CN108898697A - A kind of road surface characteristic acquisition methods and relevant apparatus - Google Patents
A kind of road surface characteristic acquisition methods and relevant apparatus Download PDFInfo
- Publication number
- CN108898697A CN108898697A CN201810825838.2A CN201810825838A CN108898697A CN 108898697 A CN108898697 A CN 108898697A CN 201810825838 A CN201810825838 A CN 201810825838A CN 108898697 A CN108898697 A CN 108898697A
- Authority
- CN
- China
- Prior art keywords
- data
- road surface
- neural network
- surface characteristic
- fuzzy neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0808—Diagnosing performance 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
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Fuzzy Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Mathematics (AREA)
- Biomedical Technology (AREA)
- Mathematical Analysis (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Automation & Control Theory (AREA)
- Traffic Control Systems (AREA)
Abstract
This application discloses a kind of road surface characteristic acquisition methods, including:It is trained processing by initial fuzzy neural network of the history pavement behavior data to building, obtains objective fuzzy neural network;Sensor information is analyzed and processed to obtain practical pavement behavior data;Calculation processing is carried out to the practical pavement behavior data using the objective fuzzy neural network, obtains preset quantity road surface characteristic probable value;Road surface characteristic is determined according to all road surface characteristic probable values.Judged by the fuzzy neural network region feature that satisfies the need, improve the accuracy of road surface characteristic acquisition, avoids the inaccurate situation directly judged by sensing data.Disclosed herein as well is a kind of road surface characteristics to obtain system, server and computer readable storage medium, has the above beneficial effect.
Description
Technical field
This application involves field of computer technology, in particular to a kind of road surface characteristic acquisition methods, road surface characteristic obtain system
System, server and computer readable storage medium.
Background technique
As information technology continues to develop, Internet technology and technology of Internet of things are that people's lives are brought greatly just
Benefit.Wherein, the operation data for acquiring the vehicle in real time from vehicle in terms of logistics sends cloud by network and carries out feature meter
It calculates, to judge what road surface the road surface that the vehicle is travelled is, that is, goes out the road on road surface at this time to the information analysis of sensor
The road conditions feature is sent to driver or is stored in logistics data library by condition feature, so that subsequent vehicle carries out speed-optimization.
But what running for vehicle sensor information carried out in the prior art is directly to judge, that is, pass through
The data that height sensor obtains, which can only obtain, jolts when the road surface travelled at this time, and makes a concrete analysis of out the feature on road surface at this time
Be how, that is, directly judged by sensing data, deterministic process not enough it is intuitive directly.Effective road can not be used as
Planar condition Information application in a practical situation so that pavement behavior judge it is inefficient and useless.It is another in the prior art, use
The identifying system of camera and corresponding analytical algorithm, parameter is relatively simple in the identifying system, and single parameter can not be fine
The judgement for continuing road surface characteristic to complicated road environment, accuracy and validity are lower.
Therefore, how to improve the validity of pavement behavior analysis is the Important Problems of those skilled in the art's concern.
Summary of the invention
The purpose of the application is to provide a kind of road surface characteristic acquisition methods, road surface characteristic obtains system, server and meter
Calculation machine readable storage medium storing program for executing is judged by the fuzzy neural network region feature that satisfies the need, and improves the accurate of road surface characteristic acquisition
Property, avoid the situation of the inaccuracy directly judged by sensing data.
In order to solve the above technical problems, the application provides a kind of road surface characteristic acquisition methods, including:
It is trained processing by initial fuzzy neural network of the history pavement behavior data to building, obtains objective fuzzy
Neural network;
Sensor information is analyzed and processed to obtain practical pavement behavior data;
Calculation processing is carried out to the practical pavement behavior data using the objective fuzzy neural network, obtains present count
Measure a road surface characteristic probable value;
Road surface characteristic is determined according to all road surface characteristic probable values.
Optionally, processing is trained by initial fuzzy neural network of the history pavement behavior data to building, obtained
Objective fuzzy neural network, including:
The initial fuzzy neural network is constructed according to the membership function received;
Processing is trained to the initial fuzzy neural network by the history pavement behavior data, obtains the mesh
Mark fuzzy neural network.
Optionally, sensor information is analyzed and processed to obtain practical pavement behavior data, including:
Vibration data processing is carried out to vibrating sensor information, obtains road vibration data;
Turning data processing is carried out according to infrared electro gate signal and gyro data, obtains traveling turning data;
To sensor, inclination data and acceleration information carry out Gradient processing forward, obtain road gradient data;
Position data is got by GPS positioning module;
By the road vibration data, the traveling turning data, the road gradient data and the position data
As the practical pavement behavior data.
Optionally, calculation processing is carried out to the practical pavement behavior data using the objective fuzzy neural network, obtained
To preset quantity road surface characteristic probable value, including:
By the road vibration data, the traveling turning data, the road gradient data and the position data
Input as the objective fuzzy neural network;
Calculating operation is executed to the objective fuzzy neural network, obtaining the preset quantity road surface characteristic may
Value.
The application also provides a kind of road surface characteristic acquisition system, including:
Fuzzy neural network training module, for the initial fuzzy neural network by history pavement behavior data to building
It is trained processing, obtains objective fuzzy neural network;
Real data obtains module, obtains practical pavement behavior data for being analyzed and processed to sensor information;
Fuzzy neural network computing module, for using the objective fuzzy neural network to the practical pavement behavior number
According to calculation processing is carried out, preset quantity road surface characteristic probable value is obtained;
Road surface characteristic obtains module, for determining road surface characteristic according to all road surface characteristic probable values.
Optionally, the fuzzy neural network training module, including:
Initial model construction unit, for constructing the initial fuzzy neural network according to the membership function received;
Model training unit, for being instructed by the history pavement behavior data to the initial fuzzy neural network
Practice processing, obtains the objective fuzzy neural network.
Optionally, the real data obtains module, including:
Road vibration data capture unit obtains road surface vibration for carrying out vibration data processing to vibrating sensor information
Dynamic data;
Turning data acquiring unit, for carrying out turning data processing according to infrared electro gate signal and gyro data,
Obtain traveling turning data;
Gradient acquiring unit, for inclination data and acceleration information to carry out at Gradient forward to sensor
Reason, obtains road gradient data;
Position data acquiring unit, for getting position data by GPS positioning module;
Road surface real data acquiring unit is used for the road vibration data, the traveling turning data, the road surface
Gradient and the position data are as the practical pavement behavior data.
Optionally, the fuzzy neural network computing module, including:
Neural network input unit is used for the road vibration data, the traveling turning data, the road gradient
The input of data and the position data as the objective fuzzy neural network;
Neural computing unit obtains described default for executing calculating operation to the objective fuzzy neural network
The quantity road surface characteristic probable value.
The application also provides a kind of server, including:
Memory, for storing computer program;
Processor, the step of road surface characteristic acquisition methods as described above are realized when for executing the computer program.
The application also provides a kind of computer readable storage medium, and calculating is stored on the computer readable storage medium
The step of machine program, the computer program realizes road surface characteristic acquisition methods as described above when being executed by processor.
A kind of road surface characteristic acquisition methods provided herein, including:By history pavement behavior data to building
Initial fuzzy neural network is trained processing, obtains objective fuzzy neural network;Sensor information is analyzed and processed
To practical pavement behavior data;The practical pavement behavior data are carried out at calculating using the objective fuzzy neural network
Reason, obtains preset quantity road surface characteristic probable value;Road surface characteristic is determined according to all road surface characteristic probable values.
By carrying out calculation processing to sensor information in trained fuzzy neural network, multiple road feature pair is obtained
The probable value answered, it would be possible to be worth maximum road surface characteristic as acquisition as a result, the membership function rule wherein obtained from expertise
Model various road surface characteristics corresponding parameter areas delimit the corresponding fuzzy value of various features with these parameter areas, thus gram
Taken by sensor information directly judged caused by non-intuitive problem, improve satisfy the need region feature judgement objectivity with
Accuracy.
The application also provides a kind of road surface characteristic and obtains system, server and computer readable storage medium, have with
Upper beneficial effect, this will not be repeated here.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of flow chart of road surface characteristic acquisition methods provided by the embodiment of the present application;
Fig. 2 is a kind of structural schematic diagram of road surface characteristic acquisition system provided by the embodiment of the present application.
Specific embodiment
The core of the application is to provide a kind of road surface characteristic acquisition methods, road surface characteristic obtains system, server and meter
Calculation machine readable storage medium storing program for executing is judged by the fuzzy neural network region feature that satisfies the need, and improves the accurate of road surface characteristic acquisition
Property, avoid the situation of the inaccuracy directly judged by sensing data.
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
It is general in the prior art that the data that sensor obtains directly are showed into user, such as the vibration data of vehicle, or
Person is the image data on road surface, and user needs to judge by these data, and it is special at what just to can analyze out vehicle at this time
The road traveling of sign does not analyze the data of sensor, and user also thinks deeply, and reduces the validity of sensor.?
Another kind is directly to judge the sensor information of vehicle acquisition, and distinguish often through sensor information in the prior art
Other road surface characteristic is fuzzy, boundary and fuzzy, directly carries out judgement and the accuracy for the region feature that satisfies the need is be easy to cause to decline,
It is unfavorable for user and vehicle driving is optimized by road surface characteristic.
Therefore, the present embodiment provides a kind of road surface characteristic acquisition methods, by trained fuzzy neural network to biography
Sensor information carries out calculation processing, obtains the corresponding probable value of multiple road feature, it would be possible to be worth maximum road surface characteristic conduct
It obtains as a result, the membership function specification wherein obtained from expertise various road surface characteristics corresponding parameter areas, with these
Parameter area delimit the corresponding fuzzy value of various features, thus overcome directly judged by sensor information caused by not
Accurate problem improves the objectivity and accuracy of region feature judgement of satisfying the need.
Specifically, referring to FIG. 1, Fig. 1 is a kind of process of road surface characteristic acquisition methods provided by the embodiment of the present application
Figure.
This method may include:
S101 is trained processing by initial fuzzy neural network of the history pavement behavior data to building, obtains mesh
Mark fuzzy neural network;
This step is intended to construct initial fuzzy neural network by the membership function that expertise forms, then passes through history road
Face status data is trained processing to initial fuzzy neural network, obtains objective fuzzy neural network.
Wherein, objective fuzzy neural network be exactly product that fuzzy theory is combined with neural network, it summarizes mind
Through network and the advantages of fuzzy theory, integrate study, association, identification, information processing.Fuzzy logic control, neural network
It is three kinds of typical intelligent control methods with Multimode Control.Usual expert system is established in expertise, is not established in work
In operation data caused by industry process, even and if inexactness, uncertainty domain expert possessed by general complication system
Also it is difficult to hold, this makes to establish expert system extremely difficult.And fuzzy logic and neural network are as two kinds of typical intelligence controls
Method processed, respectively have it is excellent lack, fuzzy logic is merged with neural network --- fuzzy neural network (Fuzzy Neural
Network the shortcomings that the advantages of) having drawn fuzzy logic and neural network part avoids the two.Therefore, it can be very good
In being analyzed and processed to road condition data, the fuzzy of sensor information is sentenced in the membership function realization being made up of the experience of expert
It is disconnected, to avoid directly carrying out judging bring inaccuracy to sensor information.
Optionally, this step may include:
Step 1: constructing initial fuzzy neural network according to the membership function received;
Step 2: being trained processing to initial fuzzy neural network by history pavement behavior data, target mould is obtained
Paste neural network.
Wherein, membership function is to be constructed to be input in system after obtaining according to expertise.Membership function mainly describes
Expertise parameter area corresponding for each characteristic value delimit the corresponding fuzzy value of each feature with these parameter areas,
And then obtain the corresponding fuzzy characteristics of some sensor information.In building fuzzy neural network, membership function is for constructing mould
The blurring layer for pasting neural network has used membership function to construct fuzzy in the structure of other layers of solemn lake neural network
Change layer.
Therefore, this optinal plan is exactly to first pass through membership function to construct initial fuzzy neural network, is gone through further according to existing
History pavement behavior data are trained processing to initial fuzzy neural network, obtain objective fuzzy neural network.
Wherein, history pavement behavior data are the given datas of known pavement behavior feature and corresponding pavement behavior data
Collection.More specifically, pavement behavior data are exactly the sensor information got, and the sensor information with defined road
Planar condition feature corresponds.Initial fuzzy neural network can be trained, be obtained by the history pavement behavior data
To objective fuzzy neural network.
Wherein, any one fuzzy neural network training side that the method for training managing can be provided using the prior art
Method is not specifically limited herein.It is exactly trained fuzznet by the obtained objective fuzzy neural network of training
Network model.
S102 is analyzed and processed sensor information to obtain practical pavement behavior data;
On the basis of step S101, this step is intended to be analyzed and processed to obtain reality to the sensor information obtained in real time
Border pavement behavior data.Such as judge whether vibration data is greater than the reasonable value of vibration, if it is, the vibration data is recorded,
If it is not, then not recording the data;Can also be according to infrared electro gate signal judge this gyro data whether be turn
It is curved to be generated in turning, if so, the gyro data is recorded, if otherwise not recording the gyro data.Be in a word
The operative sensor information that sensor records in actual environment is redundancy, the data that can not be judged as pavement behavior,
Therefore it needs to be analyzed and processed the sensor information that sensor directly acquires to obtain practical pavement behavior data.
Further, it also provides in the prior art in order to keep not mixing garbage in sensor information to biography
The method that sensor information is analyzed and processed.The side of any one analysis processing can also be provided in this step using the prior art
Method is not done specifically repeat herein.
Wherein, more specifically practical pavement behavior data may include road vibration data, traveling turning data, road surface slope
Degree is accordingly and position data.Therefore, this step may include following optinal plan:
Step 1: carrying out vibration data processing to vibrating sensor information, road vibration data are obtained;
Step 2: carrying out turning data processing according to infrared electro gate signal and gyro data, traveling number of turns is obtained
According to;
Step 3: to sensor, inclination data and acceleration information carry out Gradient processing forward, obtain road gradient
Data;
Step 4: getting position data by GPS positioning module;
Step 5: using road vibration data, traveling turning data, road gradient data and position data as practical road
Face status data.
Wherein, the sequencing relationship that step 1 to step 4 is not carried out, can be executed with random order.
S103 carries out calculation processing to practical pavement behavior data using objective fuzzy neural network, obtains preset quantity
A road surface characteristic probable value;
On the basis of step S102, this step be intended to using objective fuzzy neural network to practical pavement behavior data into
Row calculation processing obtains the corresponding road surface characteristic probable value of multiple road surface characteristics.The purpose of calculating of objective fuzzy neural network is
Calculate the probable value that the practical pavement behavior data correspond to every kind of road surface characteristic.Wherein it is possible to which setting specifically has according to demand
Certain several road surface characteristic, can also be set according to expertise specifically has certain several road surface characteristic, the sensor that can also be inputted
The type set of data specifically has certain several road surface characteristic.Specifically, road surface characteristic can for example how curved slight concave-convex blinding
Road surface, few curved slight concave-convex cement pavement, few curved severe bumps cement pavement etc..
By using, objective fuzzy neural network practical pavement behavior data calculate with to can be obtained by correspondence above-mentioned
The road surface characteristic probable value of multiple road surface characteristics.Road surface characteristic probable value illustrates a possibility that being at this time road surface characteristic in this,
Being up to 1 or 100%, minimum 0 indicates be this kind of road surface characteristic.
The representation of road surface characteristic probable value can be 0 to 1 value, be also possible to 0 to 100% percent value, also
It can select, be not specifically limited herein according to the actual situation.
Wherein, it can be used using the method that objective fuzzy neural network carries out calculation processing to practical pavement behavior data
Any one calculation method that the prior art provides, is not specifically limited herein.
Optionally, this step may include:
Step 1: using road vibration data, traveling turning data, road gradient data and position data as target mould
Paste the input of neural network;
Step 2: executing calculating operation to objective fuzzy neural network, preset quantity road surface characteristic probable value is obtained.
This optinal plan be intended to by previous step road vibration data, traveling turning data, road gradient data with
And input of the position data as objective fuzzy neural network, then calculating operation is executed to objective fuzzy neural network, it obtains pre-
If the road surface characteristic probable value of quantity.
S104 determines road surface characteristic according to all road surface characteristic probable values.
On the basis of step S103, this step can using the corresponding road surface characteristic of maximum road surface characteristic probable value as
The road surface characteristic of acquisition.The maximum corresponding road surface characteristic in the road surface characteristic probable value of preset quantity is exactly practical at this time
The corresponding most possible road surface characteristic of pavement behavior data, which is handled as road pavement status data
Processing result.
To sum up, the present embodiment is obtained by carrying out calculation processing to sensor information in trained fuzzy neural network
The corresponding probable value of multiple road feature, it would be possible to be worth maximum road surface characteristic as obtaining as a result, wherein fuzzy neural network
Given the corresponding parameter area of various road surface characteristics, the corresponding fuzzy value of various features delimited with these parameter areas, thus
The problem of inaccuracy caused by directly being judged by sensor information is overcome, the objectivity for region feature judgement of satisfying the need is improved
And accuracy.
Based on above embodiments, the embodiment of the present application also provides a kind of more specifical embodiment.
The present embodiment may include:
1, sensor acquisition module
1) using the sensor acquisition on logistics vehicles advance tire vibration number and amplitude during 100 meters, non-overtake other vehicles
When route corner change frequency, three kinds of parameters of road grade;
2) parameter is uploaded to cloud server by In-vehicle networking.
2, fuzzy neural network module
1) it before putting into operation, is trained according to existing history pavement behavior data, every power of optimization neural network
Weight parameter guarantees that the result provided when calculating in real time is accurate and reliable;
2) model after training deposits into the real-time computing subsystem in cloud, after opening the function of calculating in real time, mould
Type is calculated according to the parameter transmitted and provides the judgement of present road feature.
3, result pushing module
1) geographical location information such as longitude and latitude where when result and vehicle being sent result collect point in logistics cloud in deposit
Cloth database, including Hbase (PostgreSQL database distributed, towards column) and Redis (key-value's
Storage system), wherein deposit Hbase is counted as off line data analysis, deposit redis is the business that other need to calculate in real time
Road condition data is provided;
2) have a master control display on fraction of stream vehicle at present, show car body indices, business information and road
Line and road conditions.
In the present embodiment, road conditions judging result will be pushed directly on the master control display of former vehicle, be mentioned for driver
It is referred to for road conditions, driver can also be checked by operation master control screen selection is currently located section historical data.
Wherein, sensor acquisition module includes:
Sensor acquisition module is used for the various input parameters that gathering algorithm needs, these parameters all have by force with roadway characteristic
Correlation.Sensor network main controlled node passes through third party's cartographic information that network obtains according to master control, and every 100 meters to all sons
Node data is counted and is uploaded primary.
1, road surface material and concave-convex situation collection analysis module:
1) sensor:Three direction wireless vibration sensors are selected, four wheels are selected before and after vehicle, in tire suspension swing arm
Place's installation has four groups of sensors altogether;
2) engineer provides the number that traveling is rationally vibrated on flat-satin road surface according to vehicle, and vibrating sensor is to each axis
It is acquired to acceleration, calculates vertical vibration amplitude greater than reasonable amount and records amplitude;
3) master control notice child node is when uploading every hundred meters of data, and all vibrating sensors are by the vibration number and width of record
Value uploads to master control;
4) master control chooses maximum six amplitudes and averages.Finally vibration number and mean value are uploaded.
2, bend analysis of information collection module:
1) sensor:Select a three-axis gyroscope, two highly reliable infrared light electric switches.Three-axis gyroscope is installed on vehicle
Head, calibration initial position it is identical as direction of travel, two infrared range-measurement systems are mounted on vehicle head and tail, highly for one meter two and
Light source levels are directed toward right side;
2) in traveling process, when gyroscope detects the steering angle greater than 25 degree every time, photoelectric door, photoelectricity can all be opened
Output is low when door is not blocked, and being blocked output is high potential;
3) it is blocked in conjunction with whether receiving for entire steering procedure photoelectric door, if there is blocking, is judged as and overtakes other vehicles or have
Obstacle, this time route corner variation are disregarded;
If 4) photoelectric door whole process is not blocked, judge this data belong to it is non-overtake other vehicles or barrier under the conditions of be greater than 25
The turning of degree simultaneously records number;
5) a route corner change frequency is uploaded to master control for every hundred meters.
3, grade information collection analysis module:
1) sensor:Three-axis gyroscope, is located at headstock position, and calibration initial position is identical as direction of travel;
2) it according to the three-axis gyroscope of headstock position, acquires before vehicle to inclination angle and acceleration information;
3) it is responded according to collected acceleration time domain, if certain time measures Acceleration pulse from concussion is started to end
In this section of region, wave crest is continuous with trough, positive and negative contrary sign and amplitude differ by more than 20% or the equidirectional acceleration time less than 2
Second.So judgement is to clear the jumps, and is not counted in gradient parameter;
4) otherwise gyroscope will record this inclination angle;
5) a data are uploaded to master control for every 100 meters;
6) main controlled node uploads to server and obtained parameters is acquired and counted in this meters.
4, GPS positioning module:
1) vehicle-mounted GPS positioning module can obtain the longitude and latitude where current vehicle in real time;
2) master control just is sent by longitude and latitude when main controlled node notifies this node uploading position data.
Wherein, fuzzy neural network module includes:
1, model is established
1) input layer:Master control is uploaded into primary parameter every hundred meters:Amplitude is greater than the tire vibration time of expert's given value
Number (secondary), the mean value (cm) of 6 vertical amplitude of vibration peak values, greater than 25 ° of route corner change frequencies (secondary), road grade (°) four
Parameter divides table to be expressed as x1、x2、x3、x4Fuzzy neural network input;
2) it is blurred layer:The membership function of each roadway characteristic is provided by expert according to vehicle, is drawn according to domain of function
It is divided into several function branches, a node of the corresponding blurring layer of each branch.Such as when certain vehicle, x1、x2、x3、x4It is corresponding
Four groups of membership functions have 3,5,3,5 branches respectively;
3)μ1i(x1) i=1,2,3;μ2i(x2) i=1,2,3,4,5;…;μ4i(x4) i=1,2,3,4,5;Therefore blurring layer
16 nodes altogether.And each node is corresponding with this membership function with upper one layer according to the membership function itself represented
Feature input node is connected.I.e. the output of blurring layer is f2k=μji(xj) j=1,2,3,4;I is membership function branch serial number, k generation
Table is blurred which node of layer;
And layer 4):The number of nodes of this layer is number of fuzzy rules, is blurred in layer totally four group node, selects respectively in four groups
One connect with the node that other groups are selected out, so sharing 3*5*3*5=225 node with layer one.Each node
The product for all signals that output is input to this node for upper one layer is i.e.:K is indicated and which node of layer,
J is the serial number of second layer membership function group, and i is the serial number of a branch in every group of function;
Or layer 5):Number of nodes determines that the design has in 36 export altogether according to the number that output variable fuzziness divides
Variable, so this layer has 36 nodes, each node is totally interconnected, connection weight W with upper one layerkj, wherein k=1,
2 ..., 36 indicate which node of this layer;J=1,2 ..., 255.Weight constantly adjusts in training, therefore:
6) anti fuzzy method layer:In this layer of output 36 as a result, namely there is 36 nodes in the design, each node
Output is exactly the product of upper one layer of all node multiplied by the weight W of upper layer node to this node layerk.Then this layer of each node
Output is:
Wherein, y1To y36Corresponding 36 kinds of roadway characteristics that we set, it is assumed for example that:y1How curved corresponding slight bumps are broken
The case where stone road surface, abrupt slope.A possibility that obtained numerical value y is the decimal between 36 0 to 1, indicates the result of every kind of judgement.
It is output to supplying system server.
2, model training:
Wherein, the activation primitive used can be ReLU function (line rectification function), expression formula be y=max (x,
0)。
Specifically training process is:
1) y will be exported1The case where being set to corresponding how curved slight concave-convex macadam pavement, abrupt slope, other 36 nodes successively class
It pushes away, corresponding all 36 roadway characteristic combined results;
2) a sample (A of sample set is selectedi,Bi), AiFor data BiFor generic);
3) sample is sent into network, calculates the reality output y of network;
4) loss function is calculatedye=1 exports for target;
5) weight matrix W is adjusted according to error loss;
6) each sample is repeated the above process, until the error of entire sample set is no more than prescribed limit, (prediction is accurate
93% or more rate).
3, model calculates:
1) it by trained model, is stored in the real time computation system in middle collection logistics cloud;
2) after four parameters of vehicle parameter enter model, by operation, road surface material (cement or rubble), recessed will be obtained
Convex situation, how curved few curved, four situation all possible combinations results of gradient situation have 36 values altogether;
3) by y1To y36Value and corresponding label pass through network push system server.
Wherein, as a result pushing module includes:
This system is equally a subsystem in cloud, is deployed in several Tomcat servers of cluster, belongs to Web
Service.It, will be by network push to the server of other business after the result that system receives real time computation system push.
Include the following steps:
1) the model calculation that real time computation system is sent 36 roadway characteristic probable values and corresponding characteristic species again
Class label, background logic processing in select its intermediate value maximum one, it is possible to the maximum result of property be used for following industry
Business processing;
2) place longitude and latitude when 1) result and vehicle being sent result, is stored in the distributed data base in cloud together, including
Hbase and Redis, wherein deposit Hbase is counted as off line data analysis, deposit redis is what other needed to calculate in real time
Business provides road condition data;
3) road judging result is pushed directly on the master control display of former vehicle.If driver passes through master control panel
Transmission request, which is checked, is currently located section historical data, then the historical data inquired in HBase is pushed to vehicle again.
The embodiment of the present application provides a kind of road surface characteristic acquisition methods, can be by trained fuzzy neural network
Calculation processing is carried out to sensor information, obtains the corresponding probable value of multiple road feature, it would be possible to be worth maximum road surface characteristic
As acquisition as a result, wherein fuzzy neural network is given the corresponding parameter area of various road surface characteristics, with these parameter areas
Delimit the corresponding fuzzy value of various features, thus overcome directly judged by sensor information caused by user not side
Just the problem of understanding improves the objectivity and accuracy of region feature judgement of satisfying the need.
It obtains system to a kind of road surface characteristic provided by the embodiments of the present application below to be introduced, a kind of road described below
Region feature, which obtains system, can correspond to each other reference with a kind of above-described road surface characteristic acquisition methods.
Specifically, referring to FIG. 2, Fig. 2 is a kind of structure of road surface characteristic acquisition system provided by the embodiment of the present application
Schematic diagram.
The system may include:
Fuzzy neural network training module 100, for the initial fuzzy neural by history pavement behavior data to building
Network is trained processing, obtains objective fuzzy neural network;
Real data obtains module 200, obtains practical pavement behavior data for being analyzed and processed to sensor information;
Fuzzy neural network computing module 300, for using objective fuzzy neural network to practical pavement behavior data into
Row calculation processing obtains preset quantity road surface characteristic probable value;
Road surface characteristic obtains module 400, for determining road surface characteristic according to all road surface characteristic probable values.
Optionally, the fuzzy neural network training module 100 may include:
Initial model construction unit, for constructing initial fuzzy neural network according to the membership function received;
Model training unit, for being trained processing to initial fuzzy neural network by history pavement behavior data,
Obtain objective fuzzy neural network.
Optionally, which obtains module 200, may include:
Road vibration data capture unit obtains road surface vibration for carrying out vibration data processing to vibrating sensor information
Dynamic data;
Turning data acquiring unit, for carrying out turning data processing according to infrared electro gate signal and gyro data,
Obtain traveling turning data;
Gradient acquiring unit, for inclination data and acceleration information to carry out at Gradient forward to sensor
Reason, obtains road gradient data;
Position data acquiring unit, for getting position data by GPS positioning module;
Road surface real data acquiring unit, for by road vibration data, traveling turning data, road gradient data and
Position data is used as practical pavement behavior data.
Optionally, the fuzzy neural network computing module 300 may include:
Neural network input unit is used for road vibration data, traveling turning data, road gradient data and position
Input of the data as objective fuzzy neural network;
Neural computing unit obtains preset quantity road for executing calculating operation to objective fuzzy neural network
Region feature probable value.
The embodiment of the present application also provides a kind of server, including:
Memory, for storing computer program;
Processor, when for executing computer program the step of the realization such as road surface characteristic acquisition methods of above embodiments.
The embodiment of the present application also provides a kind of computer readable storage medium, which is characterized in that computer-readable storage medium
It is stored with computer program in matter, the road surface characteristic acquisition side such as above embodiments is realized when computer program is executed by processor
The step of method.
The computer readable storage medium may include:USB flash disk, mobile hard disk, read-only memory (Read-Only
Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. is various to deposit
Store up the medium of program code.
Each embodiment is described in a progressive manner in specification, the highlights of each of the examples are with other realities
The difference of example is applied, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part illustration
?.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond scope of the present application.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Above to a kind of road surface characteristic acquisition methods provided herein, road surface characteristic obtain system, server and
Computer readable storage medium is described in detail.Principle and embodiment of the specific case to the application used herein
It is expounded, the description of the example is only used to help understand the method for the present application and its core ideas.It should be pointed out that
For those skilled in the art, under the premise of not departing from the application principle, can also to the application into
Row some improvements and modifications, these improvement and modification are also fallen into the protection scope of the claim of this application.
Claims (10)
1. a kind of road surface characteristic acquisition methods, which is characterized in that including:
It is trained processing by initial fuzzy neural network of the history pavement behavior data to building, obtains objective fuzzy nerve
Network;
Sensor information is analyzed and processed to obtain practical pavement behavior data;
Calculation processing is carried out to the practical pavement behavior data using the objective fuzzy neural network, obtains preset quantity
Road surface characteristic probable value;
Road surface characteristic is determined according to all road surface characteristic probable values.
2. road surface characteristic acquisition methods according to claim 1, which is characterized in that by history pavement behavior data to structure
The initial fuzzy neural network built is trained processing, obtains objective fuzzy neural network, including:
The initial fuzzy neural network is constructed according to the membership function received;
Processing is trained to the initial fuzzy neural network by the history pavement behavior data, obtains the target mould
Paste neural network.
3. road surface characteristic acquisition methods according to claim 1 or 2, which is characterized in that analyze sensor information
Processing obtains practical pavement behavior data, including:
Vibration data processing is carried out to vibrating sensor information, obtains road vibration data;
Turning data processing is carried out according to infrared electro gate signal and gyro data, obtains traveling turning data;
To sensor, inclination data and acceleration information carry out Gradient processing forward, obtain road gradient data;
Position data is got by GPS positioning module;
Using the road vibration data, the traveling turning data, the road gradient data and the position data as
The practical pavement behavior data.
4. road surface characteristic acquisition methods according to claim 3, which is characterized in that use the objective fuzzy neural network
Calculation processing is carried out to the practical pavement behavior data, obtains preset quantity road surface characteristic probable value, including:
Using the road vibration data, the traveling turning data, the road gradient data and the position data as
The input of the objective fuzzy neural network;
Calculating operation is executed to the objective fuzzy neural network, obtains the preset quantity road surface characteristic probable value.
5. a kind of road surface characteristic obtains system, which is characterized in that including:
Fuzzy neural network training module, for being carried out by initial fuzzy neural network of the history pavement behavior data to building
Training managing obtains objective fuzzy neural network;
Real data obtains module, obtains practical pavement behavior data for being analyzed and processed to sensor information;
Fuzzy neural network computing module, for using the objective fuzzy neural network to the practical pavement behavior data into
Row calculation processing obtains preset quantity road surface characteristic probable value;
Road surface characteristic obtains module, for determining road surface characteristic according to all road surface characteristic probable values.
6. road surface characteristic according to claim 5 obtains system, which is characterized in that the fuzzy neural network training mould
Block, including:
Initial model construction unit, for constructing the initial fuzzy neural network according to the membership function received;
Model training unit, for being trained place to the initial fuzzy neural network by the history pavement behavior data
Reason, obtains the objective fuzzy neural network.
7. road surface characteristic according to claim 5 or 6 obtains system, which is characterized in that the real data obtains module,
Including:
Road vibration data capture unit obtains road vibration number for carrying out vibration data processing to vibrating sensor information
According to;
Turning data acquiring unit is obtained for carrying out turning data processing according to infrared electro gate signal and gyro data
Travel turning data;
Gradient acquiring unit, for obtaining to sensor inclination data and acceleration information progress Gradient processing forward
To road gradient data;
Position data acquiring unit, for getting position data by GPS positioning module;
Road surface real data acquiring unit is used for the road vibration data, the traveling turning data, the road gradient
Data and the position data are as the practical pavement behavior data.
8. road surface characteristic according to claim 7 obtains system, which is characterized in that the fuzzy neural network calculates mould
Block, including:
Neural network input unit is used for the road vibration data, the traveling turning data, the road gradient data
And input of the position data as the objective fuzzy neural network;
Neural computing unit obtains the preset quantity for executing calculating operation to the objective fuzzy neural network
A road surface characteristic probable value.
9. a kind of server, which is characterized in that including:
Memory, for storing computer program;
Processor realizes that the described in any item road surface characteristics of Claims 1-4 such as obtain when for executing the computer program
The step of method.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program realizes such as Claims 1-4 described in any item road surface characteristic acquisition sides when the computer program is executed by processor
The step of method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810825838.2A CN108898697A (en) | 2018-07-25 | 2018-07-25 | A kind of road surface characteristic acquisition methods and relevant apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810825838.2A CN108898697A (en) | 2018-07-25 | 2018-07-25 | A kind of road surface characteristic acquisition methods and relevant apparatus |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108898697A true CN108898697A (en) | 2018-11-27 |
Family
ID=64352569
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810825838.2A Pending CN108898697A (en) | 2018-07-25 | 2018-07-25 | A kind of road surface characteristic acquisition methods and relevant apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108898697A (en) |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102722989A (en) * | 2012-06-29 | 2012-10-10 | 山东交通学院 | Expressway microclimate traffic early warning method based on fuzzy neural network |
CN103176185A (en) * | 2011-12-26 | 2013-06-26 | 上海汽车集团股份有限公司 | Method and system for detecting road barrier |
CN103921743A (en) * | 2014-05-08 | 2014-07-16 | 长春工业大学 | Automobile running working condition judgment system and judgment method thereof |
CN104864878A (en) * | 2015-05-22 | 2015-08-26 | 汪军 | Electronic map based road condition physical information drawing and inquiring method |
CN105277201A (en) * | 2014-06-20 | 2016-01-27 | 昆山研达电脑科技有限公司 | Road condition providing device and method |
CN105355039A (en) * | 2015-10-23 | 2016-02-24 | 张力 | Road condition information processing method and equipment |
CN105427606A (en) * | 2015-12-24 | 2016-03-23 | 重庆云途交通科技有限公司 | Pavement condition information acquisition and release method |
CN105894843A (en) * | 2015-02-17 | 2016-08-24 | 赫克斯冈技术中心 | Method and system for determining a road condition |
CN106373397A (en) * | 2016-09-28 | 2017-02-01 | 哈尔滨工业大学 | Fuzzy neural network-based remote sensing image road traffic situation analysis method |
CN106647743A (en) * | 2016-10-26 | 2017-05-10 | 纳恩博(北京)科技有限公司 | Control method of electronic device and electronic device |
CN106845547A (en) * | 2017-01-23 | 2017-06-13 | 重庆邮电大学 | A kind of intelligent automobile positioning and road markings identifying system and method based on camera |
CN107705234A (en) * | 2017-08-31 | 2018-02-16 | 昆明理工大学 | A kind of system supervised using vehicle road pavement quality and its monitoring and managing method |
CN108241829A (en) * | 2016-12-23 | 2018-07-03 | 乐视汽车(北京)有限公司 | Vehicle travels image-recognizing method |
-
2018
- 2018-07-25 CN CN201810825838.2A patent/CN108898697A/en active Pending
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103176185A (en) * | 2011-12-26 | 2013-06-26 | 上海汽车集团股份有限公司 | Method and system for detecting road barrier |
CN102722989A (en) * | 2012-06-29 | 2012-10-10 | 山东交通学院 | Expressway microclimate traffic early warning method based on fuzzy neural network |
CN103921743A (en) * | 2014-05-08 | 2014-07-16 | 长春工业大学 | Automobile running working condition judgment system and judgment method thereof |
CN105277201A (en) * | 2014-06-20 | 2016-01-27 | 昆山研达电脑科技有限公司 | Road condition providing device and method |
CN105894843A (en) * | 2015-02-17 | 2016-08-24 | 赫克斯冈技术中心 | Method and system for determining a road condition |
CN104864878A (en) * | 2015-05-22 | 2015-08-26 | 汪军 | Electronic map based road condition physical information drawing and inquiring method |
CN105355039A (en) * | 2015-10-23 | 2016-02-24 | 张力 | Road condition information processing method and equipment |
CN105427606A (en) * | 2015-12-24 | 2016-03-23 | 重庆云途交通科技有限公司 | Pavement condition information acquisition and release method |
CN106373397A (en) * | 2016-09-28 | 2017-02-01 | 哈尔滨工业大学 | Fuzzy neural network-based remote sensing image road traffic situation analysis method |
CN106647743A (en) * | 2016-10-26 | 2017-05-10 | 纳恩博(北京)科技有限公司 | Control method of electronic device and electronic device |
CN108241829A (en) * | 2016-12-23 | 2018-07-03 | 乐视汽车(北京)有限公司 | Vehicle travels image-recognizing method |
CN106845547A (en) * | 2017-01-23 | 2017-06-13 | 重庆邮电大学 | A kind of intelligent automobile positioning and road markings identifying system and method based on camera |
CN107705234A (en) * | 2017-08-31 | 2018-02-16 | 昆明理工大学 | A kind of system supervised using vehicle road pavement quality and its monitoring and managing method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11443148B2 (en) | Multiple stage image based object detection and recognition | |
Mandhare et al. | Intelligent road traffic control system for traffic congestion: a perspective | |
CN111583639B (en) | Road traffic jam early warning method and system | |
CN103927873B (en) | Floating Car and section matching process and the parallel method obtaining real-time road | |
Rahman et al. | Review of microscopic lane-changing models and future research opportunities | |
EP2443418B1 (en) | Methods and systems for creating digital street network database | |
US9222786B2 (en) | Methods and systems for creating digital transportation networks | |
CN109643118B (en) | Influencing a function of a vehicle based on function-related information about the environment of the vehicle | |
CN102278995B (en) | Bayes path planning device and method based on GPS (Global Positioning System) detection | |
CN107945507A (en) | Travel Time Estimation Method and device | |
US11604075B2 (en) | Systems and methods for deriving planned paths for vehicles using path priors | |
CN110415517B (en) | Road congestion accurate early warning system and method based on vehicle running track | |
CN104875740B (en) | For managing the method for following space, main vehicle and following space management unit | |
EP4001843A1 (en) | Method, apparatus, and system for creating doubly-digitised maps | |
DE102022100213A1 (en) | Machine learning based framework for annotation of driveable surfaces | |
JP6914323B2 (en) | Parking lot information management system, parking lot guidance system, parking lot information management program and parking lot guidance program | |
CN117546218A (en) | Distributed multitasking machine learning for traffic prediction | |
Lin et al. | Insights into Travel Pattern Analysis and Demand Prediction: A Data-Driven Approach in Bike-Sharing Systems | |
CN108898697A (en) | A kind of road surface characteristic acquisition methods and relevant apparatus | |
JP2022517454A (en) | Providing additional instructions for difficult maneuvers during navigation | |
EP4012336A2 (en) | Method, apparatus, and system for identifying special areas and cleaning-up map data | |
US11685408B1 (en) | Driving difficulty heat maps for autonomous vehicles | |
CN113420942B (en) | Sanitation vehicle real-time route planning method based on deep Q learning | |
US20220065656A1 (en) | Method, apparatus, and computer program product for generating an automated driving capability map index | |
Habtie et al. | Road Traffic state estimation framework based on hybrid assisted global positioning system and uplink time difference of arrival data collection methods |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181127 |