CN105716620B - A kind of air navigation aid based on cloud computing and big data - Google Patents
A kind of air navigation aid based on cloud computing and big data Download PDFInfo
- Publication number
- CN105716620B CN105716620B CN201610150671.5A CN201610150671A CN105716620B CN 105716620 B CN105716620 B CN 105716620B CN 201610150671 A CN201610150671 A CN 201610150671A CN 105716620 B CN105716620 B CN 105716620B
- Authority
- CN
- China
- Prior art keywords
- mrow
- traffic
- moment
- prediction
- vehicle
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3453—Special cost functions, i.e. other than distance or default speed limit of road segments
- G01C21/3492—Special cost functions, i.e. other than distance or default speed limit of road segments employing speed data or traffic data, e.g. real-time or historical
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
- Navigation (AREA)
Abstract
A kind of air navigation aid based on cloud computing and big data, belongs to technical field of vehicle navigation.The running data that the present invention is constantly uploaded by storing vehicle, and the traffic behavior preset value of vehicle driving feature and path under different traffics is thus parsed, and calculate the new breath square of traffic normalization at outbound path each moment under each situation and store.Then when user is navigated, road prediction and roading service are carried out in real time, and timing sends navigation instruction information to navigation terminal.Navigation terminal, there is provided vehicle operation data is to terminal server and receives the navigation instruction information from server terminal.The present invention is predicted ineligible traffic tentative prediction result in real-time operation, substantially reduces the burden of server terminal real-time operation again.The future traffic condition in path can be made quick and more accurately predicted, therefrom select the air navigation aid based on cloud computing and big data of optimal guidance path.
Description
Technical field
The present invention relates to a kind of technical field of vehicle navigation, specifically a kind of navigation based on cloud computing and big data
Method.
Background technology
There are two kinds of navigation system to be widely used in life at present, one kind is that it includes locating module, offline using offline navigation
Electronic map and path calculation module.After satellite fix, road information, path computing are provided by offline electronic map
Module calculates guidance path.And the historical traffic by storing is recorded to avoid the congested link in some periods,
This navigation system when road conditions change is little with good reply effect, but due to it is no use real-time traffic
Information is difficult to play same effect as reference in the changeable city of condition of road surface.
It is now more that it possesses gsm module using second of navigation system, by GPRS connections network center,
Network center updates Real-time Traffic Information, and Real-time Traffic Information is added within limit of consideration, then carries out the choosing of guidance path
Select.Solve the problems, such as that the first navigation system does not account for Real-time Traffic Information and navigation effect is deteriorated.
But there is also a new problem for this navigation system:Transport information can be over time change and change, according to
The optimal route selection made according to traffic for the moment is often changed after a while.
To solve this problem, the Chinese patent of Application No. 201410461054.8 discloses one kind and is based on transport information
The navigation system and method for prediction, the system is according to current Real-time Traffic Information and has carried out the vehicle route of Path selection
Information, the traffic of Future Path is predicted, selects optimal path.This invents the vehicle pair that take into account path planning
The influence of subsequent vehicle navigation, avoids all vehicles from all selecting to cause optimal path most to be gathered around on the contrary in the optimal path of present moment
The problem of stifled.But some vehicles are not accounted for when being familiar with sections of road, do not navigated to caused by transport information
Influence, and the selection of vehicle of the road condition change on having been navigated influences so that the predicted distortion of the traffic in path, lead
Effect of navigating is deteriorated.
The content of the invention
Data storage capacities and computing capability for navigation terminal in the prior art can not accomplish to satisfy the need in a short time
The deficiencies of footpath situation is made a prediction, the present invention provide it is a kind of the future traffic condition in path can be made it is quick and more accurately pre-
Survey, the navigation methods and systems based on cloud computing and big data for therefrom selecting optimal guidance path.
In order to solve the above technical problems, the technical solution adopted by the present invention is:
A kind of air navigation aid based on cloud computing and big data of the present invention, comprises the following steps:
The first step:The vehicle of whole city uploads running data information in real time;
Second step:Server terminal stores and carries out algorithm simulation, draws the relevant information of whole city path and vehicle:
3rd step:After navigation terminal inputs destination, server terminal receive information simultaneously calculates all by present bit
The guidance path to destination is put, gather on guidance path and employ the information of vehicles of navigation system, and to by usually
The maximum probability run routing information of vehicle of the possibility drawn on guidance path is monitored, thus makes the tentative prediction of traffic;
4th step:The real-time traffic behavior of backward prediction based on tentative prediction, draw the new breath of Reverse Commuting normalization
Square;
5th step:The new breath square of traffic normalization in history in the same pathway traffic situation of synchronization is handed over reverse
Lead to the new breath square of normalization and make ratio λ, set threshold values L, meet in the period of prediction whole shared by number at the time of condition λ < L
The ratio of moment number is more than 50%, directly exports the preliminary traffic forecast of this period, and optimal path is selected after calculating;
If within the period of the traffic of prediction, the transport information of prediction differed compared with Real-time Traffic Information compared with
Greatly, the average speed for differing the larger vehicle flowrate for referring to prediction and car is more than or less than the average speed of real-time vehicle flowrate and car
Numerical value more than more than 6, carry out the 3rd step start;
6th step:If meet the ratio of whole moment numbers shared by number at the time of condition λ < L in the period of prediction for the first time
Less than 50%, then according to real-time traffic states, using the traffic behavior preset value after calibration, in terms of standard Kalman filtering algorithm
The traffic status prediction for drawing following a period of time is calculated, returns to the 5th step;In the period of the prediction drawn since for the second time
Meet that the ratio of whole moment numbers shared by number at the time of condition λ < L is less than 50%, then newly cease covariance by adjusting the k+1 moment
To carry out, to the traffic status prediction in following setting time, returning to the 5th step;
After setting time is reached, then above-mentioned steps are re-started:When the prediction result inaccuracy in setting time, then
The 3rd step is carried out to predict again.
Further, the relevant information of the whole city path and vehicle, including:
A1:Navigation terminal collects historical traffic state value of the whole city per paths, and free pathway is immediately arrived at according to statistics
In the case of car saturation;
A2:The traffic behavior predictive value of every paths, that is, the traffic after calibrating are calculated according to historical traffic state value
State preset value;
A3:The traffic normalization that per moment on every paths is calculated according to resulting historical traffic state value is new
Breath square simultaneously stores;
A4:Path and the storage of the daily maximum probability traveling of vehicle are drawn according to FP-tree method.
Further, the historical traffic state value refers to the average speed of vehicle flowrate and car.
Further, the computational methods of the step A2 traffic behavior preset values are:
Using the historical traffic state value at same path daily k moment and k+1 moment as model (k=0,1,2,3,4,5,
6 ... n),
Traffic behavior equation:
Wherein:F (k) is state-transition matrix, and G (k) is input control item matrix, and V (k) is process noise sequence, X (k)
The traffic behavior value at history kth moment is represented, u (k) represents control signal,Represent what is predicted based on the k moment
The predicted value at k+1 moment;X represents historical traffic state value, while using the external measurement devices such as monitor on road, ground sense
The traffic behavior value of the measurements such as coil is equally historical data with x as system measurement z;It is pre- that the amount of band " ^ " all represents system
Measured value, the amount of upper right corner band " ' " represent transposed matrix, and the amount of upper right corner band " -1 " represents inverse matrix;
Traffic measurement equation:
The systematic survey predicted value at k+1 moment predicted based on the k moment is represented, H (k+1) is measurement square
Battle array, W (k+1) are measurement noise sequence;
Filter the computing newly ceased:
Z (k+1) represents the systematic survey history value at k+1 moment,Represent the systematic survey predicted value at k+1 moment
Covariance one-step prediction:
P (k+1 | k)=F (k) P (k) F ' (k)+Q (k) (4)
F ' (k) represents F (k) transposed matrix, and Q (k) represents covariance
New breath covariance:
S (k+1)=H (k+1) P (k+1 | k) H ' (k+1)+R (k+1) (5)
S (k+1) newly ceases the k+1 moment that covariance, H (k+1) measurement matrix, P (k+1 | k) are predicted based on the k moment
Covariance one-step prediction, the transposed matrix of H ' (k+1) measurement matrix, R (k+1) and Q (k) represent different covariances
Gain:
K (k+1)=P (k+1 | k) H ' (k+1) S-1(k+1) (6)
K (k+1) represents the gain at the k+1 moment for next step
Traffic behavior renewal equation:
X (k+1) represents the traffic behavior value at the k+1 moment of the X (k+1) in history,
Covariance renewal equation:
P (k+1)=P (k+1 | k)-K (k+1) S (k+1) K ' (k+1) (8)
K ' (k+1) represents K (k+1) transposed matrix.
Further, the new breath square of step A3 traffic normalization, is tried to achieve with the traffic behavior preset value after calibration
Traffic normalization new breath square of the path in the period:
εv(k)=v ' (k) S-1(k)v(k) (9)
εv(k) the new breath square of traffic normalization, the transposed matrix that the filtering at v ' (k) k moment newly ceases, S-1(k) association side is newly ceased
The inverse matrix of difference, v (k) represent that the filtering at k moment newly ceases.
Further, the path of the daily maximum probability traveling of vehicle described in the step A4, refers to daily vehicle in difference
The route information that time uses, route information are referred to using fptree (frequent pattern tree (fp tree)) algorithms with daily traveling
The information that information is drawn.
Further, the new breath square of Reverse Commuting normalization in the 4th step, is drawn by formula 10-12:
X*(k)=F-1(k)X(k+1) (10)
K moment backward predictions newly cease:
v*(k)=z (k)-X*(k) (11)
Recalculating the k moment normalizes new breath square:
ε * (k)=v*′(k)S-1(k)v*(k) (12)
Further, the backward prediction in the 5th step is drawn by formula 10-13:
Try to achieve k moment normalizing and newly cease ratio:
λ=ε v* (k)/ε v (k) ' (13)
Wherein ... ... X*(k) the present traffic behavior value F reversely released is represented1(k) inverse matrix of state-transition matrix, X
(k 1) be prediction traffic behavior value, z (k) be real-time external measurement system measurement, ε*(k) traffic reversely to release
The new breath square of normalization, v*′(k) transposed matrix newly ceased for k moment backward predictions, v*(k) newly ceased for k moment backward predictions, S1
(k) it is the inverse matrix of new breath covariance.
Further, the 5th step:Traffic forecast is judged using the method for inspection, when traffic forecast for the second time not
When meeting λ > L conditions, then adjust the k+1 moment newly cease covariance, carry out to future traffic prediction, duplicate test and
Adjustment is obtained until the qualified traffic forecast of acquisition by formula 14 and 6-9:
S (k+1)=H (k+1) [F (k) P (k+1 | k) F ' (k)+λ Q (k)] H ' (k+1)+R (k+1) (14)
Wherein, the measurement matrix at the new breath covariance=k+1 moment at k+1 moment is multiplied by that (state-transition matrix at k moment multiplies
The k+1 predicted with the k moment covariance one-step prediction is multiplied by the normalizing at the transposed matrix+k moment of the state-transition matrix at k moment
Covariance of the new breath than being multiplied by the k moment) be multiplied by the k+1 moment measurement matrix transposed matrix+covariance.
Further, the preliminary traffic forecast of the 3rd step:It is by the detectable navigation using the vehicle to navigate
Routing information, and not using the daily maximum probability routing information of the vehicle to navigate, the traffic to every a period of time in paths future
Information is predicted.
The invention has the advantages that and advantage:
The present invention is used to store the running data that vehicle constantly uploads, and thus parses vehicle driving feature and path exists
Traffic behavior preset value under different traffics, and calculate the new breath of traffic normalization at outbound path each moment under each situation
Square and store.Then when user is navigated, road prediction and roading service are carried out in real time, and timing sends navigation and referred to
Information is led to navigation terminal.Navigation terminal, there is provided vehicle operation data is to terminal server and receives from server terminal
Navigation instruction information.The present invention is predicted ineligible traffic tentative prediction result, greatly again in real-time operation
The big burden for reducing server terminal real-time operation.
Brief description of the drawings
Fig. 1 is the flow chart of air navigation aid of the present invention;
Fig. 2 is the schematic diagram of navigation system of the present invention;
Fig. 3 is the schematic diagram of the example in the present invention illustrates to parsing trip feature.
Embodiment
With reference to embodiment and Figure of description, the present invention is further elaborated.
Embodiment:As shown in figure 1,
During system initialization:
The first step:Navigation terminal constantly uploads the running data of vehicle on server terminal storage vehicle.
Second step:The relevant information in whole city path and vehicle and storage are drawn after server terminal processing data.Whole city road
The relevant information of footpath and vehicle refers to, by historical traffic state value of the whole city per paths, counting the free pathway immediately arrived at
In the case of car saturation;The traffic behavior predictive value of every paths is calculated according to historical traffic state value, that is, after calibrating
Traffic behavior preset value;The traffic at the per moment being calculated according to resulting historical traffic state value on every paths is returned
One changes new breath square and stores;The path of the daily maximum probability traveling of vehicle is drawn according to fptree (frequent pattern tree (fp tree)) algorithm and deposited
Storage.
3rd step:After navigation terminal inputs destination, server terminal receive information simultaneously calculates all by present bit
The guidance path to destination is put, gather on guidance path and employ the information of vehicles of navigation system, and to by usually
The maximum probability run routing information of vehicle of the possibility drawn on guidance path is monitored, thus makes the tentative prediction of traffic;
4th step:The real-time traffic behavior of backward prediction based on tentative prediction, draw the new breath of Reverse Commuting normalization
Square;
5th step:The new breath square of traffic normalization in history in the same pathway traffic situation of synchronization is handed over reverse
Lead to the new breath square of normalization and make ratio λ, set threshold values L, meet in the period of prediction whole shared by number at the time of condition λ < L
The ratio of moment number is more than 50%, directly exports the preliminary traffic forecast of this period, and optimal path is selected after calculating;
If within the period of the traffic of prediction, the transport information of prediction differed compared with Real-time Traffic Information compared with
Greatly, the average speed for differing the larger vehicle flowrate for referring to prediction and car is more than or less than the average speed of real-time vehicle flowrate and car
Numerical value more than more than 6, carry out the 3rd step start;
6th step:If meet the ratio of whole moment numbers shared by number at the time of condition λ < L in the period of prediction for the first time
Less than 50%, then according to real-time traffic states, using the traffic behavior preset value after calibration, with standard Kalman filtering algorithm,
The traffic status prediction of following a period of time is calculated by formula 1-9, returns to the 5th step;The prediction drawn since for the second time
Period in meet whole moment numbers shared by number at the time of condition λ < L ratio be less than 50%, then by adjusting the k+1 moment
It is new to cease covariance to carry out, to the traffic status prediction in following setting time, returning to the 5th step;
After setting time is reached, then above-mentioned steps are re-started:3rd step, when the prediction result in setting time is forbidden
When true, then carry out the 3rd step and predict again.
The whole city path and the relevant information of vehicle, including:
A1:Navigation terminal collects historical traffic state value of the whole city per paths, and free pathway is immediately arrived at according to statistics
In the case of car saturation;Historical traffic state value refers to the average speed of vehicle flowrate and car;
A2:The traffic behavior predictive value of every paths, that is, the traffic after calibrating are calculated according to historical traffic state value
State preset value;
A3:The traffic normalization that per moment on every paths is calculated according to resulting historical traffic state value is new
Breath square simultaneously stores;
A4:Path and the storage of the daily maximum probability traveling of vehicle are drawn according to fptree FP-tree methods.
Server terminal of the present invention is the server using Hadoop or spark as Open Framework, has mass memory and meter
Calculation ability.Navigation terminal can be common automatic navigator or the smart mobile phone with locating module and on-line module
Can be positioned with GPS location, or Beidou navigation Deng, locating module, the locating module based on glonass positioning etc..
The trip feature for how parsing vehicle is specifically described below, trip feature refers to that vehicle is daily probably
The path of rate traveling, and the transport information such as speed of the daily maximum probability traveling of vehicle.The history row of server terminal combination vehicle
Data are sailed, the trip feature of vehicle is gone out by big data technology mining and stored.Big data technology is referred to from various class
In the running data of type, the quick ability for obtaining valuable information.Refer herein to by the daily driving information of vehicle
Collection, using big data mining algorithm such as Kmeans clustering algorithms, Naive Bayes NB Algorithms or fp-tree (frequencies
Numerous scheme-tree) algorithm etc. parses the trip feature of vehicle.Below by taking fp-tree (frequent pattern tree (fp tree)) algorithm as an example, knot
Close marginal data.
C1. as shown in the table from the history locality database of same vehicle upload, scanning vehicle stops in different time sections
The positional information leaned on, using the place position to stop as subitem.For convenience of description, it is assumed that vehicle can stop in fixed time period.
The number of the subitem of different time sections is pressed into frequency sort descending, and deletes the son that frequency is less than minimum support
.Minimum support can be set by actual conditions.Minimum support is set herein and is set to 1, then place j and h cast out.It can sort out
It is as follows:
Place w:6 place q:5 place t:5 place z:4 place e:4 place r:2
C2. recorded for the trip route of each, subitem is resequenced by C1 order.It can obtain:
C3. each bar order recording C2 obtained is inserted into fp-tree.Suffix when just starting is sky.Identical
The node of item will be chained up, and be formed as shown in Figure 3.
C4. frequent episode is found out from fp-tree frequent pattern tree (fp tree)s.Each single item in traversing graph 3, is carried out for items
Its ancestor node is traversed up, obtains different paths.It is complete retaining for the node on each paths by taking the e of place as an example
In the case of portion's node, while subtract identical number, it is therefore an objective to make e place number minimum.Thus obtained each road
Footpath end is all place e, it can be removed, and obtains conditional pattern base, as shown below, and suffix pattern now is place e.
C5. C4 result is run, obtained to stop position as new locality database, return C3 steps, recursive iteration
Node link is set to, the run routing information of the daily maximum probability of vehicle, accomplishing can be with the number in the place continuously driven in history
According to judging following driving path of maximum probability.
Below to the traffic behavior predictive values of every paths, and the traffic per the moment are calculated by historical traffic state value
The new breath square of normalization is specifically described, with the history at daily k moment and k+1 moment in the past period in same section
Traffic behavior value is that (k=0,1,2,3,4,5,6 ... n), substitutes into the initial traffic behavior preset value of formula alignment and meter for model
The new breath square of traffic normalization at each moment is calculated, historical traffic state value refers to the row constantly uploaded according to vehicle in history
Sail the average speed of vehicle flowrate that data parse and car.Historical traffic state value is represented with X, while using the outside on road
The traffic behavior value of the measurements such as measuring apparatus such as monitor, ground induction coil is system quantities measured value z.Inside round bracket " k | k " or
" k+1 | it is kth or the traffic behavior value at the moment of kth+1 that k+1 ", which represents the amount, as " X (k+1 | k+1) " just represents the moment of kth+1
Traffic behavior value;Be inside round bracket " k+1 | k ", just represent the predicted value at the k+1 moment predicted based on the k moment;Character
The amount of band " ^ " all represents system predicted value, and the amount of character upper right corner band " ' " represents transposed matrix, character upper right corner band " -1 "
Amount represents inverse matrix.Set initial traffic behavior preset value is state-transition matrix F (k), input control item matrix G (k), mistake
Journey noise sequence V (k), measurement matrix H (k+1), the measurement noise sequence W (k+1) with covariance R (k+1), covariance P (k |
k).And then carry out asking calculation with specific formula:
Traffic behavior equation:
Traffic measurement equation:
Filter the computing newly ceased:
Covariance one-step prediction (state error covariance, available for weigh precision of prediction):
P (k+1 | k)=F (k) P (k | k) F ' (k)+Q (k) (4)
New breath covariance (measure prediction covariance, available for weigh accuracy in measurement):
S (k+1)=H (k+1) P (k+1 | k) H ' (k+1)+R (k+1) (5)
Gain:
K (k+1)=P (k+1 | k) H ' (k+1) S-1(k+1) (6)
Traffic behavior renewal equation:
Covariance renewal equation:
P (k+1 | k+1)=P (k+1 | k)-K (k+1) S (k+1) K ' (k+1) (8)
Continuous recursive iteration operation, calculates a series of traffic behavior values using the period as interval.Same path is same
The historical traffic state value at period at each moment is compared with the actual traffic state value calculated, in error range,
When the ratio of number and total moment number exceedes default certain value at the time of differing larger with historical traffic state value, just to phase
The initial traffic behavior preset value answered is calibrated, and calibration refers to the progress of initial traffic behavior preset value is single or whole
Increase reduces, until traffic behavior preset value in time period is stored when eligible.And with the friendship after calibration
Logical state preset value tries to achieve traffic normalization new breath square of the path in the period:
εv(k)=v ' (k) S-1(k)v(k) (9)
The traffic behavior preset value in different paths is obtained in the same way, and calculates the traffic normalization of every paths
New breath square simultaneously stores.
During system operation:
D1. after user sends navigation requests, server terminal cooks up all navigation by current location to destination
Path, the All Paths planning letter of all vehicles to have been navigated that can pass through in preset time in guidance path of collection
Breath, and the maximum probability run routing information of vehicle of the possibility drawn by monitoring usually on guidance path, thus to navigation
The traffic of following a period of time carries out tentative prediction in path.
D2. enter system judgment step, gather Real-time Traffic Information, and do the traffic behavior of above-mentioned tentative prediction is worthwhile
The X (k+1 | k+1) at k+1 moment, make the reverse state prediction at k moment, be set to X*.In historical traffic state value, find out and going through
History traffic behavior value and real-time traffic states are worth the corresponding new breath square ε v of traffic normalization of similar historical traffic state value
(k) ', and traffic behavior preset value, whether correction is needed to judge tentative prediction, carrying out practically formula is as follows:
X*(k | k)=F-1(k)X(k+1|k+1) (10)
K moment backward predictions newly cease:
v*(k)=z (k)-X*(k|k) (11)
Recalculating the k moment normalizes new breath square:
ε * (k)=v*′(k)S-1(k)v*(k) (12)
Try to achieve k moment normalizing and newly cease ratio:
λ=ε v* (k)/ε v (k) ' (13)
The normalizing at each moment in the period is newly ceased and assessed than λ, if assessment threshold value is L, in the period of prediction
Meet that the ratio of whole moment numbers shared by number at the time of condition λ < L is adjusted less than 50% pair of tentative prediction.If meet bar
Part, just directly export predicting in next step for this period.It is ineligible, it is adjusted, what is be adjusted is specifically basis
Real-time traffic states numerical value, with reference to the good traffic behavior preset value of the path alignment, future one is calculated with formula (1)~(8)
The traffic forecast of section time.Traffic forecast is judged again, when traffic forecast is or else eligible, then just only passes through adjustment
The k+1 moment newly ceases covariance and comes and then the traffic of a period of time in future is predicted, and constantly returns to traffic forecast
D2 steps are judged, until obtaining qualified traffic forecast.Specific step is as follows:
The adjustment k+1 moment, which newly ceases covariance formula, is:
S (k+1)=H (k+1) [F (k) P (k+1 | k) F ' (k)+λ Q (k)] H ' (k+1)+R (k+1) (14)
The traffic forecast in the same period is recalculated according to formula (6)~(9), constantly adjustment is until obtain meeting bar
The pathway traffic of part is predicted and exported.
D4. the traffic of each guidance path of the server terminal to predicting is compared, and selects optimal path hair
Deliver to navigation terminal.The optimal path is divided with different weights, for example the time is minimum, and vehicle flowrate is minimum etc., can by car
Main unrestricted choice.
D5. timing repeats this process from D1 to D4, and acquisition can be with the optimal path of pathway traffic changed condition.If pre-
In the period of the traffic of survey, the transport information of prediction differs larger compared with Real-time Traffic Information, then repeats D1 in advance
To this process of D4.In this way, the path predicted distortion to vehicle can be reduced, and makes whole piece traffic route condition predicting
Error large effect.
A kind of navigation system based on cloud computing and big data, as shown in Fig. 2 the system includes:Server terminal, carry
Excavate and store for big data, and cloud computing service.The running data constantly uploaded for storing vehicle, and thus parse complete
The relevant information of city path and vehicle and storage.The relevant information of whole city path and vehicle refers to the going through per paths by whole city
History traffic behavior value, count the car saturation in the case of the free pathway immediately arrived at;Calculated according to historical traffic state value
Go out the traffic behavior predictive value of every paths, that is, the traffic behavior preset value after calibrating;According to resulting historical traffic state
The new breath square of traffic normalization at per moment that value is calculated on every paths simultaneously stores;According to fptree (frequent pattern tree (fp tree))
Algorithm draws the path of the daily maximum probability traveling of vehicle and storage.Return with outbound path traffic at each moment under each situation is calculated
One changes new breath square and stores.Then when user is navigated, road prediction and roading service are carried out in real time, and timing is sent out
Navigation instruction information is sent to navigation terminal.Navigation terminal, there is provided vehicle operation data is to terminal server and receives from service
The navigation instruction information of device terminal.
Claims (10)
- A kind of 1. air navigation aid based on cloud computing and big data, it is characterised in that:Comprise the following steps:The first step:The vehicle of whole city uploads running data information in real time;Second step:Server terminal stores and carries out algorithm simulation, draws the relevant information of whole city path and vehicle:3rd step:Navigation terminal input destination after, server terminal receive information and calculate it is all by current location to The guidance path of destination, gather on guidance path and employ the information of vehicles of navigation system, and to by monitoring usually The maximum probability run routing information of vehicle of the possibility drawn on guidance path, thus makes the tentative prediction of traffic;4th step:The real-time traffic behavior of backward prediction based on tentative prediction, draw the new breath square of Reverse Commuting normalization;5th step:The new breath square of traffic normalization in history in the same pathway traffic situation of synchronization is returned with Reverse Commuting One, which changes new breath square, makees ratio λ, sets threshold values L, whole moment shared by number at the time of condition λ < L are met in the period of prediction Several ratios is more than 50%, directly exports the preliminary traffic forecast of this period, and optimal path is selected after calculating;If within the period of the traffic of prediction, the transport information of prediction differs larger, phase compared with Real-time Traffic Information The average speed of the larger vehicle flowrate for referring to prediction of difference and car is more than or less than the number of the average speed of real-time vehicle flowrate and car Value carries out the 3rd step more than more than 6;6th step:If meet that the ratio of whole moment numbers shared by number at the time of condition λ < L is less than in the period of prediction for the first time 50%, then according to real-time traffic states, using the traffic behavior preset value after calibration, calculated with standard Kalman filtering algorithm Go out the traffic status prediction of following a period of time, return to the 5th step;Meet in the period of the prediction drawn since for the second time The ratio of whole moment numbers shared by number is less than 50% at the time of condition λ < L, then newly ceases covariance by adjusting the k+1 moment to enter Row returns to the 5th step to the traffic status prediction in following setting time;After setting time is reached, then the step of the above-mentioned first step~the 6th is re-started;When the prediction result in setting time is forbidden When true, then carry out the 3rd step and predict again.
- 2. the air navigation aid based on cloud computing and big data as described in claim 1, it is characterised in that:The whole city path and The relevant information of vehicle, including:A1:Navigation terminal collects historical traffic state value of the whole city per paths, and free pathway situation is immediately arrived at according to statistics Under car saturation;A2:The traffic behavior predictive value of every paths, that is, the traffic behavior after calibrating are calculated according to historical traffic state value Preset value;A3:The new breath of traffic normalization that per moment on every paths is calculated according to resulting historical traffic state value is flat Side simultaneously stores;A4:Path and the storage of the daily maximum probability traveling of vehicle are drawn according to FP-tree method.
- 3. the air navigation aid based on cloud computing and big data as described in claim 2, it is characterised in that:The historical traffic shape State value refers to the average speed of vehicle flowrate and car.
- 4. the air navigation aid based on cloud computing and big data as described in claim 2, it is characterised in that:The step A2 traffic The computational methods of state preset value are:Using the historical traffic state value at same path daily k moment and k+1 moment as model (k=0,1,2,3,4,5,6 ... N),Traffic behavior equation:<mrow> <mover> <mi>X</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mrow> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>|</mo> <mi>k</mi> </mrow> <mo>)</mo> </mrow> <mo>=</mo> <mi>F</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>X</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>G</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mi>u</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>V</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>Wherein:F (k) is state-transition matrix, and G (k) is input control item matrix, and V (k) is that process noise sequence page 8/11 arranges, X (k) the traffic behavior value at history kth moment is represented, u (k) represents control signal,Represent pre- based on the k moment The predicted value at the k+1 moment of survey;X represents historical traffic state value, while uses the external measurement devices such as monitor on road, The traffic behavior value of ground induction coil measurement is equally historical data with x as system measurement z;The amount of band " ^ " all represents system Predicted value, the amount of upper right corner band " ' " represent transposed matrix, and the amount of upper right corner band " -1 " represents inverse matrix;Traffic measurement equation:<mrow> <mover> <mi>z</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>|</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>=</mo> <mi>H</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mover> <mi>X</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>|</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>W</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>Represent the systematic survey predicted value at k+1 moment predicted based on the k moment, H (k+1) is measurement matrix, W (k+1) it is measurement noise sequence;Filter the computing newly ceased:<mrow> <mi>v</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>=</mo> <mi>z</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mover> <mi>z</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>Z (k+1) represents the systematic survey history value at k+1 moment,Represent the systematic survey predicted value at k+1 momentCovariance one-step prediction:P (k+1 | k)=F (k) P (k) F ' (k)+Q (k) (4)F ' (k) represents F (k) transposed matrix, and Q (k) represents covarianceNew breath covariance:S (k+1)=H (k+1) P (k+1 | k) H ' (k+1)+R (k+1) (5)S (k+1) newly ceases the association side at the k+1 moment that covariance, H (k+1) measurement matrix, P (k+1 | k) are predicted based on the k moment Poor one-step prediction, the transposed matrix of H ' (k+1) measurement matrix, R (k+1) and Q (k) represent different covariancesGain:K (k+1)=P (k+1 | k) H ' (k+1) S-1(k+1) (6)K (k+1) represents the gain at the k+1 moment for next stepTraffic behavior renewal equation:<mrow> <mi>X</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>=</mo> <mover> <mi>X</mi> <mo>^</mo> </mover> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>|</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>K</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mi>v</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>X (k+1) represents the traffic behavior value at the k+1 moment of the X (k+1) in history,Covariance renewal equation:P (k+1)=P (k+1 | k)-K (k+1) S (k+1) K ' (k+1) (8)K ' (k+1) represents K (k+1) transposed matrix.
- 5. the air navigation aid based on cloud computing and big data as described in claim 4, it is characterised in that:The step A3 traffic The new breath square of normalization, it is that traffic normalization new breath of the path in the period is tried to achieve with the traffic behavior preset value after calibration Square:εv(k)=v ' (k) S-1(k)v(k) (9)εv(k) the new breath square of traffic normalization, the transposed matrix that the filtering at v ' (k) k moment newly ceases, S-1(k) covariance is newly ceased Inverse matrix, v (k) represent that the filtering at k moment newly ceases.
- 6. the air navigation aid based on cloud computing and big data as described in claim 2, it is characterised in that:Institute in the step A4 The path of the daily maximum probability traveling of vehicle is stated, refers to the route information that daily vehicle uses in different time, route Information refers to the information drawn using fptree (frequent pattern tree (fp tree)) algorithms with daily driving information.
- 7. the air navigation aid based on cloud computing and big data as described in claim 1, it is characterised in that:In 4th step The new breath square of Reverse Commuting normalization, is drawn by formula (10)-(12):X*(k)=F-1(k)X(k+1) (10)K moment backward predictions newly cease:v*(k)=z (k)-X*(k) (11)Recalculating the k moment normalizes new breath square:ε * (k)=v*′(k)S-1(k)v*(k) (12)Wherein, X*(k) the present traffic behavior value F reversely released is represented-1(k) inverse matrix of state-transition matrix, X (k+1) are The traffic behavior value of prediction, the system measurement that z (k) measures for real-time external, ε*(k) traffic reversely to release normalizes new Breath square, v*′(k) transposed matrix newly ceased for k moment backward predictions, v*(k) newly ceased for k moment backward predictions, S-1(k) it is new Cease the inverse matrix of covariance.
- 8. the air navigation aid based on cloud computing and big data as described in claim 7, it is characterised in that:In 5th step Backward prediction is drawn by formula (10)-(13):Try to achieve k moment normalizing and newly cease ratio:λ=ε v* (k)/ε v (k) ' (13).
- 9. the air navigation aid based on cloud computing and big data as described in one of claim 4-8, it is characterised in that:Described 5th Step:Traffic forecast is judged using the method for inspection, when traffic forecast does not meet λ > L conditions for the second time, then adjusts k+1 Moment newly ceases covariance, carries out the prediction of the traffic to future, duplicate test and adjustment until obtaining qualified friendship Logical prediction, is obtained by formula (14) and (6)-(9):S (k+1)=H (k+1) [F (k) P (k+1 | k) F ' (k)+λ Q (k)] H ' (k+1)+R (k+1) (14)Wherein, S (k+1) is the new breath covariance at k+1 moment, and H (k+1) is the measurement matrix at k+1 moment, and F (k) is the k moment State-transition matrix, the covariance one-step prediction for the k+1 that P (k+1 | k) predict for the k moment, F ' (k) are the state transfer at k moment The transposed matrix of matrix, Q (k) be the k moment covariance, H ' (k+1) be the k+1 moment measurement matrix transposed matrix, R (k+ 1) it is covariance.
- 10. the air navigation aid based on cloud computing and big data as described in claim 1, it is characterised in that:3rd step Preliminary traffic forecast:It is the navigation route information using the vehicle to navigate by can detect, and using the vehicle to navigate Daily maximum probability routing information, the transport information of every a period of time in paths future is predicted.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610150671.5A CN105716620B (en) | 2016-03-16 | 2016-03-16 | A kind of air navigation aid based on cloud computing and big data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610150671.5A CN105716620B (en) | 2016-03-16 | 2016-03-16 | A kind of air navigation aid based on cloud computing and big data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105716620A CN105716620A (en) | 2016-06-29 |
CN105716620B true CN105716620B (en) | 2018-03-23 |
Family
ID=56159021
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610150671.5A Expired - Fee Related CN105716620B (en) | 2016-03-16 | 2016-03-16 | A kind of air navigation aid based on cloud computing and big data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105716620B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3339807B1 (en) * | 2016-12-20 | 2024-03-13 | HERE Global B.V. | An apparatus and associated methods for determining the location of a vehicle |
CA3028216A1 (en) * | 2017-12-21 | 2019-06-21 | Bce Inc. | Method and system for monitoring and assessing road conditions |
CN110807912A (en) * | 2018-08-06 | 2020-02-18 | 厦门雅迅网络股份有限公司 | Road driving speed prediction method, terminal device and storage medium |
CN109886305B (en) * | 2019-01-23 | 2021-05-04 | 浙江大学 | Multi-sensor non-sequential measurement asynchronous fusion method based on GM-PHD filtering |
CN111829538A (en) * | 2019-04-16 | 2020-10-27 | 上海博泰悦臻电子设备制造有限公司 | Traffic safety navigation method, storage medium and electronic equipment |
US11105646B2 (en) * | 2019-04-17 | 2021-08-31 | International Business Machines Corporation | Providing navigation services using context-aware trajectory analysis |
CN110160547B (en) * | 2019-05-30 | 2020-09-25 | 辽宁工业大学 | Vehicle navigation system and method based on big data cloud computing |
CN111854781A (en) * | 2020-06-17 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Navigation path recommendation method and device and electronic equipment |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1482278A2 (en) * | 2003-05-28 | 2004-12-01 | LG Electronics Inc. | System and method for estimating driving time using road traffic condition information |
CN102445209A (en) * | 2011-09-28 | 2012-05-09 | 深圳市赛格导航科技股份有限公司 | Dynamic traffic navigation method and dynamic traffic navigation equipment thereof |
CN103884344A (en) * | 2014-03-31 | 2014-06-25 | 深圳市赛格导航科技股份有限公司 | Intelligent navigation method and system based on mass vehicle data |
CN104464344A (en) * | 2014-11-07 | 2015-03-25 | 湖北大学 | Vehicle driving path prediction method and system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20020021691A (en) * | 2000-09-16 | 2002-03-22 | 이계철 | Method and Apparatus for Car Navigation Service using DSRC System |
-
2016
- 2016-03-16 CN CN201610150671.5A patent/CN105716620B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1482278A2 (en) * | 2003-05-28 | 2004-12-01 | LG Electronics Inc. | System and method for estimating driving time using road traffic condition information |
CN102445209A (en) * | 2011-09-28 | 2012-05-09 | 深圳市赛格导航科技股份有限公司 | Dynamic traffic navigation method and dynamic traffic navigation equipment thereof |
CN103884344A (en) * | 2014-03-31 | 2014-06-25 | 深圳市赛格导航科技股份有限公司 | Intelligent navigation method and system based on mass vehicle data |
CN104464344A (en) * | 2014-11-07 | 2015-03-25 | 湖北大学 | Vehicle driving path prediction method and system |
Also Published As
Publication number | Publication date |
---|---|
CN105716620A (en) | 2016-06-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105716620B (en) | A kind of air navigation aid based on cloud computing and big data | |
KR100495017B1 (en) | Traffic information providing system and method thereof | |
US8457880B1 (en) | Telematics using personal mobile devices | |
CN102023018B (en) | Method of predicting energy consumption, apparatus for predicting energy consumption, and terminal apparatus | |
Simroth et al. | Travel time prediction using floating car data applied to logistics planning | |
US8364387B2 (en) | Route search device and information control server | |
CN105702073B (en) | A kind of dynamic steering air navigation aid based on road network decision point journey time index value | |
US20080077316A1 (en) | Traffic State Predicting Apparatus | |
US8260532B2 (en) | Traffic probe in-vehicle map-based process to reduce data communications and improve accuracy | |
CN105528815B (en) | A kind of vehicle driving trace method of adjustment and vehicle driving trace adjustment system | |
CN107369318A (en) | A kind of speed predicting method and device | |
CN108604407A (en) | Method and system for generating expected gait of march | |
CN108038651A (en) | A kind of monitoring logistics transportation system for tracing and managing | |
CN105120433A (en) | WLAN indoor positioning method based on continuous sampling and fuzzy clustering | |
CN102881171A (en) | Vehicle detecting method, vehicle detecting system and vehicle path planning system | |
CN105424050B (en) | A kind of method and apparatus of determining vehicle running path | |
Lin et al. | Vehicle re-identification with dynamic time windows for vehicle passage time estimation | |
CN104794895A (en) | Multisource traffic information fusion method for expressways | |
US20230194274A1 (en) | Using Radio Frequency Signal Strength to Improve Route Options in a Navigation Service | |
CN106855878A (en) | History wheelpath display methods and device based on electronic map | |
CN109754598A (en) | A kind of congestion is formed a team recognition methods and system | |
CN112598199A (en) | Monitoring and early warning method based on decision tree algorithm | |
CN110909907A (en) | Method and device for predicting fuel consumption of truck and storage medium | |
CN115164922A (en) | Path planning method, system, equipment and storage medium | |
JP2011141678A (en) | Information processing apparatus, computer program, and information processing method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180323 Termination date: 20190316 |