CN115476857B - Method and device for controlling gear shifting based on natural driving data - Google Patents
Method and device for controlling gear shifting based on natural driving data Download PDFInfo
- Publication number
- CN115476857B CN115476857B CN202211033655.XA CN202211033655A CN115476857B CN 115476857 B CN115476857 B CN 115476857B CN 202211033655 A CN202211033655 A CN 202211033655A CN 115476857 B CN115476857 B CN 115476857B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- road
- information
- lane
- driving data
- 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
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
- B60W30/18—Propelling the vehicle
- B60W30/19—Improvement of gear change, e.g. by synchronisation or smoothing gear shift
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/02—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to ambient conditions
- B60W40/06—Road conditions
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
- B60W40/105—Speed
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/08—Interaction between the driver and the control system
- B60W50/14—Means for informing the driver, warning the driver or prompting a driver intervention
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/08—Interaction between the driver and the control system
- B60W50/14—Means for informing the driver, warning the driver or prompting a driver intervention
- B60W2050/146—Display means
-
- 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
- Y02T10/80—Technologies aiming to reduce greenhouse gasses emissions common to all road transportation technologies
- Y02T10/84—Data processing systems or methods, management, administration
Abstract
The invention discloses a method for controlling gear shifting based on natural driving data, which comprises the following steps: obtaining highway information in a specified range of a road where a running vehicle is located, wherein the highway information comprises a highway type and a lane type; generating road network information according to the road information, dynamically matching the optimal driving behavior from the optimal driving data set of the fuel economy, and generating gear shifting decision information, wherein the decision information comprises optimal gear shifting steps; outputting the gear shifting decision information to a driving end; wherein, the obtaining the lane type includes: and extracting a lane line image, performing blurring processing on the lane line image, and judging and outputting lane types, wherein the lane types comprise straight lanes and curves. According to the technical scheme, the artificial intelligence simulation control can be performed on the gear control of the commercial vehicle so as to realize the selection of the optimal gear control at the optimal time, thereby enabling the commercial vehicle to have the optimal fuel economy.
Description
Technical Field
The invention relates to the technical field of intelligent driving, in particular to a method and a device for controlling gear shifting based on natural driving data.
Background
The utility vehicle is widely applied to the fields of ports, industrial parks, cargo transportation and the like for transporting personnel and cargo, the current gear shifting rule of the automatic transmission utility vehicle is to determine the optimal gear shifting time according to three signals of a throttle opening degree, an accelerator pedal stroke and an engine rotating speed when the vehicle runs, and when the values of the three signals change to a certain value range, a control unit of the automatic transmission can be changed into a new gear. However, the changing data range of the three signals does not give consideration to road conditions and fuel consumption conditions.
The automatic shift commercial vehicle can increase the comfort of a driver, but compared with the manual shift commercial vehicle, the automatic shift commercial vehicle cannot select the optimal shift, has the defects of high fuel consumption, poor operability and the like, and therefore, a technical scheme is needed, and the optimal shift time can be grasped to shift the shift into the optimal shift according to the current road information and the environment information, so that the optimal performance and efficiency are realized.
Disclosure of Invention
In a first aspect, to achieve the above object, the present application provides a method for controlling gear shifting based on natural driving data, including the steps of:
obtaining highway information in a specified range of a road where a running vehicle is located, wherein the highway information comprises a highway type, a surrounding environment and a lane type;
generating road network information according to the road information, dynamically matching the optimal driving behavior from the optimal driving data set of the fuel economy, and generating gear shifting decision information, wherein the decision information comprises optimal gear shifting steps;
outputting the gear shifting decision information to a driving end;
wherein, the obtaining the lane type includes: and extracting a lane line image, performing blurring processing on the lane line image, and judging and outputting lane types, wherein the lane types comprise straight lanes and curves.
Further, before the optimal driving behavior is dynamically matched from the optimal driving data set of the fuel economy, acquiring a natural driving data set, and extracting the optimal driving data set of the fuel economy from the natural driving data set;
collecting a natural driving data set refers to: natural driving data of the manual-gear commercial vehicle are collected, wherein the natural driving data comprise vehicle appearance, gear system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed, steering wheel rotation angle and the like.
Extracting a fuel economy optimal driving data set from the natural driving data set includes the steps of:
extracting a result data set with accurate vehicle type identification from the natural driving data set;
performing vehicle type classification and road type classification from the result dataset;
and screening the fuel economy optimal driving data set according to the vehicle type and the road type, wherein the screening standard of the fuel economy optimal driving data set is that the fuel consumption is minimum.
Further, extracting a vehicle type identification accurate result data set from the natural driving data set includes the steps of:
identifying the appearance characteristics of the vehicle from the natural driving data set, and extracting a characteristic result set with accurate vehicle type identification;
outputting a vehicle type result according to the vehicle appearance characteristics by adopting a cluster analysis method;
extracting the appearance characteristics of the vehicle comprises preliminary screening and normalization processing;
preliminary screening means: screening the recognition results of the image data in the natural driving data set by adopting the loss function score to generate a characteristic result set, wherein the specific algorithm is as follows:
wherein L is i For loss function score, s j Is in errorMisidentification of class score, s yi A class score for correct discrimination;
if L i <=0, i.e. the recognition result is accurate, and the corresponding image data is added into the feature result set;
normalization processing means: and calculating the recognition accuracy probability of the image data in the feature result set, and clearing the corresponding data of the feature result set with low probability value according to the recognition accuracy probability.
Further, a blurring processing finger is executed on the lane line image;
performing binarization processing on the image where the lane line is located to generate a binarized image;
optimizing the binarized image;
searching black-and-white jumping points of the binarized image to obtain the boundary of the lane line image;
and judging the type of the lane according to the boundary of the image.
Wherein optimizing the binarized image refers to:
judging whether the lane line environment condition needs to execute the line supplementing operation or not; the environmental conditions in which a patch cord needs to be performed include: lane line missing, crossroads, ramps and intersections;
if the environmental condition is an intersection and the vehicle moves straight, calculating a slope to generate an optimization basis; if the environmental condition is a ramp, a fork, a crossroad left turn or right turn, calculating curvature to generate an optimization basis; if the environmental condition is lane line missing, calculating a linear track to generate an optimization basis;
further, a line supplementing operation is performed on the binarized image according to the optimization basis.
In another aspect, the present application provides an apparatus for controlling gear shifting based on natural driving data, comprising:
driving behavior acquisition module: the automatic control system is used for acquiring natural driving data of the manual-gear commercial vehicle, providing a driving data set with optimal fuel economy, wherein the natural driving data comprises vehicle appearance, gear system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed, steering wheel rotation angle and other information.
An information perception module: the method comprises the steps of interacting with a third party system, and obtaining highway information in a specified range of a road where a driving vehicle is located, wherein the highway information comprises a highway type and a lane type, and the lane information comprises a straight road and a curve;
and a matching decision module: the method comprises the steps of generating road network information according to road information, matching optimal driving behaviors from a fuel economy optimal driving data set, and generating gear shifting decision information;
and the man-machine interaction module is used for: the gear shifting control method comprises the steps of outputting gear shifting decision information, wherein the decision information comprises an optimal gear shifting position;
the information sensing module comprises a road information processing unit and is used for judging the type of the lane from the lane line image, and the information sensing module comprises the steps of performing blurring processing on the lane line image, judging and outputting the type of the lane.
Further, the driving behavior acquisition module includes:
natural driving data extraction unit: the system is used for collecting natural driving data of the manual-shift commercial vehicle;
neural network extraction unit: the method comprises the steps of extracting a result data set with accurate vehicle type identification from a natural driving data set for preliminary screening;
normalization processing unit: the method comprises the steps of calculating the recognition accuracy probability of a primarily screened result data set, and extracting a data value with the highest probability value;
clustering unit: for vehicle type classification and road type classification from the result dataset;
a heap sort unit: the method is used for screening the optimal driving data set of the fuel economy according to the type of the vehicle and the type of the road by using a heap sorting algorithm.
Further, the matching decision module includes:
a road network reorganization unit: the method comprises the steps of updating and reorganizing a road network to generate a dynamic running environment, wherein the dynamic running environment comprises a road type and a lane type;
matching unit: the driving behavior matching method is used for matching the dynamic road type and the lane type with the vehicle type of the control vehicle, and the optimal driving data set of the fuel economy accords with the driving behavior of the current driving environment;
decision information generation unit: the method is used for deciding the corresponding throttle opening, the corresponding vehicle speed, the corresponding engine torque and the corresponding output optimal gear according to the driving behavior conforming to the current driving environment.
Further, the road information processing unit further includes an image optimization sub-module: when the blurring processing is performed on the lane line image, the incomplete image is complemented.
According to the invention, based on the manual gear natural driving data of the commercial vehicle and in combination with the road running environment, the gear control of the commercial vehicle can be simulated by the artificial intelligence, so that the gear control can be carried out at the best opportunity, and the commercial vehicle has the best fuel economy.
Drawings
FIG. 1 is a method step diagram for controlling gear shifting based on natural driving data provided in accordance with an embodiment of the present invention;
FIG. 2 is a logic flow diagram of a method for controlling gear shifting based on natural driving data provided in accordance with an embodiment of the present invention;
FIG. 3 is a logic flow diagram of processing road information in a method provided in accordance with an embodiment of the present invention;
fig. 4 is a block diagram of an apparatus for controlling gear shifting based on natural driving data according to an embodiment of the present invention.
Detailed Description
According to the technical scheme, based on the manual gear natural driving data of the commercial vehicle, the road running environment is combined, the gear control of the commercial vehicle is subjected to artificial intelligence simulation control, so that the optimal gear control is selected at the optimal opportunity, and the commercial vehicle has optimal fuel economy.
The method of the present invention for controlling gear shifting based on natural driving data is described in detail below with reference to the accompanying drawings, and FIG. 1 is a specific method step diagram, as shown in the drawings, comprising the steps of:
first, the solution of the present application is a solution based on natural driving data, and therefore, it is necessary to prepare a natural driving data set before performing a control shift. The natural driving data set may be acquired by a third party, or may be generated by collecting natural driving data of a manual-shift commercial vehicle as shown in step S100 of the present application, and the fuel economy optimal driving data set may be extracted from the natural driving data set.
The natural driving data comprise vehicle appearance, gear system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed, steering wheel rotation angle and other information.
The natural driving data are unfolded in a natural driving mode, a driver drives the vehicle after the small embedded data acquisition equipment is installed on the vehicle, and the acquisition process driver presents various data in a natural operation state under the condition that the data acquisition matters are not known. The acquisition equipment mainly comprises a vehicle sensor, a special automobile CAN bus data acquisition equipment (CanOe), a vehicle-mounted video equipment and the like. The CAN bus data acquisition equipment is powered by vehicle-mounted 12V voltage, and four data acquisition channels are respectively connected to an OBD interface of the manual-gear commercial vehicle after debugging, and a CAN high/low interface. The device is electrified when the commercial vehicle is started, a plurality of operation data of the manual-gear commercial vehicle and signal data of vehicle sensors are recorded in real time, and the manual-gear commercial vehicle is automatically stored.
The natural driving data set in the step is an information cluster containing various driving information, driving data accurately identifying the type of a commercial vehicle and the type of a road are required to be extracted from the natural driving data set, the data set under the specified condition is extracted from the natural driving data set as an optimal driving data set, and a strategy for controlling gear shifting is determined in a data reference mode. The process of extracting the optimal driving dataset from the natural driving dataset is as follows:
step S101: identifying vehicle appearance characteristics from a natural driving data set through a neural network algorithm to extract a result data set with accurate vehicle type identification, in the specific process, acquiring a recognition score value of a percentile of a recognition image and the accuracy of the recognition score value through a neural network model, and carrying out preliminary screening on a recognition result according to the recognition score value and the accuracy by adopting a loss function, wherein the loss function is defined as follows:
wherein L is i Score s as a loss function j For error discrimination class score, s yi Class scores are correctly determined.
If L i The method comprises the steps that < = 0, namely the identification result is accurate, and relevant data corresponding to image data are added into a characteristic result set;
for example, a photograph of a semi-trailer is used to identify a score of 95 for the semi-trailer, a score of 80 for the passenger car, a score of 85 for the truck, and a score defined according to a loss function
L i ＝max(0，80-95+1)+max(0，85-95+1)＝max(0，-14)+max(0，-9)＝0，
If the result value of the loss function is 0, the identification result is correct, and the data corresponding to the photo can be subjected to a characteristic result set;
when a photograph of a semi-trailer is used for identification, it identifies a score of 90 for the semi-trailer, 85 for the passenger car, 92 for the truck, defined according to the loss function
L i ＝max(0，85-90+1)+max(0，92-90+1)＝max(0，-4)+max(0，3)＝3，
The result value of the loss function is 3, the identification result is inaccurate, and the corresponding data of the picture does not enter the characteristic result set.
Next, the image recognition accuracy of the data in the feature result set, that is, the data set with the loss function result value of < =0, is described with probability, that is, normalization processing, and the corresponding data in the feature result set with low probability is cleared according to the recognition accuracy probability.
First, the score value x of each image is calculated 1 ，x 2 ，x 3 ......x n The score of (2) is amplified by taking the natural logarithm e as a base, and the score value is an index. The score value of each amplified image isAccording to the normalization formulaThe probability of each image is calculated. The feature data corresponding to the image with the largest probability value is the result data of extracting the appearance feature of the driving vehicle, and the data corresponding to the image with the small probability value is cleared from the result data set.
The result data set screened in step S101 is processed in the next step.
Step S102: vehicle type classification and road type classification;
in the step, the result data obtained in the step S101 are classified according to the types of commercial vehicles (commercial vehicles are divided into passenger cars, trucks and semi-trailer traction vehicles) by adopting a cluster analysis method, so that corresponding natural driving data sets of three commercial vehicles can be obtained;
and then dividing the natural driving data set according to the type of the driving road to obtain the natural driving data set corresponding to the expressway, the urban road and the general road.
Step S103: and screening the optimal driving data set of the fuel economy according to the vehicle type and the road type, wherein the screening standard is that the fuel consumption is minimum.
In this step, the natural driving data set classified in step S102 is ranked according to the fuel consumption by using a ranking manner of the large root stacks.
The heap sort method is specifically implemented as follows: data { Q ] of fuel consumption 1 ，Q 2 ，Q 3 ，......Q n Put into a complete binary tree, since the value of each node of the large root heap is greater than or equal to the value of the corresponding child node, i.e., Q 1 ≥Q 2i +1 and Q 1 ≥Q 2i +2. Setting the current element in the arrayRepresentation, then its left child->Right childParent node is +.>First, the array is adjusted to be heap +.>Exchange->And->Second, will->Adjust to heap, exchange->And->Repeating the above steps until +.>And->Until that point. After a series of processing, an ordered result data set with the fuel economy from small to large can be obtained (the driving data set with the best fuel economy refers to the driving data set corresponding to the smallest fuel consumption, and the data set comprises a gear system, an accelerator pedal stroke, a brake pedal stroke, a throttle opening, a vehicle speed and a steering wheel angle).
The step involves a plurality of processing flows, including data acquisition, neural network feature extraction in the data set, normalization processing of extraction results, classification of commercial vehicle types by a cluster analysis method, sorting according to fuel consumption by a large-root stack sorting method, and generation of an optimal natural driving data set, wherein the sequence of the specific processing flows is shown in step S200 in fig. 2, after the above flows are completed, the optimal driving data set of fuel economy is prepared, and the following gear shifting control operations can be performed:
step S110: road information within a specified range of a road where a driving vehicle is located is obtained, for example, road information within 1 km is obtained by using an ADAS map in an intelligent networking technology, and the road information is classified into urban roads, expressways and general roads according to road types; further acquiring information such as traffic incidents, traffic signboards, traffic participants and the like distributed on a road side through V2X; the intelligent camera extracts lane line image data and classifies lane information into two types of straight roads and curves.
In the actual gear shifting operation of the vehicle, when the vehicle is in straight running and turns, the angle of the steering wheel along with the road is adjusted, and the gear of the vehicle can be adjusted along with the adjustment, so that in the application, the lane type data are required to be added into the dynamically matched condition, and the condition of the curve or the straight road is required to be considered when the road information of the running vehicle in the specified range of the road is acquired.
In this step, the classification of the road information is as shown in S300 in fig. 3, and includes the following operations: and extracting a lane line image, performing blurring processing on the lane line image, and analyzing the image subjected to blurring processing to obtain the type of the lane.
The blurring process mainly comprises the steps of performing binarization processing on an image where a lane line is located to generate a binarized image; and judging the lane type according to the binarized image and outputting a classification result.
However, the road information is judged through the lanes, and the situations of lane line missing exist in the actual situations, such as lane line abrasion missing, crossroads, ramps, intersections and the like, the reasons cause incomplete lane images, and errors exist in analyzed straight-way and curve results, so that after the binary images are generated, the binary images need to be optimized to solve the problem of incomplete lane images.
The operation of optimizing the binarized image includes:
judging whether the lane line environment condition needs to execute the line supplementing operation, if so, calculating an optimization basis, and executing the line supplementing operation according to the binarized image of the optimization basis.
And in this application, the environmental conditions for which the patch cord operation needs to be performed include: lane line missing, crossroads, ramps and intersections; if the environmental condition is an intersection and the vehicle moves straight, calculating a slope to generate an optimization basis; if the environmental condition is a ramp, a fork, a crossroad left turn or right turn, calculating curvature to generate an optimization basis; if the environmental condition is lane line missing, calculating a linear track to generate an optimization basis;
the process of judging the lane type according to the binarized image is as follows: searching black and white jumping points of the binarized image by adopting a horizontal and longitudinal line inspection method, and setting the resolution of the binarized image as M multiplied by N and the left boundary array asRight boundary array is +.>The longitudinal boundary array is vertical boundary [ M ]][N]Starting line inspection lines by using an Mth line of an image, and storing the first left and right black and white jumping points into the array to obtain a boundary array of the starting lines Wherein middle represents the central point of the M line, a represents the column number of the first black and white jumping point translation which is searched leftwards by taking the central point as a starting point, b represents the column number of the first black and white jumping point translation which is searched rightwards by taking the central point as a starting point, c represents the line number of the first black and white jumping point translation which is searched longitudinally, then the first black and white jumping point is searched by sequentially climbing the obtained array boundary, and the boundary of the image is obtained after the whole image is searched.
And then calculating the slope, curvature or linear track of the road according to the optimization basis, and supplementing the line according to the slope, curvature or linear track, wherein the road after supplementing the line is one of a straight road or a curved road.
The optimized binarized image can more accurately pre-judge the information of straight road and curve road, corresponding to steering wheel angle and the like, so that the data are more accurate when the optimal driving behavior is dynamically matched, the gear shifting effect of the automatic-shift commercial vehicle is improved, the fuel consumption of the automatic-shift commercial vehicle is reduced, and the driving comfort is improved.
Step S120: updating and reorganizing the road information acquired in the step S110 to generate road network information, and matching the road network information with the optimal driving behavior from the optimal driving data set of the fuel economy to generate gear shifting decision information;
dynamically matching the best driving behavior from the fuel economy best driving data set includes: and (2) combining the optimal driving data set with the continuously-changed road driving environment information obtained in the step (S110), combining road side traffic participants (motor vehicles, non-motor vehicles, pedestrians and animals) with an optimal driving behavior information acquisition layer, updating and reorganizing a road network, matching the driving behaviors (the stroke of stepping an accelerator pedal by a driver, the stroke of a brake pedal and the steering wheel rotation angle) which accord with the current driving environment in the natural driving data set, deciding a corresponding throttle opening degree, a vehicle speed, an engine torque and outputting an optimal gear, and sending matching decision information to a domain controller of the vehicle, wherein the gear shifting decision information comprises the optimal gear shifting position and corresponding time.
Step S130: outputting the gear shifting decision information to a driving end;
the control strategy after the domain controller is processed and matched with the decision is displayed on a display at the driving end, and a driver switches the gear to the optimal gear according to information prompt; the automatic transmission commercial vehicle can be controlled to switch gears according to logic control, and is operated according to driving behavior with optimal fuel economy, so that the engine torque, the vehicle speed and the throttle opening of the automatic transmission commercial vehicle reach decision values, and an operation closed-loop system is formed.
The method is also suitable for the automatic shift commercial vehicle, road information and fuel consumption are added on the basis of the original automatic shift to serve as shift basis, so that the shift control is performed on the basis of considering the traditional dynamic property, the oil consumption is further improved, and the economic cost is reduced.
In the step, the information after the operation is executed is collected and fed back to the domain controller of the vehicle.
The device for controlling gear shifting based on natural driving data provided by the application interacts with a vehicle central domain controller of a controlled vehicle. Fig. 4 provides a block diagram of a device for controlling gear shifting based on natural driving data, which includes the following parts:
p400 driving behavior acquisition module: the natural driving data comprises vehicle appearance, gear system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed, steering wheel rotation angle and other information.
The driving behavior acquisition module comprises a natural driving data extraction unit: the method is used for collecting basic information such as natural driving data of the manual transmission commercial vehicle.
The driving behavior acquisition module also needs to screen and classify natural driving data, extracts a driving data set with optimal fuel economy for controlling gear shifting, and comprises the following parts for supporting the function:
neural network extraction unit: the method comprises the steps of performing preliminary screening on accurate image data for identifying the vehicle type from a natural driving data set, and adding the image data with normal identification into a result data set;
normalization processing unit: the method comprises the steps of calculating the identification accuracy probability of a primarily screened result data set, extracting a data value with the highest probability value, and reserving a data result with the highest probability value in the result data set for processing in the next step;
clustering unit: the method is used for classifying the vehicle type and the road type from the result data set after the preliminary screening and the normalization processing, the vehicle type involved in the step is a commercial vehicle type, and the method comprises the following steps: passenger cars, trucks, semi-trailer traction vehicles, road types including expressways, urban roads, general roads;
a heap sort unit: the fuel consumption index sorting method comprises the steps of sorting fuel consumption indexes by using a heap sorting algorithm according to a result data set which is classified according to the vehicle type and the road type; thus, the heap sort unit may process the result data set into a ordered data set sorted by fuel consumption. During the execution of the control shift, a data set with the lowest fuel consumption can be extracted from the ordered data collection set.
P410 information awareness module: the method comprises the steps of interacting with a third party system, and obtaining highway information and lane information in a specified range of a road where a driving vehicle is located, wherein the lane information comprises a straight road and a curve;
in the module, a third party system such as an ADAS map platform, a Y2X communication platform, an intelligent camera interface and the like is supported.
The information sensing module is used for identifying lane line images of lane information through the road information processing unit, performing blurring processing on the lane line images, and judging and outputting lane types.
Since the lane line image has incomplete lanes, such as abrasion of the lane line, and in the case of an intersection, the lane information cannot be directly identified, the road information processing unit further includes an image optimization sub-module: when the blurring processing is performed on the lane line image, the incomplete image is complemented, so that the situation that the lane line image is incomplete is solved.
P420 match decision module: the method comprises the steps of generating road network information according to road information, matching optimal driving behaviors from natural driving data, and generating gear shifting decision information;
the matching decision module comprises the following parts:
a road network reorganization unit: the road network updating and reorganizing method is used for dynamically integrating road running environment information, road side traffic environment information and lane information in a specified range, and generating a dynamic running environment which comprises a road type and a lane type;
matching unit: the driving behavior matching method is used for matching the dynamic road type and the lane type with the vehicle type of the control vehicle, and the optimal driving data set of the fuel economy accords with the driving behavior of the current driving environment;
decision information generation unit: the method is used for deciding the corresponding throttle opening, the corresponding vehicle speed, the corresponding engine torque and the corresponding output optimal gear according to the driving behavior conforming to the current driving environment.
The matching decision module is used for continuously acquiring new road information and continuously updating gear shifting decision information in the vehicle driving process, so that the whole matching process is a dynamic processing process.
P430 man-machine interaction module: and the method is used for outputting gear shifting decision information such as the optimal gear and the like.
The man-machine interaction module displays a control strategy after the domain controller processes and matches the decision on a display, and a driver switches the gear to the optimal gear according to information prompt; the automatic transmission commercial vehicle is also supported to switch gears according to logic control, and is operated according to driving behavior with optimal fuel economy, so that the engine torque, the vehicle speed and the throttle opening of the automatic transmission commercial vehicle reach decision values, and an operation closed-loop system is formed.
The man-machine interaction module also supports the feedback of information after the execution of the operation to the domain controller.
According to the technical scheme, based on the data sample of the manual gear natural driving data set, driving data comprise driving behavior information of a plurality of vehicles of various types in various road environments, and abundant information provides data support for the method, so that the method is wide in application range and high in reliability, meanwhile, defects possibly occurring in road environment acquisition are optimized, the effectiveness and accuracy of predictive humanoid intelligent control gear shifting are further improved, optimal gear shifting time can be maximally achieved, fuel consumption is reduced, and an optimal driving state is achieved.
The foregoing disclosure is merely illustrative of some embodiments of the present invention, and the present invention is not limited thereto, as modifications may be made by those skilled in the art without departing from the scope of the present invention.
Claims (10)
1. A method for controlling gear shifting based on natural driving data, comprising the steps of:
obtaining highway information in a specified range of a road where a running vehicle is located, wherein the highway information comprises a road type and a lane type;
generating road network information according to the road information, dynamically matching the optimal driving behavior from a driving data set with optimal fuel economy, and generating gear shifting decision information, wherein the decision information comprises an optimal gear shifting gear;
outputting the gear shifting decision information to a driving end;
wherein obtaining the lane type comprises: and extracting a lane line image, performing blurring processing on the lane line image, and judging and outputting lane types, wherein the lane types comprise straight lanes and curves.
2. The method of controlling a gear shift according to claim 1, wherein a natural driving dataset is collected before dynamically matching an optimal driving behavior from a fuel economy optimal driving dataset, and a fuel economy optimal driving dataset is extracted from the natural driving dataset;
collecting a natural driving data set refers to: natural driving data of the manual transmission commercial vehicle is collected, wherein the natural driving data comprises vehicle appearance, a transmission system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed and steering wheel rotation angle.
3. A method of controlling a gear shift according to claim 2, characterized in that extracting a fuel economy optimal driving data set from the natural driving data set comprises the steps of:
extracting a result data set with accurate vehicle type identification from the natural driving data set;
performing vehicle type classification and road type classification from the result dataset;
and screening a fuel economy optimal driving data set according to the vehicle type and the road type, wherein the screening standard of the fuel economy optimal driving data set is that the fuel consumption is minimum.
4. A method of controlling a gear shift according to claim 3, wherein the extracting a vehicle type identification accurate result data set from the natural driving data set comprises:
identifying the appearance characteristics of the vehicle from the natural driving data set, and extracting a characteristic result set with accurate vehicle type identification;
outputting a vehicle type result according to the vehicle appearance characteristics by adopting a cluster analysis method;
the method comprises the steps of extracting vehicle appearance characteristics, wherein the extracting of the vehicle appearance characteristics comprises preliminary screening and normalization processing;
the primary screening refers to: screening the recognition results of the image data in the natural driving data set by adopting the loss function score to generate a characteristic result set, wherein the specific algorithm is as follows:
wherein L is i For loss function score, s j For error discrimination class score, s yi A class score for correct discrimination;
if L i <=0, i.e. the recognition result is accurate, and the corresponding image data is added into the feature result set;
the normalization processing refers to: calculating the recognition accuracy probability of the image data in the characteristic result set, and clearing the data corresponding to the characteristic result set according to the recognition accuracy probability.
5. The method of controlling shifting according to claim 1, characterized in that the performing blurring processing on the lane line image means;
performing binarization processing on the image where the lane line is located to generate a binarized image;
optimizing the binarized image;
searching black and white jumping points of the binarized image to obtain the boundary of the lane line image;
and judging the type of the lane according to the boundary of the image.
6. The method of controlling shifting according to claim 5, wherein the optimizing the binarized image refers to:
judging whether the lane line environment condition needs to execute the line supplementing operation or not; the environmental conditions for which the repair line needs to be executed include: lane line missing, crossroads, ramps and intersections;
if the environmental condition is an intersection and the vehicle moves straight, calculating a slope to generate an optimization basis; if the environmental condition is a ramp, a fork, a crossroad left turn or right turn, calculating curvature to generate an optimization basis; if the environmental condition is lane line missing, calculating a linear track to generate an optimization basis;
and executing line supplementing operation on the binarized image according to the optimization basis.
7. A device for controlling gear shifting based on natural driving data, comprising:
driving behavior acquisition module: the method comprises the steps of acquiring natural driving data of a manual-gear commercial vehicle and providing a driving data set with optimal fuel economy; the natural driving data comprise vehicle appearance, a gear system, fuel consumption, accelerator pedal stroke, brake pedal stroke, throttle opening, vehicle speed and steering wheel rotation angle;
an information perception module: the method comprises the steps of interacting with a third party system, and obtaining highway information in a specified range of a road where a driving vehicle is located, wherein the highway information comprises a road type and a lane type, and the lane type comprises a straight road and a curve;
and a matching decision module: the method comprises the steps of generating road network information according to road information, dynamically matching optimal driving behaviors from a driving data set with optimal fuel economy, and generating gear shifting decision information;
and the man-machine interaction module is used for: the gear shifting decision information is used for outputting the gear shifting decision information, and the decision information comprises an optimal gear shifting position;
the information sensing module comprises a road information processing unit and is used for judging the type of the lane from the lane line image, and the information sensing module comprises the steps of performing blurring processing on the lane line image, judging and outputting the type of the lane.
8. The apparatus for controlling gear shifting according to claim 7, characterized in that the driving behavior acquisition module includes:
natural driving data extraction unit: the system is used for collecting natural driving data of the manual-shift commercial vehicle;
neural network extraction unit: the result data set is used for extracting the vehicle type identification accuracy from the natural driving data set for preliminary screening;
normalization processing unit: the data value with the highest probability value is extracted;
clustering unit: for vehicle type classification and road type classification from the result dataset;
a heap sort unit: and the method is used for screening the optimal driving data set of the fuel economy by using a heap sorting algorithm according to the vehicle type and the road type.
9. The shift control device of claim 7, wherein the match decision module comprises:
a road network reorganization unit: the road network updating and reorganizing method is used for dynamically integrating road running environment information, road side traffic environment information and lane information in a specified range, and generating a dynamic running environment, wherein the dynamic running environment comprises a road type and a lane type;
matching unit: the driving behavior of the current dynamic driving environment is matched with the optimal driving data set of the fuel economy by combining the road type and the lane type and combining the vehicle type of the control vehicle;
decision information generation unit: the method is used for generating the throttle opening, the vehicle speed, the engine torque and outputting the optimal gear according to the driving behavior conforming to the current dynamic driving environment.
10. The gear shifting control apparatus according to claim 7, wherein the road information processing unit further comprises an image optimizing sub-module: and when the blurring processing is performed on the lane line image, the incomplete image is complemented.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211033655.XA CN115476857B (en) | 2022-08-26 | 2022-08-26 | Method and device for controlling gear shifting based on natural driving data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211033655.XA CN115476857B (en) | 2022-08-26 | 2022-08-26 | Method and device for controlling gear shifting based on natural driving data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115476857A CN115476857A (en) | 2022-12-16 |
| CN115476857B true CN115476857B (en) | 2023-08-04 |
Family
ID=84421012
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211033655.XA Active CN115476857B (en) | 2022-08-26 | 2022-08-26 | Method and device for controlling gear shifting based on natural driving data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115476857B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107131296A (en) * | 2017-05-25 | 2017-09-05 | 重庆大学 | A kind of pure electric automobile two towards energy consumption keeps off speed change system control strategy |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102011080712A1 (en) * | 2011-08-10 | 2013-02-14 | Zf Friedrichshafen Ag | Method for influencing a transmission shift strategy of a motor vehicle |
| CN203237137U (en) * | 2013-04-12 | 2013-10-16 | 北汽福田汽车股份有限公司 | Device used for reminding economic gear-shifting operation in stick-shift car and stick-shift car |
| KR101826550B1 (en) * | 2015-11-30 | 2018-02-07 | 현대자동차 주식회사 | Device for controlling shift of vehicle and method for controlling shift using the same |
| CN205554180U (en) * | 2016-02-15 | 2016-09-07 | 潍柴动力股份有限公司 | Self -adaptation driver assistance system |
| CN109237011A (en) * | 2018-09-11 | 2019-01-18 | 江苏大学 | A kind of automatic transmission gear shift control method containing driving behavior prediction model |
| CN110395268B (en) * | 2019-07-05 | 2021-06-18 | 吉林大学青岛汽车研究院 | Truck economic driving reminding system based on front road information |
| CN111703427B (en) * | 2020-06-30 | 2021-06-22 | 潍柴动力股份有限公司 | Method and device for switching gear shifting strategy, electronic equipment and storage medium |
| US11608067B2 (en) * | 2020-08-12 | 2023-03-21 | Honda Motor Co., Ltd. | Probabilistic-based lane-change decision making and motion planning system and method thereof |
| KR102393138B1 (en) * | 2020-10-26 | 2022-05-04 | 주식회사 현대케피코 | Gear shift indication apparatus and method based on driver desired speed |
| CN112818910B (en) * | 2021-02-23 | 2022-03-18 | 腾讯科技（深圳）有限公司 | Vehicle gear control method and device, computer equipment and storage medium |
-
2022
- 2022-08-26 CN CN202211033655.XA patent/CN115476857B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107131296A (en) * | 2017-05-25 | 2017-09-05 | 重庆大学 | A kind of pure electric automobile two towards energy consumption keeps off speed change system control strategy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115476857A (en) | 2022-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2019101142A4 (en) | A pedestrian detection method with lightweight backbone based on yolov3 network | |
| CN109993082B (en) | Convolutional neural network road scene classification and road segmentation method | |
| CN108985194B (en) | Intelligent vehicle travelable area identification method based on image semantic segmentation | |
| US20210197851A1 (en) | Method for building virtual scenario library for autonomous vehicle | |
| CN112508977A (en) | Deep learning-based semantic segmentation method for automatic driving scene | |
| CN113343461A (en) | Simulation method and device for automatic driving vehicle, electronic equipment and storage medium | |
| CN108830254B (en) | Fine-grained vehicle type detection and identification method based on data balance strategy and intensive attention network | |
| CN112990065A (en) | Optimized YOLOv5 model-based vehicle classification detection method | |
| WO2020007589A1 (en) | Training a deep convolutional neural network for individual routes | |
| CN110991523A (en) | Interpretability evaluation method for unmanned vehicle detection algorithm performance | |
| CN110304068A (en) | Acquisition method, device, equipment and the storage medium of running car environmental information | |
| CN115019279A (en) | Context feature fusion method based on MobileNet lightweight network | |
| CN112009491B (en) | Deep learning automatic driving method and system based on traffic element visual enhancement | |
| CN116205024A (en) | Self-adaptive automatic driving dynamic scene general generation method for high-low dimension evaluation scene | |
| CN115476857B (en) | Method and device for controlling gear shifting based on natural driving data | |
| CN116630702A (en) | Pavement adhesion coefficient prediction method based on semantic segmentation network | |
| CN115273003A (en) | Traffic sign recognition and navigation decision method and system combining character positioning | |
| CN115019263A (en) | Traffic supervision model establishing method, traffic supervision system and traffic supervision method | |
| CN113147781A (en) | Intelligent driving information display system for automobile | |
| Bezawada et al. | Modeling lane and road characteristics for drive quality rating | |
| CN113343817A (en) | Unmanned vehicle path detection method and device for target area and medium | |
| CN113989780A (en) | Sign board detection method and device | |
| CN112785863B (en) | Merging decision classification early warning method based on K-Means and entropy weighting | |
| Gautam et al. | Image Segmentation for Self-Driving Car | |
| CN115188205B (en) | Road information-based automobile driving condition correction method |
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 |