CN109829573A - A kind of intelligent paths planning method merging user driving habits - Google Patents
A kind of intelligent paths planning method merging user driving habits Download PDFInfo
- Publication number
- CN109829573A CN109829573A CN201910034710.9A CN201910034710A CN109829573A CN 109829573 A CN109829573 A CN 109829573A CN 201910034710 A CN201910034710 A CN 201910034710A CN 109829573 A CN109829573 A CN 109829573A
- Authority
- CN
- China
- Prior art keywords
- road
- user
- time
- sample
- economy
- 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
Landscapes
- Traffic Control Systems (AREA)
Abstract
The present invention relates to a kind of intelligent paths planning methods for merging user driving habits, it is related to vehicle route navigation field, solve the problems, such as that single navigation information frequently can lead to cluster effect and cause crowded elegant so that increasing traffic dispersion pressure and being difficult to the problem of meeting the personalized trip requirements of user comprising: S100: obtain the driving characteristics vector of user;S200: criterion and quantity is carried out to the time of road, economy, comfort property;S300: the road personalization performance number in the user visual field is obtained;S400: carrying out asking processing reciprocal to road personalization performance each in the user visual field, consumes cost as road;And cost value is consumed as standard using this, seek optimal path using self-adaptive genetic operator.The present invention has and can be suitable only for the navigation information of oneself for its customized one group according to the driving habit of user, and the effect of the personalized trip experience of user is increased substantially while alleviating traffic pressure.
Description
Technical field
The present invention relates to vehicle route navigation fields, advise more particularly, to a kind of intelligent path for merging user driving habits
The method of drawing.
Background technique
With flourishing for Hyundai Motor industry, self-driving group grows stronger day by day, and the current pressure of road network is growing day by day,
It is unable to satisfy the growing driving comfort demand of user.Current navigation system is provided using global positioning system (GPS)
Geography information and electronic map detect vehicle location, and then provide popular " distance-steering " formula navigation Service.
However, single navigation information frequently can lead to cluster effect, cause it is crowded elegant, to increase traffic dispersion pressure
Power, simultaneously, it is also difficult to meet the personalized trip requirements of user, there is room for improvement.
Summary of the invention
The object of the present invention is to provide it is a kind of with can according to the driving habit of user be its customized one group only fit
The navigation information of oneself is closed, and increase substantially while alleviating traffic pressure the effect of the personalized trip experience of user
Merge the intelligent paths planning method of user driving habits.
Foregoing invention purpose of the invention has the technical scheme that
A kind of intelligent paths planning method merging user driving habits characterized by comprising
S100: from the Car log track of user itself extract user driving characteristics, and carry out feature clustering screening to
Obtain the driving characteristics vector of user;
S200: in conjunction with each influence factor of road, using the three-dimensional T-S fuzzy neural network of extension to the time of road, economy, relax
Adaptive can be carried out criterion and quantity;
S300: validity screening is carried out to the characteristic item in driving characteristics vector, validity feature item characteristic value remains unchanged, in vain
Characteristic item characteristic value sets 0;The time of acquisition, economy, amenity standards performance are added item by item according to the feature vector value after screening
Power summation, obtains the road personalization performance number in the user visual field;
S400: carrying out asking processing reciprocal to road personalization performance number each in the user visual field, consumes cost as road;And with this
Consumption cost value is standard, seeks optimal path using self-adaptive genetic operator.
It, can be according to the Car log track of user itself by the setting of step S100 by using above-mentioned technical proposal
It determines the driving habit of user, and can determine the driving factor that user is paid attention to, step according to the driving habit of user
The setting of S200 and step S300 can carry out standard to time, economy, comfort property based on three-dimensional T-S fuzzy neural network
Quantization, and personalized road performance value is obtained by weighted average, and determine based on step S300 by the setting of step S400
Optimal path, and current user is recommended, to increase substantially the personalized trip experience of user.
The present invention is further arranged to, and: S100 includes:
S110: a large amount of log track of user is divided into the countless number of dropouts members comprising driving characteristics, by time geography
Space-time path fashion excavates the driving characteristics of log path locus;
S120: limiting and screens feature clustering center, and then using log fuzzy C-means clustering to the arrow for respectively including driving characteristics
Amount member is clustered, and has carried out Cluster Validity screening and quantity statistics to cluster sample;
S130: the driving characteristics vector of user is obtained using time, economy, the effective sample quantity of comfortable cluster centre as standard.
By using above-mentioned technical proposal, can effectively be determined based on the log track of user by the setting of step S110
The driving characteristics of user out, effective driving that the setting of step S120 can filter out user in the driving characteristics of user are special
Sign, can obtain corresponding to each driving characteristics finally by the setting of step S130 after determining the driving characteristics of S120
Driving characteristics vector.
The present invention is further arranged to, and: S110 includes:
S111: Historic space position and vehicle based on vehicle reach at the time of each spatial position corresponds to and establish three-dimensional coordinate
System, arbitrarilyThe running speed at moment, travel accelerationWithThe tangent slope of moment oriented trend curveRelationship such as
Under:,;
S112: comfort level selection: pass through total weighted root mean square accelerationApproximation is made to the level of comfort by bus of human body to retouch
It states,Shown in being defined as follows:, in formula,Respectively represent vehicle advance, water
Flat, vertical direction weighted root mean square acceleration and directivity factor, calculate for evaluating natural environment quality to the path of user
Select the evaluation number of structure:, in formula,RespectivelyThe environment matter of kind factor of natural environment
Volume index, coverage rate, evaluation criterion;Represent the type of natural environment influence factor, including roadside greening, river distribution, sky
Makings amount etc.,It is bigger, indicate that road synthetic environmental quality is better,, whereinConventional is 0.63,According to
User demand adjustment;
Selection of time: being screened by the average speed v arrived at the destination every time,, wherein、According to tool
The adjustment of body situation;
Economical selection: definitionOil consumption, road consumption are respectively represented, is screened by oil consumption, road consumption, wherein。
It, can be according to vehicle Historic space position and vehicle by the setting of step S121 by using above-mentioned technical proposal
It reaches at the time of each spatial position corresponds to and to establish corresponding history coordinate system, S121 can be based on by the setting of step S122
Acquired data effectively carry out road comfort level, road time and road economics and make corresponding selection.
The present invention is further arranged to, and: S120 includes:
S121: using travel speed, unit gas mileage+road consumption, total weighted root mean square acceleration+Multiple environmental quality as standard,
It establishes three-dimensional cluster coordinate system and characterizes each log road vectors member in cluster coordinate systemIt is time for containing, economy, comfortable
Information, wherein definition is vector characteristic member by the minimum unit for containing driving characteristics that the sampling of log track obtains, and is referred to as sweared
Amount member;
S122: according to the selection principle of cluster centre, meet region initialization time, economy, comfortable cluster in constraint condition
Center, it is ensured that three cluster centres moderately separate;
S123: log road vectors member is calculatedTo time, economy, comfortable cluster centreHausdorff distance:(5);
Wherein:For the finite aggregate of vector member;For time, economy, the set of comfortable three classes cluster centre;
S124: each vector member is calculatedTo three kinds of cluster centresDegree of membership:(6);
S125: according to degree of membership renewal time, economy, comfortable cluster centre vector element:(7);
S126: if meeting condition shown in formula (8), otherwise iteration stopping jumps back to S122,(8);Wherein,Respectively represent the cluster centre variable of current time and previous moment;
S127: normalized, as the formula:(9), in formulaIt respectively represents any
The time of sample, economy, comfortable dimension,Respectively represent sampleIn dimensionUnder performance number, minimality
Energy value and maximum performance value,For sampleIn dimension after normalizedUnder performance number.
By using above-mentioned technical proposal, by step S121 to step S127 can effectively to vector member clustered into
Row Effective selection and corresponding quantity statistics carry out corresponding preparation for the calculating of subsequent driving characteristics vector.
The present invention is further arranged to, and: S130 includes:
S131: the effective range of initialization sample randomly selects lesser effective radius of a ball;
S132: using the Euclidean distance of sample local density and more high density point sample as standard, effective radius is carried out by formula (9)
Forward direction adjustment, and finally determine the effective radius of cluster, corresponding plane projection(10),
In:、Respectively represent sampleTo cluster centre, more high density point sampleEuclidean distance;For distance is truncated,
The size of value is related with the average percent of the total sample book of neighbour's number Zhan of sample point, for the feelings of a large amount of flow vector member sample
Condition,Selection have robustness;Respectively sampleLocal density;For sampleTo more high density point sampleMost short Euclidean distance;For relevant boundary parameter, frequently with empirical value;
S133: erasing time, economy, the invalid sample of comfortable cluster centre count each cluster centre effective radiusWithin
Number of samples is denoted as:.;
S134: pass through normalized, difference acquisition time, economy, the distribution of weights of comfort index, calculated
Journey is as the formula (10):(11);The driving style feature vector of user is。
By using above-mentioned technical proposal, can be gone out based on the data screening acquired in S127 by the setting of step S131
Reasonable data, then can effectively to calculate user corresponding for the setting based on step S132, step S133 and step S134
Driving style feature vector is ready for the calculating of subsequent road personalization performance number.
The present invention is further arranged to, and: S200 includes:
S210: fuzzy reasoning is carried out using " IF-THEN " rule format: , in formula,Respectively
Time, comfortably significantly affects set of factors at economy,For indexArbitrarily significantly affect factor
Variable,For indexFuzzy set;For the relevant parameter of fuzzy rule;For fuzzy rule output, former piece network is based on
The relevance grade for calculating fuzzy rule, for any, its degree of membership is asked according to Gauss member function in blurring layer first:(12);In formulaRespectively represent function center and width;,ForFuzzy point
Every number, 7 are enabled, is represented ---, -- ,-, ignore ,+, ++, +++ six kinds of impact effects;
S220: in the relevance grade that fuzzy reasoning layer calculates every rule using operator is even multiplied:(13);
S230: calculating is normalized first in anti fuzzy method layer:(14);Consequent network is identical by three structures
Sub-network constitute, for exporting the fuzzy rule of time, economy, comfort index.Input layer is except inputOutside, it also needs to supplement
The input value 1 of 0th node, for generating the constant term in road performance rating calculation result.Fuzzy reasoning layer calculates each
The consequent of fuzzy rule, it may be assumed that(15);Total output is the weighted sum of each fuzzy rule,
Weighting coefficient is the relevance grade of each rule, exports road performance level results after sharpening;(16)。
By using above-mentioned technical characteristic, the three-dimensional T-S of extension can use by the setting of S210, S220 and S230
Fuzzy neural network carries out criterion and quantity to the time of road, economy, comfort property, carries out for subsequent road personalization performance number
Prepare.
The present invention is further arranged to, and: S300 includes:
S310: the driving style of user is divided into according to the harmony of feature weight distributionThree kinds of classifications, whereinThe species number of validity feature in user characteristics vector is respectively represented, the validity of feature is defined as shown in theorem 1, fixed
Reason 1: for driving characteristics vectorArbitrary elementIf meeting following any condition:
a.;b.;Then claimIt is no for invalid feature
It is then validity feature;
S320: the weighted sum for carrying out validity feature standard path quantized value corresponding to its obtains personalized road performance value w;
S330: to road performanceDeviation standardization is carried out, linear variation is carried out to original performance number, makes the road of the network of communication lines
The codomain range of road property is adjusted to [0,1]:(17);
By using above-mentioned technical proposal, the road individual character in the user visual field can effectively be obtained by the setting of S310 to S330
Change performance number, seeks optimal path for later use self-adaptive genetic operator and be ready.
The present invention is further arranged to: step S400 is further comprising the steps of:
S4a0: unit gas mileage situation of driving a vehicle before obtaining car owner and the sought optimal path out of self-adaptive genetic operator
Milimeter number;
S4b0: with the product of driving the unit gas mileage situation and milimeter number of car owner, as the estimated required amount of gasoline of car owner, and
With real-time detection to automobile in amount of gasoline be compared;
S4c0: if amount of gasoline is less in the automobile that real-time detection arrives, current optimal path is excluded, is calculated by self-adapting ant colony
Method finds out optimal path in residual paths and re-starts differentiation;Conversely, if amount of gasoline is more in the automobile that real-time detection arrives,
Using the optimal path that is currently found as guidance path.
By using above-mentioned technical proposal, can be obtained most by self-adaptive genetic operator by the setting of step S4a0
After good route, horizontal by the oil consumption of step S4b0 and S4c0 reasonable consideration current vehicle, selection is suitble to current oil consumption most
Good path.
The present invention is further arranged to: S4a0 the following steps are included:
S4a1: the unit gas mileage energy of driving unit gas mileage ability under different traffic congestion indexes before car owner is obtained
Force data library;
S4a2: by the optimal path currently cooked up according to current path traffic congestion index different demarcation be several roads
Section, and inquired with the length in each section and with the traffic congestion index of corresponding road section in unit gas mileage capability database
The achievement of unit gas mileage ability phase out is as oil consumption needed for each section;
S4a3: using the summation of oil consumption needed for all sections as dividend, the total length of optimal path is as divisor, with the two
Quotient is as the unit gas mileage ability in optimal path.
Led to by using above-mentioned technical proposal by the current oil consumption situation of the available optimal path of step S4a1, S4a2
The setting for crossing S4a3 can effectively analyze the unit gas mileage ability of optimal path, so that whole path analysis be made to select
It is more accurate.
The present invention is further arranged to: S4c0 is further comprising the steps of:
S4c1: if real-time detection to automobile in amount of gasoline it is less and optimal path can not be inquired in residual paths, with
The gas station closest to destination nearby is inquired in radiation around centered on vehicle, while being cooked up vehicle and being reached gas station most
Short path;
S4c2: it jumps back to S100 and cooks up vehicle by the optimal path of gas station to destination;
S4c3: after cooking up optimal path by S4c2, with the driving unit gas mileage situation of car owner and by gas station
The distance product arrived at the destination is as required oil consumption, as the oil mass refueled needed for gas station needed for car owner.
By using above-mentioned technical proposal, by the setting of step S4C1 can not inquire best route when
It waits, first selects a gas station nearest from destination to refuel, and cook up vehicle by the best road of gas station to destination
Diameter meets everyone personalized trip requirements as far as possible, cooks up most preferably by S4c2 while satisfaction arrives at the destination
After path, distance product that S4c3 is arrived at the destination using the driving unit gas mileage situation of car owner and by gas station is as institute
Oil consumption is needed, as the oil mass refueled needed for gas station needed for car owner, car owner can be allowed to reach purpose according to present case
Minimum volume read-out needed for ground, thus guaranteeing to arrive at the destination simultaneously meeting user, moreover it is possible to meet user individual trip
Demand.
Detailed description of the invention
Fig. 1 is the general frame schematic diagram of the intelligent paths planning method of present invention fusion user driving habits.
Fig. 2 is the step schematic diagram of S100 in Fig. 1.
Fig. 3 is the partial schematic diagram of S400 in Fig. 1.
Fig. 4 is the schematic diagram of S4a0 in Fig. 3.
Fig. 5 is the schematic diagram of S4c0 in Fig. 3.
Fig. 6 is the table schematic diagram of personalized road performance value w in step S320.
Specific embodiment
Below in conjunction with attached drawing, invention is further described in detail.
As shown in Figure 1, a kind of intelligent paths planning method for merging user driving habits, whole to realize that steps are as follows:
S100: extracting the driving characteristics of user from the Car log track of user itself, and carries out feature clustering screening to obtain
The driving characteristics vector of user;S200: in conjunction with each influence factor of road, using the three-dimensional T-S fuzzy neural network of extension to road
The time on road, economy, comfort property carry out criterion and quantity;S300: validity sieve is carried out to the characteristic item in driving characteristics vector
Choosing, validity feature item characteristic value remain unchanged, and invalid characteristic item characteristic value sets 0;According to the feature vector value after screening to acquisition
Time, economy, the weighted sum item by item of amenity standards performance, obtain the user visual field in road personalization performance number;S400:
Road personalization performance number each in the user visual field is carried out asking processing reciprocal, consumes cost as road;And cost is consumed with this
Value is standard, seeks optimal path using self-adaptive genetic operator.
As shown in Figure 2, wherein the process that the driving characteristics vector of user is obtained in step S100 includes the following steps:
S110: a large amount of log track of user is divided into the countless number of dropouts members comprising driving characteristics, by time suboptimal control
Space-time path fashion excavates the driving characteristics of log path locus, and driving characteristics include drive speed, surrounding enviroment quality, oil
It consumes, expense of passing by one's way, road quality, degree of jolting etc.;S120: limiting and screens feature clustering center, and then utilizes log Fuzzy C
Mean cluster clusters the vector member for respectively including driving characteristics, and has carried out Cluster Validity screening and number to cluster sample
Amount statistics;S130: using the time, economy, the effective sample quantity of comfortable cluster centre as standard obtain user driving characteristics to
Amount.
Wherein, the driving according to log path locus is dug in step S110 about the space-time path fashion by time suboptimal control
Feature, what is mainly excavated is road time, economy, comfort, therefore S110 includes: S111: the Historic space based on vehicle
Position and vehicle reach at the time of each spatial position corresponds to and establish three-dimensional system of coordinate, arbitrarilyThe running speed at moment, driving
AccelerationWithThe tangent slope of moment oriented trend curveRelationship it is as follows:,;S112: it relaxes
Appropriateness selection: pass through total weighted root mean square accelerationApproximate description is made to the level of comfort by bus of human body,It is defined as follows
It is shown:, in formula,Vehicle is respectively represented to advance, is horizontal, vertical direction
Weighted root mean square acceleration and directivity factor are calculated for evaluating evaluation of the natural environment quality to the Path selection structure of user
Index:, in formula,RespectivelyThe environmental quality index of kind of factor of natural environment, coverage rate,
Evaluation criterion;Type of natural environment influence factor, including roadside greening, river distribution, air quality etc. are represented,It is bigger,
Indicate that road synthetic environmental quality is better,, whereinConventional is 0.63,It is adjusted according to user demand;When
Between select: screened by the average speed v that arrives at the destination every time,, wherein、As the case may be
Adjustment;Economical selection: definitionOil consumption, road consumption are respectively represented, is screened by oil consumption, road consumption, wherein。
Wherein, the vector member for respectively including driving characteristics is clustered using log fuzzy C-means clustering about in S120,
And Cluster Validity screening and quantity statistics are carried out to cluster sample, the specific steps are as follows: S120 includes:
S121: using travel speed, unit gas mileage+road consumption, total weighted root mean square acceleration+Multiple environmental quality as standard,
It establishes three-dimensional cluster coordinate system and characterizes each log road vectors member in cluster coordinate systemIt is time for containing, economy, comfortable
Information, wherein definition is vector characteristic member by the minimum unit for containing driving characteristics that the sampling of log track obtains, and is referred to as sweared
Amount member.
S122: according to the selection principle of cluster centre, meet region initialization time, economy, comfortable in constraint condition
Cluster centre, it is ensured that three cluster centres moderately separate.
S123: log road vectors member is calculatedTo time, economy, comfortable cluster centreHausdorff distance:(5);Wherein:For the finite aggregate of vector member;For the time, economy,
The set of comfortable three classes cluster centre.
S124: each vector member is calculatedTo three kinds of cluster centresDegree of membership:(6)。
S125: according to degree of membership renewal time, economy, comfortable cluster centre vector element:(7)。
S126: if meeting condition shown in formula (8), otherwise iteration stopping jumps back to S122,(8);Its
In,Respectively represent the cluster centre variable of current time and previous moment.
S127: normalized, as the formula:(9), in formulaIt respectively represents
The time of arbitrary sample, economy, comfortable dimension,Respectively represent sampleIn dimensionUnder performance number, most
Small performance number and maximum performance value,For sampleIn dimension after normalizedUnder performance number.
Log road vectors member feature can be constructed on the basis of step S120 relative to time, economy, comfortable cluster
The coordinate system of the cluster structure at center has step S130, step S130 to specifically include S131 on the basis of coordinate system: initialization
The effective range of sample randomly selects lesser effective radius of a ball;S132: with sample local density and more high density point sample
Euclidean distance be standard, the positive adjustment of effective radius is carried out by formula (9), and finally determine the effective radius clustered, right
The plane projection answered(10), wherein:、Respectively represent sampleTo cluster centre, it is highly denser
Degree point sampleEuclidean distance;For distance is truncated, neighbour's number total sample book of Zhan of the size and sample point of value is averaged
Percentage is related, the case where for a large amount of flow vector member sample,Selection have robustness;Respectively sampleLocal density;For sampleTo more high density point sampleMost short Euclidean distance;For relevant boundary parameter,
Frequently with empirical value;S133: erasing time, economy, the invalid sample of comfortable cluster centre count each cluster centre effective radiusWithin number of samples, be denoted as:.;S134: by normalized, difference acquisition time, comfortably refers to economy
Target distribution of weights, calculating process is as the formula (10):(11);The driving style feature of user
Vector is。
Under the premise of determining the driving style feature vector of user, in conjunction with each influence factor of road, extension is utilized
Three-dimensional T-S fuzzy neural network carries out criterion and quantity to the time of road, economy, comfort property.
Wherein each influence factor of road specifically includes road oneself factor, surrounding enviroment factor, ancillary factors, road itself
Factor includes following factor: road alignment, link length, road intersection quantity, category of roads, road width, company on the way
Continuous turning number, greenbelt isolation, scribing line isolation, split, road circulating density, road flatness, is intersected at road flatness
Crossing number, intersection blocking rate, average traffic delay, disobeys occupancy lane rate at major trunk roads average speed;Wherein, peripheral ring
Border factor includes following factor: roadside greening and landscape arrangement, refueling station, parking lot, traffic light, way-finding sign master
Arterial highway bright light rate, parking guidance, system for traffic guiding setting, rubbish cleaning room and the setting at operating station, cross-road passages for pedestrians are set
Set the influence of rate, yield signs, big density populations Circulation Area.
Wherein, ancillary factors can be divided into traffic legal system regulation factor, Driver's Factors, other dynamic factors, traffic method
Preparation method rule factor includes following factor: the limitation of high speed is walked during 1. habits;2. speed limit, limit for height, freight weight limit, limitation enter;3. uniline
Line setting, odd-and-even license plate rule;4, the influence, limitation of divided roadway transit time or no through traffic region and type of vehicle are drawn
It is fixed.Driver's Factors include following factor: the limitation of 1. driving abilities;2, the driving age limits.Other dynamic factors include driver
The fluctuation of mood, emergency (large-scale public activity, major traffic accidents, temporary traffic control) road occupying rate, bad weather,
Natural calamity, road construction, urban planning.
Criterion and quantity is carried out to the time of road, economy, comfort property using the three-dimensional T-S fuzzy neural network of extension
Step S200 comprising the following specific steps
S210: fuzzy reasoning is carried out using " IF-THEN " rule format: , in formula,Respectively
Time, comfortably significantly affects set of factors at economy,For indexArbitrarily significantly affect factor
Variable,For indexFuzzy set;For the relevant parameter of fuzzy rule;For fuzzy rule output, former piece network is based on
The relevance grade for calculating fuzzy rule, for any, its degree of membership is asked according to Gauss member function in blurring layer first:(12);In formulaRespectively represent function center and width;,ForFuzzy point
Every number, 7 are enabled, is represented ---, -- ,-, ignore ,+, ++, +++ six kinds of impact effects.
S220: in the relevance grade that fuzzy reasoning layer calculates every rule using operator is even multiplied:(13)。
S230: calculating is normalized first in anti fuzzy method layer:(14);Consequent network is by three structure phases
Same sub-network is constituted, for exporting the fuzzy rule of time, economy, comfort index.Input layer is except inputOutside, it also needs to mend
The input value 1 for filling the 0th node, for generating the constant term in road performance rating calculation result.Fuzzy reasoning layer calculates every
The consequent of one fuzzy rule, it may be assumed that(15);Total output is the weighting of each fuzzy rule
It is the relevance grade of each rule with, weighting coefficient, exports road performance level results after sharpening;
(16)。
It completes to quantify the time to road, economy, amenity standards, subsequent progress characteristic item carries out validity screening
And the acquisition of road personalization performance number, it is specific as step S300, step S300 specifically comprise the following steps:
S310: the driving style of user is divided into according to the harmony of feature weight distributionThree kinds of classifications, whereinThe species number of validity feature in user characteristics vector is respectively represented, the validity of feature is defined as shown in theorem 1, fixed
Reason 1: for driving characteristics vectorArbitrary elementIf meeting following any condition:
a.;b.;Then claimIt is no for invalid feature
It is then validity feature;
S310: the driving style of user is divided into according to the harmony of feature weight distributionThree kinds of classifications, whereinThe species number of validity feature in user characteristics vector is respectively represented, the validity of feature is defined as shown in theorem 1, fixed
Reason 1: for driving characteristics vectorArbitrary elementIf meeting following any condition:
a.;b.;Then claimIt is no for invalid feature
It is then validity feature.
S320: the weighted sum for carrying out validity feature standard path quantized value corresponding to its obtains personalized road performance value
W, it is specific as shown in Figure 6.
S330: to road performanceDeviation standardization is carried out, linear variation is carried out to original performance number, makes the network of communication lines
The codomain range of road is adjusted to [0,1]:(17)。
Wherein, step S400 specifically comprises the following steps: S410: carrying out linear turn between road performance and road consumption
It changes, as shown in formula (18):(18);S420: with consumption valueIt is solved any two in road traffic net for index
Personalized optimal path between point, optimum path problems are described as follows: it is assumed thatIt is oriented for a road network
Figure, whereinIt is branch's point set of road network,It isRoad set,Disappear for road
Consumption set, be arbitrarily connected branch pointBetween directed pathRoad consumption be denoted as, wherein.For, optimal path is exactly to find in road network digraph from pointIt arrivesRoad
Road consumption and the smallest path.
As shown in figure 3, further consider car owner about automobile unit gas mileage ability, plan best road
Milimeter number and automobile itself institute the band oil consumption that diameter arrives at the destination determine whether to continue to use through self-adaptive genetic operator institute
The optimal path sought, the specific steps are as follows: S4a0: drive a vehicle before obtaining car owner unit gas mileage situation and adaptive ant
The milimeter number of the sought optimal path out of group's algorithm;S4b0: with the driving unit gas mileage situation of car owner and milimeter number
Product, as car owner it is estimated needed for amount of gasoline, and with real-time detection to automobile in amount of gasoline be compared;S4c0: if in real time
Amount of gasoline is less in the automobile detected, then current optimal path is excluded, through self-adaptive genetic operator in residual paths
It finds out optimal path and re-starts differentiation;Conversely, if amount of gasoline is more in the automobile that real-time detection arrives, with what is currently found
Optimal path is as guidance path.
As shown in figure 4, the car owner's oil consumption ability further contemplated under different traffic congestion indexes is different, therefore S4a0
The following steps are included: S4a1: obtaining the unit of driving unit gas mileage ability under different traffic congestion indexes before car owner
Gas mileage capability database;S4a2: by the optimal path currently cooked up according to current path traffic congestion index not
It is same to be divided into several sections, and with the length in each section and with the traffic congestion index of corresponding road section in unit gas mileage
The achievement for the unit gas mileage ability phase that capability database is inquired is as oil consumption needed for each section;S4a3: with all
The summation of oil consumption needed for section is as dividend, and the total length of optimal path is as divisor, using the quotient of the two as optimal path
In unit gas mileage ability.
As shown in figure 5, further contemplating can not sought by self-adaptive genetic operator to reasonable optimal path
When, it pays the utmost attention at this time to automobile fuel filler, automobile gasoline is added to and comes back to S100 again after suitable amount and cooks up by vapour
Vehicle gas station is to the best route of destination, therefore S4c0 is further comprising the steps of: S4c1: if vapour in the automobile that real-time detection arrives
Oil mass is less and optimal path can not be inquired in residual paths, then radiation inquiry nearby most connects around centered on vehicle
The gas station of nearly destination, while cooking up the shortest path that vehicle reaches gas station;S4c2: it jumps back to S100 and cooks up vehicle
By the optimal path of gas station to destination;S4c3: after cooking up optimal path by S4c2, with the driving unit of car owner
Gas mileage situation and the distance product arrived at the destination by gas station as required oil consumption, as needed for car owner in institute, gas station
The oil mass that need to be refueled.
The embodiment of present embodiment is presently preferred embodiments of the present invention, not limits protection of the invention according to this
Range, therefore: the equivalence changes that all structures under this invention, shape, principle are done, should all be covered by protection scope of the present invention it
It is interior.
Claims (10)
1. a kind of intelligent paths planning method for merging user driving habits characterized by comprising
S100: from the Car log track of user itself extract user driving characteristics, and carry out feature clustering screening to
Obtain the driving characteristics vector of user;
S200: in conjunction with each influence factor of road, using the three-dimensional T-S fuzzy neural network of extension to the time of road, economy, relax
Adaptive can be carried out criterion and quantity;
S300: validity screening is carried out to the characteristic item in driving characteristics vector, validity feature item characteristic value remains unchanged, in vain
Characteristic item characteristic value sets 0;The time of acquisition, economy, amenity standards performance are added item by item according to the feature vector value after screening
Power summation, obtains the road personalization performance number in the user visual field;
S400: carrying out asking processing reciprocal to road personalization performance number each in the user visual field, consumes cost as road;And with this
Consumption cost value is standard, seeks optimal path using self-adaptive genetic operator.
2. the intelligent paths planning method of fusion user driving habits according to claim 1, which is characterized in that S100 packet
It includes:
S110: a large amount of log track of user is divided into the countless number of dropouts members comprising driving characteristics, by time geography
Space-time path fashion excavates the driving characteristics of log path locus;
S120: limiting and screens feature clustering center, and then using log fuzzy C-means clustering to the arrow for respectively including driving characteristics
Amount member is clustered, and has carried out Cluster Validity screening and quantity statistics to cluster sample;
S130: the driving characteristics vector of user is obtained using time, economy, the effective sample quantity of comfortable cluster centre as standard.
3. the intelligent paths planning method of fusion user driving habits according to claim 2, which is characterized in that S110 packet
It includes:
S111: Historic space position and vehicle based on vehicle reach at the time of each spatial position corresponds to and establish three-dimensional coordinate
System, arbitrarilyThe running speed at moment, travel accelerationWithThe tangent slope of moment oriented trend curveRelationship such as
Under:,;
S112: comfort level selection: pass through total weighted root mean square accelerationApproximate description is made to the level of comfort by bus of human body,Shown in being defined as follows:, in formula,Respectively represent vehicle advance, be horizontal,
The weighted root mean square acceleration and directivity factor of vertical direction are calculated for evaluating natural environment quality to the Path selection of user
The evaluation number of structure:, in formula,RespectivelyThe environmental quality of kind factor of natural environment refers to
Number, coverage rate, evaluation criterion;Represent the type of natural environment influence factor, including roadside greening, river distribution, air matter
Amount etc.,It is bigger, indicate that road synthetic environmental quality is better,, whereinConventional is 0.63,According to user
Demand adjustment;
Selection of time: being screened by the average speed v arrived at the destination every time,, wherein、According to tool
The adjustment of body situation;
Economical selection: definitionOil consumption, road consumption are respectively represented, is screened by oil consumption, road consumption, wherein。
4. the intelligent paths planning method of fusion user driving habits according to claim 3, which is characterized in that S120 packet
It includes:
S121: using travel speed, unit gas mileage+road consumption, total weighted root mean square acceleration+Multiple environmental quality as standard,
It establishes three-dimensional cluster coordinate system and characterizes each log road vectors member in cluster coordinate systemIt is time for containing, economy, comfortable
Information, wherein definition is vector characteristic member by the minimum unit for containing driving characteristics that the sampling of log track obtains, and is referred to as sweared
Amount member;
S122: according to the selection principle of cluster centre, meet region initialization time, economy, comfortable cluster in constraint condition
Center, it is ensured that three cluster centres moderately separate;
S123: log road vectors member is calculatedTo time, economy, comfortable cluster centreHausdorff distance:(5);
Wherein:For the finite aggregate of vector member;For time, economy, the set of comfortable three classes cluster centre;
S124: each vector member is calculatedTo three kinds of cluster centresDegree of membership:(6);
S125: according to degree of membership renewal time, economy, comfortable cluster centre vector element:(7);
S126: if meeting condition shown in formula (8), otherwise iteration stopping jumps back to S122,(8);Wherein,Respectively represent the cluster centre variable of current time and previous moment;
S127: normalized, as the formula:(9), in formulaIt respectively represents any
The time of sample, economy, comfortable dimension,Respectively represent sampleIn dimensionUnder performance number, minimality
Energy value and maximum performance value,For sampleIn dimension after normalizedUnder performance number.
5. the intelligent paths planning method of fusion user driving habits according to claim 4, which is characterized in that S130 packet
It includes:
S131: the effective range of initialization sample randomly selects lesser effective radius of a ball;
S132: using the Euclidean distance of sample local density and more high density point sample as standard, effective radius is carried out by formula (9)
Forward direction adjustment, and finally determine the effective radius of cluster, corresponding plane projection(10),
In:、Respectively represent sampleTo cluster centre, more high density point sampleEuclidean distance;For distance is truncated,
The size of value is related with the average percent of the total sample book of neighbour's number Zhan of sample point, for the feelings of a large amount of flow vector member sample
Condition,Selection have robustness;Respectively sampleLocal density;For sampleTo more high density point sampleMost short Euclidean distance;For relevant boundary parameter, frequently with empirical value;
S133: erasing time, economy, the invalid sample of comfortable cluster centre count each cluster centre effective radiusWithin sample
This number is denoted as:;
S134: pass through normalized, difference acquisition time, economy, the distribution of weights of comfort index, calculating process
It is as the formula (10):(11);The driving style feature vector of user is。
6. the intelligent paths planning method of fusion user driving habits according to claim 5, which is characterized in that S200 packet
It includes:
S210: fuzzy reasoning is carried out using " IF-THEN " rule format: , in formula,Respectively
Time, comfortably significantly affects set of factors at economy,For indexArbitrarily significantly affect factor
Variable,For indexFuzzy set;For the relevant parameter of fuzzy rule;For fuzzy rule output, former piece network is based on
The relevance grade for calculating fuzzy rule, for any, its degree of membership is asked according to Gauss member function in blurring layer first:(12);In formulaRespectively represent function center and width;,ForFuzzy point
Every number, 7 are enabled, is represented ---, -- ,-, ignore ,+, ++, +++ six kinds of impact effects;
S220: in the relevance grade that fuzzy reasoning layer calculates every rule using operator is even multiplied:(13);
S230: calculating is normalized first in anti fuzzy method layer:(14);Consequent network is identical by three structures
Sub-network is constituted, and for exporting the fuzzy rule of time, economy, comfort index, input layer is except inputOutside, supplement the 0th is also needed
The input value 1 of a node, for generating the constant term in road performance rating calculation result, fuzzy reasoning layer calculates each mould
Paste the consequent of rule, it may be assumed that(15);Total output is the weighted sum of each fuzzy rule, weighting
Coefficient is the relevance grade of each rule, exports road performance level results after sharpening;(16)。
7. the intelligent paths planning method of fusion user driving habits according to claim 6, which is characterized in that S300 packet
It includes:
S310: the driving style of user is divided into according to the harmony of feature weight distributionThree kinds of classifications, whereinThe species number of validity feature in user characteristics vector is respectively represented, the validity of feature is defined as shown in theorem 1, fixed
Reason 1: for driving characteristics vectorArbitrary elementIf meeting following any condition:
a.;b.;Then claimIt is no for invalid feature
It is then validity feature;
S320: the weighted sum for carrying out validity feature standard path quantized value corresponding to its obtains personalized road performance value w;
S330: to road performanceDeviation standardization is carried out, linear variation is carried out to original performance number, makes the road of the network of communication lines
The codomain range of property is adjusted to [0,1]:(17)。
8. the intelligent paths planning method of fusion user driving habits according to claim 7, which is characterized in that step
S400 is further comprising the steps of:
S4a0: unit gas mileage situation of driving a vehicle before obtaining car owner and the sought optimal path out of self-adaptive genetic operator
Milimeter number;
S4b0: with the product of driving the unit gas mileage situation and milimeter number of car owner, as the estimated required amount of gasoline of car owner, and
With real-time detection to automobile in amount of gasoline be compared;
S4c0: if amount of gasoline is less in the automobile that real-time detection arrives, current optimal path is excluded, is calculated by self-adapting ant colony
Method finds out optimal path in residual paths and re-starts differentiation;Conversely, if amount of gasoline is more in the automobile that real-time detection arrives,
Using the optimal path that is currently found as guidance path.
9. the intelligent paths planning method of fusion user driving habits according to claim 8, which is characterized in that S4a0 packet
Include following steps:
S4a1: the unit gas mileage energy of driving unit gas mileage ability under different traffic congestion indexes before car owner is obtained
Force data library;
S4a2: by the optimal path currently cooked up according to current path traffic congestion index different demarcation be several roads
Section, and inquired with the length in each section and with the traffic congestion index of corresponding road section in unit gas mileage capability database
The achievement of unit gas mileage ability phase out is as oil consumption needed for each section;
S4a3: using the summation of oil consumption needed for all sections as dividend, the total length of optimal path is as divisor, with the two
Quotient is as the unit gas mileage ability in optimal path.
10. the intelligent paths planning method of fusion user driving habits according to claim 9, it is characterised in that: S4c0
It is further comprising the steps of:
S4c1: if real-time detection to automobile in amount of gasoline it is less and optimal path can not be inquired in residual paths, with
The gas station closest to destination nearby is inquired in radiation around centered on vehicle, while being cooked up vehicle and being reached gas station most
Short path;
S4c2: it jumps back to S100 and cooks up vehicle by the optimal path of gas station to destination;
S4c3: after cooking up optimal path by S4c2, with the driving unit gas mileage situation of car owner and by gas station
The distance product of optimal path is arrived at the destination as required oil consumption, as the oil mass refueled needed for car owner in gas station.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910034710.9A CN109829573A (en) | 2019-01-15 | 2019-01-15 | A kind of intelligent paths planning method merging user driving habits |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910034710.9A CN109829573A (en) | 2019-01-15 | 2019-01-15 | A kind of intelligent paths planning method merging user driving habits |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109829573A true CN109829573A (en) | 2019-05-31 |
Family
ID=66860899
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910034710.9A Pending CN109829573A (en) | 2019-01-15 | 2019-01-15 | A kind of intelligent paths planning method merging user driving habits |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109829573A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110398960A (en) * | 2019-07-08 | 2019-11-01 | 浙江吉利汽车研究院有限公司 | A kind of paths planning method of intelligent driving, device and equipment |
CN111080202A (en) * | 2019-12-13 | 2020-04-28 | 拉货宝网络科技有限责任公司 | Efficiency management method and system of working vehicle for oil quantity saving |
CN111860928A (en) * | 2020-01-21 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Method and device for making driving plan, electronic equipment and storage medium |
CN112686444A (en) * | 2020-12-29 | 2021-04-20 | 北京中交兴路信息科技有限公司 | Truck route planning method and device based on similar vehicle driving characteristic data |
CN113919188A (en) * | 2021-07-27 | 2022-01-11 | 北京理工大学 | Relay unmanned aerial vehicle path planning method based on context-MAB |
CN114137973A (en) * | 2021-11-26 | 2022-03-04 | 湖北亿纬动力有限公司 | Path planning method, device, equipment and storage medium |
CN114485709A (en) * | 2022-01-19 | 2022-05-13 | 武汉大学 | Vehicle navigation path planning method |
CN114518124A (en) * | 2022-04-20 | 2022-05-20 | 四川九通智路科技有限公司 | Vehicle driving route planning method and system |
CN115578879A (en) * | 2022-09-14 | 2023-01-06 | 江苏天一航空工业股份有限公司 | Vehicle optimal path navigation control method and system based on Internet of things |
CN117650844A (en) * | 2024-01-29 | 2024-03-05 | 江西开放大学 | VLC relay system safety performance optimization method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102878998A (en) * | 2011-07-13 | 2013-01-16 | 上海博泰悦臻电子设备制造有限公司 | Vehicle fueling prompting method based on path programming |
CN105679037A (en) * | 2016-03-29 | 2016-06-15 | 重庆云途交通科技有限公司 | Dynamic path planning method based on user travel habit |
US20170138744A1 (en) * | 2015-11-12 | 2017-05-18 | International Business Machines Corporation | Personalized optimal travel route planning |
CN107270925A (en) * | 2017-07-27 | 2017-10-20 | 三星电子(中国)研发中心 | A kind of user's Vehicular navigation system, device and method |
CN107305128A (en) * | 2016-04-21 | 2017-10-31 | 斑马网络技术有限公司 | Navigation processing method, navigation equipment, vehicles control device and operating system |
CN107490386A (en) * | 2017-08-29 | 2017-12-19 | 广州小鹏汽车科技有限公司 | A kind of method and system for planning of electric automobile optimal path and drive manner |
-
2019
- 2019-01-15 CN CN201910034710.9A patent/CN109829573A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102878998A (en) * | 2011-07-13 | 2013-01-16 | 上海博泰悦臻电子设备制造有限公司 | Vehicle fueling prompting method based on path programming |
US20170138744A1 (en) * | 2015-11-12 | 2017-05-18 | International Business Machines Corporation | Personalized optimal travel route planning |
CN105679037A (en) * | 2016-03-29 | 2016-06-15 | 重庆云途交通科技有限公司 | Dynamic path planning method based on user travel habit |
CN107305128A (en) * | 2016-04-21 | 2017-10-31 | 斑马网络技术有限公司 | Navigation processing method, navigation equipment, vehicles control device and operating system |
CN107270925A (en) * | 2017-07-27 | 2017-10-20 | 三星电子(中国)研发中心 | A kind of user's Vehicular navigation system, device and method |
CN107490386A (en) * | 2017-08-29 | 2017-12-19 | 广州小鹏汽车科技有限公司 | A kind of method and system for planning of electric automobile optimal path and drive manner |
Non-Patent Citations (1)
Title |
---|
张晓燕: "融合用户习惯的车辆导航路径规划策略研究", 《中国优秀硕士学位论文全文数据库(工程科技Ⅱ辑)》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110398960A (en) * | 2019-07-08 | 2019-11-01 | 浙江吉利汽车研究院有限公司 | A kind of paths planning method of intelligent driving, device and equipment |
CN111080202A (en) * | 2019-12-13 | 2020-04-28 | 拉货宝网络科技有限责任公司 | Efficiency management method and system of working vehicle for oil quantity saving |
CN111860928A (en) * | 2020-01-21 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Method and device for making driving plan, electronic equipment and storage medium |
CN112686444B (en) * | 2020-12-29 | 2024-02-09 | 北京中交兴路信息科技有限公司 | Truck route planning method and device based on similar vehicle driving characteristic data |
CN112686444A (en) * | 2020-12-29 | 2021-04-20 | 北京中交兴路信息科技有限公司 | Truck route planning method and device based on similar vehicle driving characteristic data |
CN113919188A (en) * | 2021-07-27 | 2022-01-11 | 北京理工大学 | Relay unmanned aerial vehicle path planning method based on context-MAB |
CN113919188B (en) * | 2021-07-27 | 2024-05-24 | 北京理工大学 | Relay unmanned aerial vehicle path planning method based on context-MAB |
CN114137973A (en) * | 2021-11-26 | 2022-03-04 | 湖北亿纬动力有限公司 | Path planning method, device, equipment and storage medium |
CN114137973B (en) * | 2021-11-26 | 2024-05-07 | 湖北亿纬动力有限公司 | Path planning method, device, equipment and storage medium |
CN114485709A (en) * | 2022-01-19 | 2022-05-13 | 武汉大学 | Vehicle navigation path planning method |
CN114518124A (en) * | 2022-04-20 | 2022-05-20 | 四川九通智路科技有限公司 | Vehicle driving route planning method and system |
CN115578879B (en) * | 2022-09-14 | 2023-12-08 | 江苏天一航空工业股份有限公司 | Vehicle optimal path navigation control method and system based on Internet of things |
CN115578879A (en) * | 2022-09-14 | 2023-01-06 | 江苏天一航空工业股份有限公司 | Vehicle optimal path navigation control method and system based on Internet of things |
CN117650844A (en) * | 2024-01-29 | 2024-03-05 | 江西开放大学 | VLC relay system safety performance optimization method |
CN117650844B (en) * | 2024-01-29 | 2024-04-26 | 江西开放大学 | VLC relay system safety performance optimization method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109829573A (en) | A kind of intelligent paths planning method merging user driving habits | |
CN109311478B (en) | Automatic driving vehicle speed control method based on comfort level | |
CN105865472B (en) | A kind of navigation method based on optimum oil consumption | |
CN107340772B (en) | Unmanned-oriented anthropomorphic reference trajectory planning method | |
CN104024800B (en) | Can coverage area computing device, method | |
CN103674046B (en) | Vehicle travel is planned | |
US20230113298A1 (en) | Autonomous vehicle fleet management for reduced traffic congestion | |
CN104157139B (en) | A kind of traffic congestion Forecasting Methodology and method for visualizing | |
CN103927873B (en) | Floating Car and section matching process and the parallel method obtaining real-time road | |
EP3555569A1 (en) | Vehicle route guidance | |
CN107305131A (en) | Navigation optimization centered on node | |
CN105355049B (en) | A kind of highway evaluation of running status method based on macroscopical parent map | |
CN103366557A (en) | Traffic congestion evaluation method based on congestion index | |
CN101694706A (en) | Modeling method of characteristics of population space-time dynamic moving based on multisource data fusion | |
US20200116515A1 (en) | Autonomous Vehicle Capability and Operational Domain Evaluation and Selection for Improved Computational Resource Usage | |
CN109191849A (en) | A kind of traffic congestion Duration Prediction method based on multi-source data feature extraction | |
CN111879329B (en) | Customized public transport passable shortest path calculation method based on A-x algorithm | |
Henson | Levels of service for pedestrians | |
US20220005140A1 (en) | Method, apparatus, and system for providing a contextually relevant vehicle comparison | |
CN110363985A (en) | A kind of traffic data analyzing method, apparatus, storage medium and equipment | |
WO2020120789A1 (en) | Method and system for vehicle routing on a road segment | |
CN104574966B (en) | Variable information identity device and variable information identification method | |
CN108710996A (en) | Gather region hotel addressing appraisal procedure in hotel based on tourism trip time and space usage | |
Si et al. | Optimization model and algorithm for mixed traffic of urban road network with flow interference | |
CN103065460B (en) | Multi-vehicle speed fusion method and device based on floating vehicle data |
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 |
Application publication date: 20190531 |
|
RJ01 | Rejection of invention patent application after publication |