CN112036630A - Highway pavement rainfall distribution estimation method, storage medium and computing device - Google Patents
Highway pavement rainfall distribution estimation method, storage medium and computing device Download PDFInfo
- Publication number
- CN112036630A CN112036630A CN202010864891.0A CN202010864891A CN112036630A CN 112036630 A CN112036630 A CN 112036630A CN 202010864891 A CN202010864891 A CN 202010864891A CN 112036630 A CN112036630 A CN 112036630A
- Authority
- CN
- China
- Prior art keywords
- sampling
- rainfall
- estimation
- sensor
- road surface
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000009826 distribution Methods 0.000 title claims abstract description 28
- 238000003860 storage Methods 0.000 title claims abstract description 10
- 238000005070 sampling Methods 0.000 claims abstract description 154
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 238000001914 filtration Methods 0.000 claims abstract description 28
- 230000003068 static effect Effects 0.000 claims abstract description 16
- 230000007246 mechanism Effects 0.000 claims abstract description 10
- 238000011478 gradient descent method Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 61
- 230000006870 function Effects 0.000 claims description 50
- 239000013598 vector Substances 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 5
- 238000011217 control strategy Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 8
- 238000010586 diagram Methods 0.000 description 14
- 238000004088 simulation Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 206010039203 Road traffic accident Diseases 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000005314 correlation function Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a method for estimating rainfall distribution on a road surface, a storage medium and a computing device, wherein a sampling database is constructed according to sampling information of sampling points in a target area, and relevant parameters are initialized; a dynamic and static mixed sensor sampling model under a self-triggering mechanism is adopted, the sampling interval is adjusted according to the sampling error of the movable sensor, and the sampling data is dynamically updated; on the basis, an estimation model of the rainfall distribution of the road surface is designed; according to an improved Kalman filtering state estimation algorithm, combining a weight coefficient in surface fitting with a sampling position of a sensor network to form a state variable in Kalman filtering, and giving a road rainfall distribution result based on current sampling information; according to the gradient descent method, the rainfall distribution estimation result in the road domain is further optimized by means of adjusting the sampling position of the mobile sensor. The invention has flexible sampling position, effectively improves the operation efficiency and reduces the calculation load of the estimation system; the accuracy is improved.
Description
Technical Field
The invention belongs to the technical field of spatial information distribution estimation, and particularly relates to a road surface rainfall distribution estimation method based on a dynamic and static hybrid sensor, a storage medium and a computing device.
Background
With the vigorous development of the transportation industry, road traffic safety accidents frequently occur, and road surface rainwater is one of the main reasons for accidents. The road surface friction coefficient is obviously reduced due to the rain on the road surface, the braking performance of the vehicle is poor, the out-of-control phenomena of vehicle slipping, side slipping or wheel idling and the like are easy to occur, and serious traffic accidents are easy to be induced. Therefore, as long as the information of rainfall capacity in the area is mastered, the vehicles in the target area can be effectively managed and controlled in time, traffic accidents are avoided, and driving safety is guaranteed.
At present, the rainfall in the target area is generally predicted by adopting a traditional estimation method such as an iterative least square method or Gaussian estimation. These conventional estimation algorithms are relatively slow in estimation speed and generally have low estimation accuracy. In the face of such problems, a general method improves accuracy by setting a large number of sampling points. But this requires a large amount of data sampling and computation costs. Moreover, the algorithms often cannot eliminate the influence of noise, and further, the accuracy of the estimation result is influenced.
In addition, the rainfall variation in the target area is generally non-linear, but in the conventional estimation algorithm sampling mode, a fixed-interval sampling mode is generally adopted. In such a sampling mode, the calculation resources of the estimation system are wasted in a time period when the road rainfall information changes slowly, and the calculation cost is increased. For the condition that the rainfall information of the road surface changes rapidly, the estimation result and the following effect are poor, and the real-time road surface condition cannot be reflected timely. Therefore, how to improve the efficiency of the estimation system is an urgent problem to be solved.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a method, a storage medium and a computing device for estimating the distribution of rainfall on a road surface based on a dynamic and static hybrid sensor, wherein the dynamic and static hybrid sensor is used for acquiring observation data, and a target area rainfall estimation algorithm based on a dynamic sampling mechanism and an improved Kalman filtering method is provided under the condition of considering sampling noise, so as to effectively realize the efficient and accurate estimation of the rainfall on the road surface in the target area.
The invention adopts the following technical scheme:
a method for estimating rainfall distribution of a road surface comprises the following steps:
s1, constructing a sampling database containing sampling positions and road rainfall according to sampling information of sampling points in the target area, and initializing parameters;
s2, adopting a dynamic and static mixed sensor sampling model under a self-triggering mechanism, defining a sampling threshold value delta of the rainfall on the road surface by normalization according to the error between the sampling value of the sensor in the current execution period and the average sampling value in the previous execution period, dividing the threshold value delta of the rainfall on the road surface to form a plurality of sampling error intervals, determining the sampling intervals of the sensor in different sampling error intervals, and dynamically updating the sampling data in the sampling database of the step S1;
s3, designing a road surface rainfall estimation model according to the dynamically updated sampling data of the step S2 by adopting a surface fitting method;
s4, combining the weight coefficient of the rainfall estimation model of the road surface after the curved surface is fitted in the step S3 with the sampling position of the sensor network by using a Kalman filtering algorithm to form a state variable in Kalman filtering, then adjusting parameters of an observation equation in the Kalman filtering algorithm, improving the Kalman filtering state estimation algorithm, and estimating the rainfall at any position in the target area according to the state variable of the Kalman filtering at the current moment;
and S5, optimizing the estimation result of the rainfall in the target area in the step S4 by means of adjusting the sampling position of the mobile sensor according to a gradient descent method, and finishing the rainfall distribution estimation.
Specifically, in step S1, the sampling information of the sampling points in the target region is:
Y=[S,M]T
wherein ,the sampling information of the stationary sensor is represented,the data set representing the sampled information of the mobile sensor, the region estimation algorithm is:
D={P,Y}
wherein D is a sampling data set containing sampling position information, and P is a sensor sampling position set.
Specifically, in step S2, separating the estimated execution period of the sensor network from the information sampling period, determining the sampling times of the sensor in a fixed execution period, giving the fixed execution period of the sensor, and using the error between the current sampling value and the sampling average value of the previous execution period in one execution period as the threshold Δ of the rainfall on the road surface; normalizing the threshold value delta according to the sampling interval of the current momentAnd adjusting the sampling time of the device sampling, and updating the sampling data Y of the estimation system.
Further, the threshold Δ is expressed as follows:
wherein ,the ith sample value representing the mth execution cycle,represents the mean of the samples of the m-1 execution cycle; the normalization process for the threshold value Δ is as follows:
wherein, Delta is more than 0, and the value range of e is [0,1]](ii) a Divide e into three segments, sampling intervalRespectively, the upper limit and the lower limit ofAndsampling intervalThe change rule according to e is as follows:
wherein ,0≤c1<c2Is less than or equal to 1, and theta is the minimum step length for increasing the sampling interval.
Specifically, in step S3, a radial basis function is defined for the rainfall estimation of the target area, and the radial basis center points on both sides of the highway and on the highway are determined with the highway surface as the center, where the radial basis function specifically is:
wherein ,ψj(P) representing the jth radial basis function in the set of radial basis function vectors; l is the dimension of the vector set, σjIs the width of the radial basis function, betajIs a normalization constant, ξjIs the center point of the radial basis function and P represents any point within the target region.
Further, the central points of the two sides of the road are selected as follows:
wherein, i is 1, …, nsRepresenting the number of stationary sensors, n representing the selected number of center points, d1Indicating the length of the road in the target area, d2Representing the width of two sides of the highway, wherein m represents the total row number of the central points of the two sides of the highway, and j represents the jth row;
the central point on the road surface is selected as follows:
wherein, i is 1, …, nmIndicating the number of motion sensors;
the vector form of the radial basis function is denoted as psi (P)T=[ψ1(P),…,ψl(P)]The sampling set of rainfall in the target area is expressed as:
φ(P,t)=ψ(P)Tx(t)
wherein x (t) ═ x1(t),…xl(t)]TRepresenting the weight coefficients of the radial basis functions.
Specifically, step S4 specifically includes:
s401, taking a weight coefficient x in the radial basis function as a state variable in a linear road rainfall estimation model, wherein a linear dynamic system comprises the following components:
x(k+1)=Ax(k)+ω(k)
wherein ,representing system matrix, noiseThe white Gaussian noise is zero mean value, and the variance is recorded as W;
s402, determining a sampling value set of each sampling device in an execution cycleGiving an observation equation Si(k) The following were used:
wherein ,a matrix of the system is represented,representing the observation matrix, W being noiseThe variance of (a);
s403, setting a Kalman gain matrix Ki(k) The following were used:
wherein ,Ki(k) For the Kalman gain matrix obtained at time k, Q (k | k-1) is the estimated prediction covariance matrix at time k-1, Si(k) Is a sampling noise covariance matrix;
s404, setting the state variable at the initial time asCalculating the state variable of the next moment
S405, setting an initial value of an error covariance matrix in a Kalman filtering algorithm to be Q (0|0), and updating an estimated prediction covariance matrix;
s406, setting an initial value of the noise covariance matrix S (0), and calculating a covariance matrix Q (k | k);
Specifically, in step S5, the control strategy for optimizing the rainfall estimation result in the target area in step S4 is:
wherein O is a target region, W is a system noise covariance, J is an estimation error cost function, and piFor the sampling position of the sensor, Q is the estimation error covariance matrix, psi (p) is the radial basis function, K is the Kalman gain matrix, and S is the noise covariance matrix.
Another aspect of the invention is a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform any of the methods described.
Another technical solution of the present invention is a computing device, including:
one or more processors, memory, and one or more programs stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods.
Compared with the prior art, the invention has at least the following beneficial effects:
the invention discloses a road surface rainfall estimation method based on a dynamic and static mixed sensor. By adopting a curved surface fitting method and an improved Kalman filtering method, a weight coefficient in curved surface fitting is combined with a sampling position of a mobile sensor network to form a state variable in Kalman filtering, and the rainfall capacity of a target area is estimated in real time according to a correlation function of the state variable and the rainfall capacity information of the target area. Compared with the prior art, the method can greatly improve the estimation precision and the estimation speed, and can reflect the change of rainfall in the target area in real time.
Furthermore, the rainfall information of the target area is acquired by using the fixed sensor with the rainfall detection function arranged near the road surface and the mobile sensor arranged on the road surface, and the rainfall information of the target area can be detected in real time by detecting the area information through the dynamic and static mixed sensors. Compared with the existing detection method, the method has the advantages that the sampling position of the sensor is not limited to the roadside, and the flexibility is realized.
Furthermore, the dynamic sampling mechanism designed by the invention can dynamically adjust the sampling period of the sampling network by judging the variation degree of the rainfall sampling value, more reasonably allocate the calculation load of the estimation system, expand the application range of the invention and improve the accuracy and the real-time performance of the rainfall information estimation of the target area.
Furthermore, the estimation model adopts a Gaussian function as a radial basis function in surface fitting, and aiming at the narrow and long space distribution characteristics of the road surface, the central point position of the radial basis function is reasonably selected by the method and mainly concentrated around the road, so that the accuracy of the estimation algorithm is effectively improved. In addition, a Gaussian function is adopted as a radial basis function, and the linear combination of the radial basis functions has better continuity and smoothness in a two-dimensional space, so that the generation of a nonlinear term is effectively avoided.
Furthermore, a dynamic sampling mechanism and an improved Kalman filtering method are utilized to combine the weight coefficient in the surface fitting with the sampling position of the sensor network, and the rainfall capacity of any point in the target area is estimated. Compared with the prior art, the method can improve the estimation precision and the estimation speed, effectively inhibit the interference of environmental noise on the estimation result, and reflect the change of rainfall in a target area in real time.
Furthermore, by means of adjusting the sampling position of the mobile sensor, the estimation result of the rainfall in the target area is further optimized by using a gradient descent method. Compared with the prior art, the method can automatically adjust the sampling position, and further improves the accuracy of the rainfall estimation result of the target area.
In conclusion, the method for estimating the rainfall capacity of the road surface based on the dynamic and static mixed sensors, disclosed by the invention, has the advantages that the sampling position is flexible, the operation efficiency is effectively improved, and the calculation load of an estimation system is reduced; the accuracy of the estimation algorithm is improved.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
FIG. 1 is a schematic view of a sampling instant;
FIG. 2 is a schematic diagram of the distribution of sensors for detecting the rainfall on the road surface;
FIG. 3 is a flow chart of the present invention;
fig. 4 is a simulation diagram of the estimated distribution of rainfall on the road surface, wherein (a) is a simulation diagram of the estimated distribution of the rainfall in the target area at the time when t is 1s, (b) is a simulation diagram of the estimated distribution of the rainfall in the target area at the time when t is 50s, and (c) is a simulation diagram of the estimated distribution of the rainfall in the target area at the time when t is 100 s;
fig. 5 is a simulation diagram of the rainfall estimation error of the present invention.
Detailed Description
The invention provides a road surface rainfall distribution estimation method based on a dynamic and static mixed sensor, which comprises the steps of constructing a sampling database according to sampling information of a sensor network in a target area and initializing relevant parameters; secondly, a dynamic and static mixed sensor sampling model under a self-triggering mechanism is adopted, the sampling interval is adjusted according to the sampling error of the sensor, and the sampling data is dynamically updated; on the basis, a road rainfall distribution estimation model is designed; then, according to an improved Kalman filtering state estimation algorithm, combining a weight coefficient in surface fitting with a sampling position of a sensor network to form a state variable in Kalman filtering, and providing a road rainfall distribution algorithm based on current sampling information; and finally, according to a gradient descent method, further optimizing an estimation result of rainfall distribution in the road domain by means of adjusting the sampling position of the mobile sensor.
Referring to fig. 3, the method for estimating the distribution of rainfall on the road surface based on the dynamic and static hybrid sensors according to the present invention includes the following steps:
s1, arranging sensor devices with a function of detecting the rainfall on the road surface at two sides of the road, wherein the sensor devices are divided into fixed sensor devices and movable sensor devices, collecting the rainfall information of the current road surface in real time through a wireless communication technology, and preliminarily constructing a sampling database of the rainfall on the road surface and initializing relevant parameters by combining the position information of sampling devices as shown in figure 2;
the sampling information of the rainfall in the target area is expressed as:
Y=[S,M]T
wherein ,the sampling information of the stationary sensor is represented,representing the sampled information of the mobile sensor, on the basis of which the data set of the rainfall estimation algorithm is
D={P,Y}
Where D denotes a sample data set containing sample position information, and P denotes a sample position.
S2, adopting a dynamic and static mixed sensor sampling model under a self-triggering mechanism, utilizing a fixed sensor and a mobile sensor to combine with collected region information, setting a road rainfall correlation threshold value by comparing errors of front and rear sampling information of the sensors, carrying out normalization processing on the threshold value, and dynamically updating sampling data by dividing the threshold value, adjusting sampling intervals;
setting a fixed sensor execution period, taking the difference value of the current sampling value and the sampling mean value of the previous execution period in one execution period as a threshold value delta, and changing the next sampling interval;
the threshold Δ is expressed as follows:
wherein ,the ith sample value representing the mth execution cycle,representing the mean of the samples for the m-1 execution cycle.
On the basis, the threshold value delta is normalized, and the specific calculation mode is as follows:
wherein, Delta is more than 0, and the value range of e is [0,1 ].
Divide e into three segments, sampling intervalRespectively, the upper limit and the lower limit ofAndsampling intervalThe change rule according to e is as follows:
wherein ,0≤c1<c2And the division boundary of the threshold value is less than or equal to 1, and theta is the minimum step length for increasing the sampling interval.
Sampling interval according to current timeThe sampling time of the device sampling is adjusted and the sampling data Y of the estimation system is updated as shown in fig. 1.
S3, designing an estimation model of the rainfall of the road surface according to a dynamic sampling mechanism and a curved surface fitting method;
the radial basis functions are defined as follows:
ψ1(P)=1
wherein ,ψj(P) representing the jth radial basis function in the set of radial basis function vectors; l is the dimension of the vector set, σjIs the width of the radial basis function, betajIs a normalization constant, ξjIs the center point of the radial basis function and P represents any point within the target region.
Aiming at the rainfall estimation of a target area, the important focus is the rainfall of the road surface, and a radial basis function psi is reasonably selectedl(pi) The central point xi of the road surface rainfall estimation value can be ensured to be more accurate. Taking a road surface as a center, giving a central point xi and selecting the central point xi mainly comprises two parts:
(1) the formula for selecting the central points of the two sides of the highway is as follows:
wherein, i is 1, …, nsRepresenting the number of stationary sensors, n representing the selected number of center points, d1Indicating the length of the road in the target area, d2The width of the two sides of the highway is shown, m represents the total number of rows where the center points of the two sides of the highway are located, and j represents the jth row.
(2) The formula for selecting the central point on the road surface is as follows:
wherein, i is 1, …, nmIndicating the number of motion sensors.
The vector form of the radial basis function may be expressed as psi (P)T=[ψ1(P),…,ψl(P)]Then the sampling set of the rainfall in the target area can be expressed as
φ(P,t)=ψ(P)Tx(t)
Wherein x (t) ═ x1(t),…xl(t)]TRepresenting the weight coefficients of the radial basis functions.
S4, combining the weight coefficient in surface fitting with the sampling position of the sensor network according to a Kalman filtering algorithm to form a state variable in Kalman filtering, then reasonably selecting related parameters of an observation equation in the Kalman filtering, improving a Kalman filtering state estimation algorithm, and finally estimating the rainfall at any point in a target area according to the Kalman filtering state variable at the current moment, wherein the method specifically comprises the following steps:
s401, estimating state x ═ x in model1(t),…xl(t)]TCan be obtained from a linear dynamic system:
x(k+1)=Ax(k)+ω(k)
wherein ,representing system matrix, noiseIs white gaussian noise with zero mean and the variance is denoted as W.
S402, for each sampling device, the set of sampling values in the execution period is expressed as:
wherein ,NiRepresents the sampling times of the ith device in one execution cycle, and the total sampling times is
On this basis, the following observation equation is given:
wherein, I is 1,2, …, n, InRepresents an n-dimensional identity matrix of the cell,representing the observation error, assuming zero mean white Gaussian noise, the variance is denoted as V, and the noise is denoted asIs white gaussian noise with zero mean value,representing system moments
A matrix of the system is represented,represents the observation matrix, whose j-th row represents the following:
Cj=Ψ(pi)=[ψ1(pi)ψ2(pi)…ψl(pi)]
observing noiseThe correlation matrix is related to the system noise ω (k-1) and is denoted as S (k) as follows:
wherein ,Si(k) A correlation matrix representing the observed noise of the ith sampling device and the system noise,a matrix of the system is represented,representing the observation matrix, W being noiseThe variance of (c).
S403, giving a Kalman gain matrix Ki(k) The specific calculation is as follows:
wherein ,Ki(k) For the Kalman gain matrix obtained at time k, Q (k | k-1) is the estimated prediction covariance matrix at time k-1, Si(k) Is a sampling noise covariance matrix;
s404, setting the state variable at the initial time asCalculating the state variable of the next moment
wherein ,take the value of [ -10,10]Random number between, Ki(k) Kalman gain matrix, y, obtained for time ki(k) The sample value representing the time instant k is,andthe state estimation values of the next time and the current time are respectively.
S405, setting an error covariance matrix Q (0|0) in a Kalman filtering algorithm, and updating an estimated prediction covariance matrix Q (k | k-1);
Q(k|k-1)=AQ(k-1|k-1)AT+W
wherein Q (k-1| k-1) represents an error covariance matrix at the time of k-1,representing the system matrix, W being noiseThe variance of (c).
S406, setting an initial value of the noise covariance matrix S (0), and calculating an error covariance matrix Q (k | k);
wherein Q (k | k-1) is an estimated prediction covariance matrix,representing the system matrix, W being noiseVariance of (S)i(k) Is a sampled noise covariance matrix.
And S5, according to the gradient descent method, further optimizing the estimation result of the rainfall in the region by means of adjusting the sampling position of the mobile sensor.
S501, enabling J to represent a performance cost function of an estimation effect, and estimating error variance by meanDefining;
the cost function is as follows:
on the basis of the method, the mobile sensor network estimates the state information of the target area in each execution period.
S502, adjusting the position of the mobile sensor network by a gradient descent method, so as to reduce an error cost function and obtain a better estimation effect;
the specific control strategy is as follows:
pi(k+1)=pi(k)+fi(k)
wherein ,KfA positive gain is indicated by 10.
The method is simplified and can be obtained:
wherein O represents a target region, W is a system noise covariance, J represents an estimation error cost function, and piDenotes the sampling position of the sensor, Q is the estimation error covariance matrix, ψ (p) denotes the radial basis function, K denotes the kalman gain matrix, and S denotes the noise covariance matrix.
S503, obtaining a sampling position p at the next moment according to a control strategy of the sensori(k +1), the updated set of sample values for each sampling device is represented as:
wherein ,Ki(k +1) is the Kalman gain matrix obtained at the time k + 1, yi(k +1) represents the sample value at time k + 1,andthe state estimation values at the time k +1 and the current time are respectively.
S505, updating the estimated prediction covariance matrix Q (k +1| k) at the moment k + 1;
Q(k+1|k)=AQ(k|k)AT+W
wherein Q (k | k) represents an error covariance matrix at time k,representing the system matrix, W being noiseThe variance of (c).
S506, calculating a k +1 moment error covariance matrix Q (k +1| k + 1);
wherein Q is(k +1| k) estimates the prediction covariance matrix for time k + 1,representing the system matrix, W being noiseVariance of (S)iAnd (k +1) is a sampling noise covariance matrix at the k +1 moment.
S508, calculating the variance of the error estimated by the meanAnd if the defined cost function J meets the precision, ending the algorithm, otherwise, returning to the step S1.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 4, the result of the estimated simulation of the rainfall in the target road area of 10km × 10 km. The highway pavement is represented by a rectangular area of the graph, a colored background in the area represents the rainfall of the pavement, wherein the rainfall value corresponding to each color is displayed by a color amplitude corresponding table beside the area, for example, a yellow area indicates that the rainfall in the area is higher, a red star in the graph represents a sampling position of a fixed sensor, a yellow circle represents a sampling position of an initial moment of a moving sensor, a moving track of the moving sensor is represented by a blue line, and a blue circle represents a final position of the moving sensor.
From the simulation results, it is found that the region with high rainfall is concentrated on the right side of the road in the target region. As time goes by, as shown in fig. 4b and 4c, the rainfall value of the target area is higher and higher, and the area with higher rainfall moves to the center of the road, and the simulation result shows that the road surface rainfall estimation method based on the dynamic and static mixed sensors can estimate the rainfall distribution under the time-varying condition in real time.
Referring to fig. 5, the estimation error of the rainfall estimation method of the present invention is shown. According to the simulation result, the estimation error keeps converging with the increase of time, and the method for estimating the rainfall on the road surface can accurately estimate the rainfall distribution under the time-varying condition.
Therefore, the rainfall estimation method can effectively realize the dynamic estimation of the rainfall information distribution of the given road.
In summary, according to the method for estimating the rainfall distribution of the road surface based on the dynamic and static hybrid sensors, the dynamic and static hybrid sensors are used for collecting the regional information, the sampling position is more flexible, and the method is not limited to two sides of the road. Then, the sampling period of the estimation system is adjusted in real time through a designed dynamic sampling mechanism, so that the operation efficiency of the algorithm disclosed by the invention can be effectively improved, and the calculation load of the estimation system is reduced. On the basis, aiming at the rainfall estimation of a target area, the rainfall of a road surface is focused, the position of a central point in a radial basis function is reasonably selected to be mainly concentrated on the road surface, the accuracy of the estimation algorithm is effectively improved, then an improved Kalman filtering algorithm is introduced, a state variable in the Kalman filtering algorithm is formed by combining a weight coefficient of surface fitting and a sampling position, and the rainfall of any point in the target area is estimated. And finally, the sampling position of the mobile sensor is automatically adjusted by adopting a gradient descent method, so that the accuracy is further improved.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.
Claims (10)
1. A method for estimating rainfall distribution on a road surface is characterized by comprising the following steps:
s1, constructing a sampling database containing sampling positions and road rainfall according to sampling information of sampling points in the target area, and initializing parameters;
s2, adopting a dynamic and static mixed sensor sampling model under a self-triggering mechanism, defining a sampling threshold value delta of the rainfall on the road surface by normalization according to the error between the sampling value of the sensor in the current execution period and the average sampling value in the previous execution period, dividing the threshold value delta of the rainfall on the road surface to form a plurality of sampling error intervals, determining the sampling intervals of the sensor in different sampling error intervals, and dynamically updating the sampling data in the sampling database of the step S1;
s3, designing a road surface rainfall estimation model according to the dynamically updated sampling data of the step S2 by adopting a surface fitting method;
s4, combining the weight coefficient of the rainfall estimation model of the road surface after the curved surface is fitted in the step S3 with the sampling position of the sensor network by using a Kalman filtering algorithm to form a state variable in Kalman filtering, then adjusting parameters of an observation equation in the Kalman filtering algorithm, improving the Kalman filtering state estimation algorithm, and estimating the rainfall at any position in the target area according to the state variable of the Kalman filtering at the current moment;
and S5, optimizing the estimation result of the rainfall in the target area in the step S4 by means of adjusting the sampling position of the mobile sensor according to a gradient descent method, and finishing the rainfall distribution estimation.
2. The method according to claim 1, wherein in step S1, the sampling information of the sampling points in the target area is:
Y=[S,M]T
wherein ,the sampling information of the stationary sensor is represented,the data set representing the sampled information of the mobile sensor, the region estimation algorithm is:
D={P,Y}
wherein D is a sampling data set containing sampling position information, and P is a sensor sampling position set.
3. The method according to claim 1, characterized in that in step S2, the estimated execution period of the sensor network is separated from the information sampling period, the sampling times of the sensors in a fixed execution period are determined, given the fixed sensor execution period, the error between the current sampling value and the sampling average value of the previous execution period in one execution period is used as the road rainfall threshold value Δ; normalizing the threshold value delta according to the sampling interval tau i of the current momentmAnd adjusting the sampling time of the device sampling, and updating the sampling data Y of the estimation system.
4. A method according to claim 3, characterized in that the threshold value Δ is expressed as follows:
wherein ,the ith sample value representing the mth execution cycle,represents the mean of the samples of the m-1 execution cycle; the normalization process for the threshold value Δ is as follows:
wherein, Delta is more than 0, and the value range of e is [0,1]](ii) a Divide e into three segments, sampling intervalRespectively, the upper limit and the lower limit ofAndsampling intervalThe change rule according to e is as follows:
wherein ,0≤c1<c2Is less than or equal to 1, and theta is the minimum step length for increasing the sampling interval.
5. The method according to claim 1, wherein in step S3, for the rainfall estimation of the target area, radial basis functions are defined, and the radial basis center points on both sides of the road and on the road surface are determined with the road surface as the center, and the radial basis functions are specifically:
wherein ,ψj(P) representing the jth radial basis function in the set of radial basis function vectors; l is the dimension of the vector set, σjIs the width of the radial basis function, betajIs a normalization constant, ξjIs the center point of the radial basis function and P represents any point within the target region.
6. The method of claim 5, wherein the center points of the two sides of the roadway are selected as follows:
wherein, i is 1, …, nsRepresenting the number of stationary sensors, n representing the selected number of center points, d1Indicating the length of the road in the target area, d2Representing the width of two sides of the highway, wherein m represents the total row number of the central points of the two sides of the highway, and j represents the jth row;
the central point on the road surface is selected as follows:
wherein, i is 1, …, nmIndicating the number of motion sensors;
the vector form of the radial basis function is denoted as psi (P)T=[ψ1(P),…,ψl(P)]The sampling set of rainfall in the target area is expressed as:
φ(P,t)=ψ(P)Tx(t)
wherein x (t) ═ x1(t),…xl(t)]TRepresenting the weight coefficients of the radial basis functions.
7. The method according to claim 1, wherein step S4 is specifically:
s401, taking a weight coefficient x in the radial basis function as a state variable in a linear road rainfall estimation model, wherein a linear dynamic system comprises the following components:
x(k+1)=Ax(k)+ω(k)
wherein ,representing system matrix, noiseThe white Gaussian noise is zero mean value, and the variance is recorded as W;
s402, determining a sampling value set of each sampling device in an execution cycleGiving an observation equation Si(k) The following were used:
wherein ,a matrix of the system is represented,representing the observation matrix, W being noiseThe variance of (a);
s403, setting a Kalman gain matrix Ki(k) The following were used:
wherein ,Ki(k) For the Kalman gain matrix obtained at time k, Q (k | k-1) is the estimated prediction covariance matrix at time k-1, Si(k) Is a sampling noise covariance matrix;
s404, setting the state variable at the initial time asCalculating the state variable of the next moment
S405, setting an initial value of an error covariance matrix in a Kalman filtering algorithm to be Q (0|0), and updating an estimated prediction covariance matrix;
s406, setting an initial value of the noise covariance matrix S (0), and calculating a covariance matrix Q (k | k);
8. The method according to claim 1, wherein in step S5, the control strategy for optimizing the rainfall estimation result in the target area of step S4 is:
wherein O is a target region, W is a system noise covariance, J is an estimation error cost function, and piFor the sampling position of the sensor, Q is the estimation error covariance matrix, psi (p) is the radial basis function, K is the Kalman gain matrix, and S is the noise covariance matrix.
9. A computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform any of the methods of claims 1-8.
10. A computing device, comprising:
one or more processors, memory, and one or more programs stored in the memory and configured for execution by the one or more processors, the one or more programs including instructions for performing any of the methods of claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010864891.0A CN112036630B (en) | 2020-08-25 | 2020-08-25 | Highway pavement rainfall distribution estimation method, storage medium and computing equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010864891.0A CN112036630B (en) | 2020-08-25 | 2020-08-25 | Highway pavement rainfall distribution estimation method, storage medium and computing equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112036630A true CN112036630A (en) | 2020-12-04 |
CN112036630B CN112036630B (en) | 2023-08-04 |
Family
ID=73581338
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010864891.0A Active CN112036630B (en) | 2020-08-25 | 2020-08-25 | Highway pavement rainfall distribution estimation method, storage medium and computing equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112036630B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0961546A (en) * | 1995-08-21 | 1997-03-07 | Toshiba Corp | Device for predicting amount of rainfall |
JPH1123727A (en) * | 1997-07-08 | 1999-01-29 | Nippon Telegr & Teleph Corp <Ntt> | Maximum likelihood estimation device for rainfall amount |
JP2011002266A (en) * | 2009-06-17 | 2011-01-06 | Nec Corp | Target tracking processor and method for correcting error covariance matrix used therefor |
CN102183802A (en) * | 2011-03-10 | 2011-09-14 | 西安交通大学 | Short-term climate forecast method based on Kalman filtering and evolution modeling |
CN106526542A (en) * | 2016-10-17 | 2017-03-22 | 西南大学 | Augmentation Kalman filtering method based on deterministic sampling |
WO2020105812A1 (en) * | 2018-11-22 | 2020-05-28 | 제주대학교 산학협력단 | Prediction system and method on basis of parameter improvement through learning |
-
2020
- 2020-08-25 CN CN202010864891.0A patent/CN112036630B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0961546A (en) * | 1995-08-21 | 1997-03-07 | Toshiba Corp | Device for predicting amount of rainfall |
JPH1123727A (en) * | 1997-07-08 | 1999-01-29 | Nippon Telegr & Teleph Corp <Ntt> | Maximum likelihood estimation device for rainfall amount |
JP2011002266A (en) * | 2009-06-17 | 2011-01-06 | Nec Corp | Target tracking processor and method for correcting error covariance matrix used therefor |
CN102183802A (en) * | 2011-03-10 | 2011-09-14 | 西安交通大学 | Short-term climate forecast method based on Kalman filtering and evolution modeling |
CN106526542A (en) * | 2016-10-17 | 2017-03-22 | 西南大学 | Augmentation Kalman filtering method based on deterministic sampling |
WO2020105812A1 (en) * | 2018-11-22 | 2020-05-28 | 제주대학교 산학협력단 | Prediction system and method on basis of parameter improvement through learning |
Non-Patent Citations (3)
Title |
---|
T. WARDAH等: "WRF model input for improved radar rainfall estimates using Kalman Filter", 2014 INTERNATIONAL SYMPOSIUM ON TECHNOLOGY MANAGEMENT AND EMERGING TECHNOLOGIES * |
张岳: "多普勒天气雷达定量估测降水方案改进", 中国优秀硕士学位论文全文数据库, no. 1 * |
黄伯强;: "基于遗传算法优化BP神经网络的降雨量预测", 信息与电脑(理论版), no. 11 * |
Also Published As
Publication number | Publication date |
---|---|
CN112036630B (en) | 2023-08-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110991272B (en) | Multi-target vehicle track recognition method based on video tracking | |
CN107341819B (en) | Target tracking method and storage medium | |
CN110415277B (en) | Multi-target tracking method, system and device based on optical flow and Kalman filtering | |
CN111192284B (en) | Vehicle-mounted laser point cloud segmentation method and system | |
CN112347993B (en) | Expressway vehicle behavior and track prediction method based on vehicle-unmanned aerial vehicle cooperation | |
CN104318258A (en) | Time domain fuzzy and kalman filter-based lane detection method | |
CN104517275A (en) | Object detection method and system | |
CN105046714A (en) | Unsupervised image segmentation method based on super pixels and target discovering mechanism | |
CN112731334A (en) | Method and device for positioning vehicle by laser | |
Yuan et al. | Image feature based GPS trace filtering for road network generation and road segmentation | |
CN109447161A (en) | A kind of real-time incremental and adaptive clustering scheme based on car radar data | |
CN115545281A (en) | Risk scene identification method, device, equipment and storage medium | |
CN113176562B (en) | Multi-target tracking method, device, electronic equipment and readable storage medium | |
CN114609609A (en) | Speed estimation method for extracting static point cloud by FMCW laser radar random sampling | |
Busch et al. | Classification of road surface and weather-related condition using deep convolutional neural networks | |
CN112036630B (en) | Highway pavement rainfall distribution estimation method, storage medium and computing equipment | |
Jin et al. | Road friction estimation method based on fusion of machine vision and vehicle dynamics | |
CN114730494A (en) | Method for estimating the coverage of a space of a traffic scene | |
CN112163521A (en) | Vehicle driving behavior identification method, device and equipment | |
CN115691140B (en) | Analysis and prediction method for space-time distribution of automobile charging demand | |
CN114639238B (en) | Urban expressway traffic state estimation method based on normalized spectral clustering algorithm | |
CN116628531A (en) | Crowd-sourced map road object element clustering method, system and storage medium | |
CN110619254A (en) | Target tracking method and device based on disparity map and terminal | |
Galanis et al. | Weather-based road condition estimation in the era of Internet-of-Vehicles (IoV) | |
CN112200831A (en) | Dense connection twin neural network target tracking method based on dynamic template |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |