3-2) in the server.
y(t)＝[y(t-1)，y(t-2)，…，y(t-n)] (3-1)
Screening data through correlation analysis, and selecting parameter data with higher weight value of vehicle speed influence as exogenous input;
data normalization processing according to equation 3-3
Wherein x is n Refers to each element, x, in the constituent row vectors max Refers to the maximum value, x, in each row vector min Refers to the minimum value in each row vector;
sample data is divided into three parts: the sample capacities of the training data, the test data and the verification data are respectively set to be 70%, 15% and are used as input layers for training the neural network;
selecting a delay order, the number of neurons in a hidden layer and an excitation function, training and verifying the neural network, and determining an optimal network structure;
predicting the vehicle speed according to the trained neural network structure, predicting the vehicle speed 60s after the time point, performing inverse normalization on the data, and finally storing the vehicle speed prediction result as a vector y predict(t) And fitting a speed-time curve:
y predict (t)＝[y(t)，y(t+1)，y(t+2)，…，y(t+60)] (3-4)
analyzing and calculating a suggested vehicle speed according to the acquired information such as the signal lamp information and the speed-time curve;
the cycle time allocation of the signal lamps is respectively recorded as: red light T _ red, green light T _ green and yellow light T _ yellow; here, the total signal lamp period T is set to 60s, that is:
T＝T_red+T_green+T_yellow＝60s (3-5)
signal light information sent by the marking road side unit: when the phase is red, it is recorded as sec _ red; when the phase is green, it is recorded as sec _ green; when the phase is yellow, it is recorded as sec _ yellow;
calculating the distance between the vehicle and the stop line as distance _ stop according to the vehicle position information and the stop line position information;
judging whether the intersection can be smoothly passed or not;
when the phase of the signal lamp is red, the red lamp remaining time is sec _ red, the estimated driving distance0 and distance1 are obtained by integrating curves in the intervals of [0, sec _ red ], [0, (sec _ red + T _ green) ] according to the speed-time curve, and if distance _ stop belongs to (distance0 and distance1), the driver can drive smoothly through the intersection; if distance _ stop ═ distance1, the driver can be prompted that the vehicle can be accelerated appropriately; if distance _ stop is less than distance0, the driver is prompted to properly decelerate and send (V _ min, V _ max) to the vehicle-mounted end, and if V _ max is more than V _ limit road speed limit value, the speed limit value is taken as the maximum value to be sent to the vehicle-mounted end; if V _ min > V _ limit, the driver needs to be reminded that the driver has overspeed, slows down, stops before a stop line and waits "
V_min＝distance/(sec_red+T_green) (3-6)
V_max＝distance/sec_red (3-7)
When the phase of the signal lamp is green, obtaining the green lamp remaining time as sec _ green, integrating the curve in the interval of [0, sec _ green ] according to the speed-time curve to obtain the predicted driving distance1, and if the distance1 is greater than distance _ stop, the driver can smoothly pass through the intersection by normal driving; if distance is less than distance _ stop, the driver can be prompted to properly accelerate and send (V _ min, V _ max) to the vehicle-mounted end. If V _ min is more than V _ limit, the driver needs to be reminded that you have overspeed, please slow down, and stop for waiting before the stop line "
V_min＝distance_stop/sce_green (3-8)
V_max＝V_limit (3-9)
When the phase of the signal lamp is yellow, the remaining time of the yellow lamp is sec _ yellow, and the 'please slow down and slow down, need to wait for passing before the stop line' is sent to the vehicle-mounted end "
Following cruise driving: when the vehicle runs on a certain road section, n vehicles in front of the vehicle do not cross the stop line, at the moment, the vehicle is numbered n, the current speed of the vehicle numbered n is taken as V _ max, and information of ' suggestion of driving speed [ V _ min, V _ max ] is sent to a vehicle end, the distance is noticed, and the following driving is kept ', wherein V _ min is the numerical value obtained in the step, if V _ min ═ V _ max, the vehicle end needs to be sent with the distance noticed, the following driving is kept, the front crossing is about to stop and wait ', and the vehicle end numbered n +1 is sent with the information of ' noticing the parking brake of the vehicle in front '
1. Road side unit
(1) Carrying a positioning module and a 5G communication unit at a signal lamp end of the intersection, wherein the positioning module provides position information of a signal lamp and provides guidance for a user vehicle of the signal lamp; the 5G communication unit is used for realizing real-time communication between the signal lamp and the vehicle-mounted end and the edge server, transmitting the phase information and the remaining time of the signal lamp to the vehicle-mounted end and the edge server end and providing data support;
(2) installing a positioning module and a communication unit at one side of the corresponding stop line, transmitting the position information of the stop line to an edge server end, and providing data support;
(3) the roadside processing module is a running vehicle number, and the number is started from the vehicle closest to the stop line along the reverse direction of the running direction: 0. 1, 2 …, wherein the vehicle with number 0 is set as a head vehicle, when the head vehicle crosses the stop line, the number is released, and the number of the rear vehicle is decreased.
2. Vehicle-mounted unit
(1) Similarly, a positioning module and a 5G communication unit are carried, signal lamp information transmitted by the roadside end is received, and the phase and the remaining time of the signal lamp are displayed on a glass screen in the form of images by utilizing the holographic projection technology;
(2) collecting historical data of relevant parameters required by vehicle speed prediction, wherein the relevant parameters comprise vehicle speed, acceleration, engine speed, accelerator, brake pedal and gear lever positions;
training of neural network models requires a large amount of data, and therefore, a large amount of data needs to be collected. The vehicle-mounted self-diagnosis system reads the vehicle CAN bus information through an OBD interface connected to the vehicle. The CAN bus is an international standardized serial communication protocol, is developed by German Bosch company for automobiles, has the characteristics of strong real-time and reliable data transmission among network nodes, and becomes an automobile network data transmission standard protocol. The CAN bus analyzer is connected to an OBD interface of a vehicle, so that the actual numerical value of the sensor in the internal part of the vehicle CAN be analyzed, and parameters required by a training network are extracted, wherein the parameters comprise: the method comprises the steps of generating an Excel file in real time according to vehicle speed, vehicle acceleration, engine speed, throttle position and accelerator pedal position, and sending the Excel file to an edge server.
(3) And sending the positioning information of the vehicle to an edge server.
(4) According to the Vehicle number, V2V (Vehicle to Vehicle) information intercommunication is realized through a D2D (Device to Device) communication technology, and information of a front Vehicle is received and sent to a rear Vehicle;
(5) the vehicle speed suggestion information sent by the edge server is received and projected on a glass screen in a text mode, and meanwhile, the vehicle speed suggestion information is prompted to a driver through a voice broadcaster.
3. Edge server
(1) Receiving signal lamp information (including phase and remaining time second), signal lamp position information and stop line position information sent by a road side unit;
(2) receiving historical data of vehicle speed prediction related parameters and the current position of the vehicle sent by a vehicle-mounted unit, and storing the historical data and the current position of the vehicle in a server;
(3) the artificial neural network has excellent nonlinear fitting capability and great advantages in processing nonlinear and time-varying problems, and various neural networks are applied to analysis and prediction of time series at present.
Due to the strong computing power of the edge server, the vehicle speed prediction and vehicle speed suggestion decision part is put in the server for operation. Compared with a nonlinear autoregressive neural Network (NAR), the nonlinear autoregressive exogenous neural Network (NARX) increases the consideration of relevant influence factors, so that the data fitting is more accurate and the efficiency is higher. The NARX basic structure is divided into an input layer, an input delay layer, a hidden layer and an output layer, and the network structure has memory capacity and can return the output of the neural network as the input through external feedback to the neural network to form a closed-loop system.
The invention has the technical effects that: and judging whether the vehicle can smoothly pass through the intersection by reducing the starting and stopping times or not according to the actual predicted vehicle speed of the vehicle, the actual vehicle speed of the front vehicle, the distance between the vehicle and the intersection and the remaining time of a signal lamp, and broadcasting the recommended vehicle speed for economic driving to a driver. The model has high prediction precision and high transmission efficiency, and can be used for urban road vehicles. The invention can effectively improve the driving safety and improve the traffic jam and the environmental pollution.
Drawings
FIG. 1 is a schematic diagram of a technical scheme according to the present invention;
FIG. 2 is a schematic diagram of information interaction among a road side unit, a vehicle-mounted unit and an edge server according to the present invention;
FIG. 3 is a schematic diagram of the module composition of the on-board unit according to the present invention;
FIG. 4 is a schematic diagram of the roadside unit module according to the present invention;
FIG. 5 is a block diagram of an edge server according to the present invention;
fig. 6 is a diagram of a NARX neural network structure according to the present invention.
Detailed Description
In the system, under the 5G network environment, the most economical driving scheme is provided for a driver to stably pass through a crossing Road through information exchange and processing among a Road Side Unit (RSU), an On Board Unit (OBU) and an Edge server (Edge computing).
The technical route of the embodiment of the invention is clearly and completely described below with reference to the accompanying fig. 1-6, and the technical principle of the invention is explained. As shown in figure 1 of the drawings, in which,
the method comprises the following steps: the method comprises the steps that a Road Side Unit (RSU) acquires signal lamp information and sends the signal lamp information to a side cloud server (Edge) and an on-board unit (OBU), and meanwhile position information of a road junction stop line in the direction is acquired;
step two: the method comprises the steps that a vehicle-mounted unit obtains vehicle running information including position, vehicle speed, acceleration, engine rotating speed, throttle position and accelerator pedal position and sends the vehicle running information to an edge server;
step three: at the edge server side, training a suitable network structure by using a NARX neural network, and predicting the possible driving speed within 60s in the future;
step four: judging whether the starting and stopping times can be reduced or not according to the distance between the vehicle and the stop line of the intersection and the phase information of the signal lamp, and smoothly passing through the intersection;
step five: if the vehicle can smoothly pass through the intersection under the condition of single vehicle passing or vehicle following, a suggested vehicle speed area is given, and if the vehicle needs to stop for waiting, warning information is given;
step six: and sending the prompt information or the warning information to the vehicle-mounted unit, and displaying the vehicle in different forms such as a vehicle central control screen, HUD head-up display, voice broadcast and the like.
The system mainly comprises three units: roadside units, on-board units, and edge servers. The road side unit provides signal lamp related information and stop line information; the vehicle-mounted unit provides vehicle running data and displays the obtained signal lamp information and the information comprehensive processing result in the vehicle; the edge server is used as a calculation unit for performing information comprehensive processing.
The road side unit comprises a positioning module, a road side processing module and a 5G communication module. The positioning module acquires signal lamp position information and stop line position information; the roadside processing module is responsible for acquiring the phase and the remaining time of the signal lamp and numbering the running vehicles; the 5G communication module is responsible for establishing real-time communication between the road side end and the vehicle-mounted unit and between the road side end and the edge server, and information interaction between the road side end and the vehicle-mounted end and between the road side end and the edge server is achieved.
The vehicle-mounted unit comprises a positioning module, a vehicle-mounted processing module, a 5G communication module, a display module and a voice unit. The positioning module is responsible for positioning the vehicle, providing the position information of the vehicle and used for judging which intersection the vehicle is at and driving to which direction; the vehicle-mounted processing module is used for acquiring vehicle running data and historical running data; the 5G communication module is responsible for real-time communication between the vehicle end and the road side end, and between the vehicle end and the server end, and exchanges required information; the display module is responsible for displaying the phase and the remaining time of the signal lamp and the comprehensive processing suggestion or warning of the server side in the vehicle so that a driver can make a reasonable driving decision; the voice unit transmits the obtained suggestion or warning to the driver in a voice broadcasting mode.
The edge server comprises a database, a data processing unit and a 5G communication module. The database stores vehicle running data and vehicle numbers sent by a vehicle end and a neural network structure trained by a NARX neural network; the data processing unit is responsible for training the NARX neural network, predicting the vehicle speed and making a reasonable driving suggestion according to the vehicle speed suggestion mathematical model and the vehicle speed prediction result; the 5G communication module is responsible for real-time communication between the server side and the vehicle side and the road side to transmit data.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an illustrative embodiment," "an example," "a specific example," or "some examples" or the like mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
Claims (5)
1. A method for passing through an intersection based on a NARX neural network and a C-V2X technology is characterized by comprising the following steps:
the method comprises the following steps: the road side unit RSU acquires signal lamp information and sends the signal lamp information to the Edge server Edge and the on board unit OBU, and meanwhile, the road side unit RSU acquires position information of a road junction stop line in the forward driving direction of a vehicle;
step two: the method comprises the steps that a vehicle-mounted unit obtains vehicle running information including position, vehicle speed, acceleration, engine rotating speed, throttle position and accelerator pedal position and sends the vehicle running information to an edge server;
step three: at the edge server side, training a suitable network structure by using a NARX neural network, and predicting the possible driving speed within 60s in the future;
step four: judging whether the number of start-stop times can be reduced or not according to the distance between the vehicle and the stop line of the intersection and the phase information of the signal lamp, and smoothly passing through the intersection;
step five: if the vehicle can smoothly pass through the intersection under the condition of single vehicle passing or vehicle following, a suggested vehicle speed interval is given, and if the vehicle needs to stop for waiting, warning information is given;
step six: sending the prompt information or the warning information to a vehicle-mounted unit, and displaying the prompt information or the warning information in different forms of an automobile central control screen, HUD head-up display and voice broadcast;
the specific process of the third step to the fifth step is as follows:
data acquisition, namely receiving data of relevant parameters sent by the vehicle-mounted unit, storing vehicle speed data in a row vector y (t) as shown in a formula (3-1), storing other relevant parameter data in a matrix X (t) as shown in a formula (3-2) in a server,
y(t)＝[y(t-1)，y(t-2)，…，y(t-n)] (3-1)
screening data through correlation analysis, and selecting parameter data with higher weight value of vehicle speed influence as exogenous input;
data normalization processing according to equation 3-3
Wherein x is n Refers to each element, x, in the constituent row vectors max Refers to the maximum value, x, in each row vector min Refers to the minimum value in each row vector;
sample data is divided into three parts: the sample capacities of training data, testing data and verification data are respectively set to be 70%, 15% and 15%, and the sample capacities are used as input layers for training a neural network;
selecting a delay order, the number of neurons in a hidden layer and an excitation function, training and verifying a neural network, and determining an optimal network structure;
predicting the vehicle speed according to the trained neural network structure, predicting the vehicle speed 60s after the time point, performing inverse normalization on the data, and finally storing the vehicle speed prediction result as a vector y predict (t) and fitting a speed-time curve:
y predict (t)＝[y(t)，y(t+1)，y(t+2)，…，y(t+60)] (3-4)
analyzing and calculating a suggested vehicle speed according to the acquired signal lamp information and the speed-time curve information;
the cycle time allocation of the signal lamps is respectively recorded as: red light T _ red, green light T _ green and yellow light T _ yellow; here, the total signal lamp period T is set to 60s, that is:
r＝T_red+Tgreen+T_yellow＝60s (3-5)
signal light information sent by the marking road side unit: when the phase is red, it is recorded as sec _ red; when the phase is green, recording as sec _ green; when the phase is yellow, it is recorded as sec _ yellow;
calculating the distance between the vehicle and the stop line as distance _ stop according to the vehicle position information and the stop line position information;
judging whether the intersection can be smoothly passed or not;
a) when the phase of the signal lamp is red, the red lamp remaining time is sec _ red, the estimated driving distance0 and distance1 are obtained by integrating curves in the intervals of [0, sec _ red ], [0, (sec _ red + T _ green) ] according to the speed-time curve, and if distance _ stop belongs to (distance0 and distance1), the driver can drive smoothly through the intersection; if distance _ stop > is distance1, the driver can be prompted to drive with proper acceleration; if distance _ stop < > distance0, the driver can be prompted to properly decelerate and drive, and simultaneously (V _ min, V _ max) is sent to the vehicle-mounted end, and if V _ max > V _ limit road speed limit value, the speed limit value is used as the maximum value to be sent to the vehicle-mounted end; if V _ min > V _ limit, the driver needs to be reminded that you have overspeed, please slow down, and stop for waiting before the stop line "
b) When the phase of the signal lamp is green, obtaining the green lamp remaining time as sec _ green, integrating the curve in the interval of [0, sec _ green ] according to the speed-time curve to obtain the predicted driving distance1, and if the distance1 is greater than distance _ stop, the driver can smoothly pass through the intersection by normal driving; if distance1< (distance _ stop), the driver can be prompted to properly accelerate and send (V _ min, V _ max) to the vehicle-mounted end, if V _ min > V _ limit, the driver needs to be reminded to 'you have overspeed, please decelerate and walk slowly, and stop and wait before the stop line'
V_min＝distance_stop/sce_green (3-8)
V_max＝V_limit (3-9)
c) When the phase of the signal lamp is yellow, the remaining time of the yellow lamp is sec _ yellow, and the 'please slow down and wait for passing' before the stop line is required to be sent to the vehicle-mounted end "
Following cruise driving: when the vehicle runs on a certain road section, u vehicles in front of the vehicle do not cross the stop line, at this moment, the vehicle is numbered u, the current speed of the vehicle numbered u is taken as V _ max, and information "suggest running speed [ V _ min, V _ max ] is sent to the vehicle end, the distance is noticed, and the following driving is kept", wherein V _ min is a numerical value obtained in the steps, if V _ min > is equal to V _ max, the vehicle end needs to be sent with the distance noticed, the following driving is kept, the front road junction is about to stop and wait ", and information" notice the front vehicle parking brake "is sent to the vehicle end numbered u + 1.
2. The method for passing through the intersection based on the NARX neural network and the C-V2X technology, according to claim 1, wherein the road side unit comprises a positioning module, a road side processing module and a 5G communication module which are connected in sequence, and the positioning module acquires signal lamp position information and stop line position information; the road side processing module is used for acquiring the phase and the remaining time of the signal lamp and numbering a running vehicle; and the 5G communication module is responsible for establishing real-time communication between the vehicle-mounted unit and the edge server, and realizing information interaction between the road side end and the vehicle-mounted end as well as the server side.
3. The crossing traffic method based on NARX neural network and C-V2X technology of claim 1, wherein said vehicle-mounted unit comprises a positioning module, a vehicle-mounted processing module, a 5G communication module, a display module, and a voice unit, said positioning module, said vehicle-mounted processing module, said display module, and said voice unit are connected in turn, and at the same time, said vehicle-mounted processing module is further connected with said 5G communication module; the positioning module is responsible for positioning the vehicle, providing the position information of the vehicle and used for judging which intersection the vehicle is at and driving to which direction; the vehicle-mounted processing module is used for acquiring vehicle running data and historical running data; the 5G communication module is responsible for real-time communication between the vehicle end and the road side end, and between the vehicle end and the server end, and required information is exchanged; the display module is responsible for displaying the phase and the residual time of the signal lamp and the comprehensive processing suggestion or warning of the server side in the vehicle so that a driver can make reasonable driving decision; the voice unit transmits the obtained suggestion or warning to the driver in a voice broadcasting mode.
4. The method for passing through an intersection based on NARX neural network and C-V2X technology of claim 1, wherein the edge server comprises a database, a data processing unit, and a 5G communication module, the database and the data processing unit are connected to each other and then connected to a base station and the 5G communication module; the database stores vehicle running data and vehicle numbers sent by a vehicle end and a neural network structure trained by a NARX neural network; the data processing unit is responsible for training the NARX neural network, predicting the vehicle speed and making a reasonable driving suggestion according to the vehicle speed suggestion mathematical model and the vehicle speed prediction result; the 5G communication module is responsible for real-time communication between the server side and the vehicle side and the road side to transmit data.
5. The method for crossing traffic based on NARX neural network and C-V2X technology of claim 1, wherein in the neural network model, the CAN bus analyzer is connected to the OBD interface of the vehicle to analyze the actual values of the sensors in the interior of the vehicle and extract the parameters required by the training network, comprising: the method comprises the steps of generating an Excel file in real time according to vehicle speed, vehicle acceleration, engine speed, throttle position and accelerator pedal position, and sending the Excel file to an edge server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110559099.9A CN113450564B (en) | 2021-05-21 | 2021-05-21 | Intersection passing method based on NARX neural network and C-V2X technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110559099.9A CN113450564B (en) | 2021-05-21 | 2021-05-21 | Intersection passing method based on NARX neural network and C-V2X technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113450564A CN113450564A (en) | 2021-09-28 |
| CN113450564B true CN113450564B (en) | 2022-08-23 |
Family
ID=77809990
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110559099.9A Active CN113450564B (en) | 2021-05-21 | 2021-05-21 | Intersection passing method based on NARX neural network and C-V2X technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113450564B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113611122B (en) * | 2021-10-09 | 2022-01-14 | 北京理工大学深圳汽车研究院(电动车辆国家工程实验室深圳研究院) | Vehicle speed guiding method, device and equipment |
| CN116476828A (en) * | 2023-03-08 | 2023-07-25 | 华中科技大学 | Vehicle speed planning method and system considering influence of front vehicles and traffic light information |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100864451B1 (en) * | 2007-04-19 | 2008-10-22 | 학교법인 영광학원 | Vehicle detection system based vehicle to roadside communication and traffic signal control method using the vehicle detection system in crossroad |
| CN105590144A (en) * | 2015-12-30 | 2016-05-18 | 国电联合动力技术有限公司 | Wind speed prediction method and apparatus based on NARX neural network |
| CN109461320A (en) * | 2018-12-20 | 2019-03-12 | 清华大学苏州汽车研究院（吴江） | Intersection speed planing method based on car networking |
| CN110444015A (en) * | 2019-07-08 | 2019-11-12 | 清华大学 | Intelligent network based on no signal crossroad subregion joins car speed decision-making technique |
| CN111508254A (en) * | 2020-03-17 | 2020-08-07 | 宁波吉利汽车研究开发有限公司 | Control method and control system for efficiently passing through traffic intersection |
| CN112289059A (en) * | 2020-10-22 | 2021-01-29 | 中电智能技术南京有限公司 | Vehicle-road cooperative road traffic system |
| US10930146B1 (en) * | 2019-12-23 | 2021-02-23 | Continental Automotive Systems, Inc. | Traffic control system and method for operating same |
| CN112466115A (en) * | 2019-12-30 | 2021-03-09 | 北京航空航天大学 | Bus intersection priority passing control system and method based on edge calculation |
| CN112509343A (en) * | 2020-10-26 | 2021-03-16 | 长沙理工大学 | Roadside type guiding method and system based on vehicle-road cooperation |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021524087A (en) * | 2018-05-16 | 2021-09-09 | ノートラフィック リミテッド | Systems and methods for using V2X and sensor data |
| CN111080018B (en) * | 2019-12-20 | 2024-02-20 | 南京航空航天大学 | Intelligent network-connected automobile speed prediction method based on road traffic environment |
| CN112562326A (en) * | 2020-11-26 | 2021-03-26 | 上汽通用五菱汽车股份有限公司 | Vehicle speed guiding method, server and readable storage medium |
-
2021
- 2021-05-21 CN CN202110559099.9A patent/CN113450564B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100864451B1 (en) * | 2007-04-19 | 2008-10-22 | 학교법인 영광학원 | Vehicle detection system based vehicle to roadside communication and traffic signal control method using the vehicle detection system in crossroad |
| CN105590144A (en) * | 2015-12-30 | 2016-05-18 | 国电联合动力技术有限公司 | Wind speed prediction method and apparatus based on NARX neural network |
| CN109461320A (en) * | 2018-12-20 | 2019-03-12 | 清华大学苏州汽车研究院（吴江） | Intersection speed planing method based on car networking |
| CN110444015A (en) * | 2019-07-08 | 2019-11-12 | 清华大学 | Intelligent network based on no signal crossroad subregion joins car speed decision-making technique |
| US10930146B1 (en) * | 2019-12-23 | 2021-02-23 | Continental Automotive Systems, Inc. | Traffic control system and method for operating same |
| CN112466115A (en) * | 2019-12-30 | 2021-03-09 | 北京航空航天大学 | Bus intersection priority passing control system and method based on edge calculation |
| CN111508254A (en) * | 2020-03-17 | 2020-08-07 | 宁波吉利汽车研究开发有限公司 | Control method and control system for efficiently passing through traffic intersection |
| CN112289059A (en) * | 2020-10-22 | 2021-01-29 | 中电智能技术南京有限公司 | Vehicle-road cooperative road traffic system |
| CN112509343A (en) * | 2020-10-26 | 2021-03-16 | 长沙理工大学 | Roadside type guiding method and system based on vehicle-road cooperation |
Non-Patent Citations (3)
| Title |
|---|
| A Generative Learning Approach for Spatio-temporal Modeling in Connected Vehicular Network;Rong Xia Et al.;《ICC 2020 - 2020 IEEE International Conference on Communications (ICC)》;20200727;第1-6页 * |
| In-Vehicle Traffic Light based on Local Area Network;Yunshun Zhang Et al.;《2021 IEEE International Conference on Mechatronics (ICM)》;20210330;第1-6页 * |
| 车路协同环境下信号交叉口速度引导策略;刘欢等;《重庆交通大学学报（自然科学版）》;20191231;第38卷(第12期);第8-17页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113450564A (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ersal et al. | Connected and automated road vehicles: state of the art and future challenges | |
| Sciarretta et al. | Energy-efficient driving of road vehicles | |
| Araújo et al. | Driving coach: A smartphone application to evaluate driving efficient patterns | |
| d'Orey et al. | ITS for sustainable mobility: A survey on applications and impact assessment tools | |
| CN102819954B (en) | Traffic region dynamic map monitoring and predicating system | |
| CN109242251A (en) | Vehicular behavior safety detecting method, device, equipment and storage medium | |
| CN113450564B (en) | Intersection passing method based on NARX neural network and C-V2X technology | |
| CN107844613A (en) | Electric car continuation of the journey management method based on data analysis | |
| Pandazis et al. | Study of Intelligent Transport Systems for reducing CO2 emissions for passenger cars | |
| CN115270475A (en) | Vehicle road cloud bus priority online simulation platform based on 5G and CV2X | |
| Samaras et al. | Development of a methodology and tool to evaluate the impact of ICT measures on road transport emissions | |
| Fang et al. | Change in microscopic traffic simulation practice with respect to the emerging automated driving technology | |
| Ruan et al. | A modularized electric vehicle model-in-the-loop simulation for transportation electrification modeling and analysis | |
| Du et al. | Impacts of vehicle-to-everything enabled applications: literature review of existing studies | |
| Pan et al. | Research on multi-lane energy-saving driving strategy of connected electric vehicle based on vehicle speed prediction | |
| Boccolini | Development of a speed profile prediction algorithm based on navigation data for energy management optimization | |
| CN117238131B (en) | Traffic flow characteristic analysis method in Internet of vehicles environment | |
| Manzie et al. | A comparison of fuel consumption between hybrid and intelligent vehicles during urban driving | |
| Qi | Next generation intelligent driver-vehicle-infrastructure cooperative system for energy efficient driving in connected vehicle environment | |
| Saint Pierre et al. | D43. 1: Eco Driving in the Real-world: Behavioural, Environmental and Safety Impacts | |
| Hu | Simulated Influence of Electric Vehicles on Traffic Flow and Analysis of Energy Consumption Using the Example of the City of Duisburg | |
| Marinelli et al. | Assessing the Impact of Eco-ADAS Systems on Vehicle Fuel Consumption and Emissions at Signalised Intersections | |
| Barth et al. | ECO-ITS: intelligent transportation system applications to improve environmental performance | |
| Chen et al. | Research on evaluation system of simulation test based on driving scenarios for L1-L2 autonomous driving | |
| Chen et al. | On the key technologies of internet of vehicles and its innovative application of integration with 5G |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |