Vehicle key driving state estimation method based on low-orbit satellite and vehicle-mounted sensing multi-source information fusion

Abstract

The invention discloses a vehicle key driving state estimation method based on low-orbit satellite and vehicle-mounted sensing multi-source information fusion, which comprises the following steps: acquiring the real-time longitudinal speed of a vehicle by combining a Doppler frequency shift formula based on a low-orbit satellite measurement signal; acquiring real-time lateral force of the tire based on the intelligent tire measurement signal in combination with a Gaussian process regression algorithm; a tire slip angle is calculated based on a vehicle-mounted sensing signal, a tire cornering stiffness estimation algorithm is established in combination with lateral force prior information, and a real-time cornering stiffness adaptive value is obtained; vehicle parameters and real-time sensing signals are obtained, and a vehicle dynamics model is established in combination with the self-adaptive value of the cornering stiffness; a lateral force estimation value provided by an intelligent tire system, a longitudinal vehicle speed value measured by a low-orbit satellite, a yaw velocity and a transverse acceleration sensing measurement value serve as measurement signals, an innovation self-adaptive expansion Kalman filter is combined, and a vehicle key driving state estimation algorithm is designed. And the real-time side slip angle, the yaw velocity and the longitudinal velocity of the vehicle are obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle control, and relates to a method for estimating a key driving state of a vehicle, and in particular to a method for estimating a key driving state of a vehicle based on multi-source information fusion of low-orbit satellites and vehicle-mounted sensors. Background Art

[0002] The key driving states of a vehicle (including longitudinal speed, sideslip angle and yaw rate) are important parameters describing the vehicle's motion characteristics, among which the longitudinal speed (v x ) is used to adjust the vehicle's acceleration, braking and cruise control, and is the core input for the longitudinal control of autonomous vehicles; the sideslip angle (β) describes the angle between the vehicle's actual driving direction and the longitudinal axis of the vehicle body, and is an important parameter for lateral stability analysis; the yaw rate (γ) describes the angular velocity of the vehicle's rotation around the center of mass, and is an important control variable for trajectory tracking and path planning.

[0003] Traditional state estimation methods usually rely on the acceleration and angular velocity data provided by on-board sensors (such as the on-board inertial measurement unit IMU), combined with the vehicle dynamics model to calculate the key driving state of the vehicle. However, these methods often fail to meet the estimation requirements under special driving conditions (such as slippery roads, high-speed curves, or emergency obstacle avoidance). The main reason is that it is difficult to accurately and concisely express tire forces. In other words, although the use of a simplified tire model can accurately describe the tire forces under most driving conditions, when the vehicle is in special driving conditions, the simplified tire model is often difficult to accurately characterize the actual movement behavior of the tire. Although the use of a more complex tire model can improve the accuracy of the tire force representation under special driving conditions, this will also increase the complexity of the vehicle driving state estimation system, making it difficult to meet the real-time requirements of the vehicle control system.

[0004] Smart tire technology is to install sensors inside the tire, directly measure the dynamic parameters of the tire during movement, and combine machine learning or theoretical models to estimate the key state of the tire. Compared with traditional tire state estimation methods, this technology can obtain more accurate and reliable tire states. Therefore, if the prediction of tire force can be achieved based on smart tire technology, the process of traditional tire force prediction methods can be simplified, and the estimation accuracy and reliability of tire force can be improved, especially the estimation accuracy and reliability of tire force under special driving conditions, so as to achieve the purpose of improving the overall performance of the vehicle driving state estimation system. Although many scholars have proposed some tire force prediction methods based on smart tire technology, there are still some areas that can be optimized, mainly including the improvement of the signal acquisition quality of the smart tire system and the optimization of the mapping relationship between signal characteristics and tire force. In terms of signal acquisition, the current field of smart tire force prediction lacks sufficient theoretical support for the scientific selection of sensor installation positions, and it is difficult to ensure the continuous and high-quality acquisition of tire dynamic characteristics during tire operation. In terms of mapping relationship construction, the wrong or inappropriate selection of mapping methods will also lead to poor performance of the final tire force prediction model.

[0005] Although smart tires can directly measure the lateral force and slip rate of the tire and enhance the estimation of the lateral dynamics of the vehicle, their performance may also be limited by road conditions and tire contact conditions. Against the backdrop of the rapid development of satellite technology, low earth orbit (LEO) satellites have become a disruptive innovation. These satellites are located at an altitude of about 100 to 500 miles above the earth's surface, completely changing our communication methods, data collection methods and earth monitoring modes. At the same time, the rapid development of LEO technology has also made it possible to obtain high-precision vehicle longitudinal speed and position information. As the key to estimating vehicle speed, the accurate measurement of longitudinal vehicle speed can obviously further improve the accuracy and reliability of vehicle driving state estimation.

[0006] However, neither academia nor industry has yet proposed a vehicle key driving state estimation solution based on multi-source information fusion of low-orbit satellites (for longitudinal vehicle speed measurement) and on-board sensors (traditional on-board sensors + intelligent tire sensors). Summary of the invention

[0007] In view of the deficiencies in the above-mentioned prior art, the purpose of the present invention is to provide a vehicle key driving state estimation method based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors, taking the intelligent tire lateral force estimation value, the low-orbit satellite longitudinal vehicle speed measurement value, and the yaw angular velocity and lateral acceleration measurement values ​​as measurement signals, and combining the new information adaptive extended Kalman filter to design a vehicle key driving state estimation model, obtain the vehicle's real-time center of mass sideslip angle, yaw angular velocity and longitudinal speed, realize the vehicle's key driving state estimation based on multi-source information fusion, and improve the accuracy and reliability of vehicle driving state estimation.

[0008] The technical solution adopted by the present invention is as follows:

[0009] The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors is characterized by comprising the following steps:

[0010] S1. Based on the low-orbit satellite measurement signal and the Doppler frequency shift formula, the real-time longitudinal velocity of the vehicle is obtained;

[0011] S2, based on the intelligent tire measurement signal, combined with the Gaussian process regression algorithm, obtain the real-time tire lateral force;

[0012] S3. Based on the real-time lateral force of the tire and combined with the vehicle-mounted sensor information, a tire cornering stiffness adaptive algorithm is designed to obtain a real-time cornering stiffness adaptive value;

[0013] S4, acquiring vehicle parameters and real-time sensor signals, and establishing a vehicle dynamics model in combination with the real-time cornering stiffness adaptive value;

[0014] S5. Establish a key driving state estimation algorithm for the vehicle based on the vehicle dynamics model and the new information adaptive extended Kalman filter to obtain the real-time vehicle longitudinal speed, center of mass sideslip angle and yaw angular velocity;

[0015] Further, S1 comprises the following steps:

[0016] S1-1, based on the low-orbit satellite ephemeris data received by the vehicle-mounted receiver, calculate the position of each satellite in the geodetic coordinate system;

[0017] S1-2, based on the position of each low-orbit satellite, determine the visible range of each low-orbit satellite to obtain the intersection of the visible range, and divide the grid to select candidate initial value points;

[0018] S1-3. Use the relative motion between the vehicle and the satellite to establish the Doppler frequency shift equation, linearize the Doppler frequency shift equation, and substitute the candidate initial value points; use the least squares method to solve the real-time position of the vehicle and calculate the real-time longitudinal velocity of the vehicle.

[0019] Furthermore, the constructed Doppler frequency shift equation is expressed as:

[0020]

[0021] Where Δf i is the ith frequency shift observation value, f0 is the satellite’s signal transmission frequency, P u represents the desired position of the vehicle, P s is the satellite position vector, v u represents the desired speed of the vehicle, v srepresents the speed of the satellite, c represents the signal propagation speed, τ i represents the frequency shift measurement error;

[0022] Linearize the Doppler frequency shift equation and substitute the candidate initial value points into it. The Doppler observation equation is approximated as:

[0023]

[0024] Where x is the variable to be determined; Δx is the vehicle position and speed correction to be determined, denoted as Δx=(Δx u ,Δy u ,Δz u ,Δv x , u ,Δv y,u ,Δv z,u ), Δx u , Δy u , Δz u , Δv x,u , Δv y,u , Δv z,u They are respectively the vehicle longitudinal position correction, lateral position correction, vertical position correction, longitudinal speed correction, lateral speed correction and vertical speed correction; Represents the Jacobian matrix of the observation equation for the variable to be determined; v u (0) is the initial velocity vector of the ground vehicle.

[0025] Furthermore, taking the subsatellite point of each low-orbit satellite as the center, calculate its visible radius r on the ground v :

[0026]

[0027] The visible projection range of each low-orbit satellite in the longitude and latitude directions is:

[0028]

[0029] Take the visible range of all low-orbit satellites and calculate their intersection area, which is represented as a rectangle:

[0030]

[0031] Divide the intersection matrix into N×N grids and use the center point of each grid as a candidate initial value point:

[0032]

[0033] Where R is the radius of the Earth, h s is the altitude of satellite s, L lonis the vertical boundary of the satellite visible projection range, L lat is the lateral boundary of the satellite visible projection range, ΔΦ is the span of the rectangular area in the latitude direction, Φ min is the southernmost boundary of the intersection area, that is, the lowest latitude value within the rectangular range, Φ max is the northernmost boundary of the intersection area, that is, the highest latitude value within the rectangular range, Δλ is the span of the rectangular area in the longitude direction, and λ min is the westernmost boundary of the intersection area, that is, the lowest longitude value within the rectangular range, λ max is the easternmost boundary of the intersection area, that is, the highest longitude value within the rectangular range, Φ i is the latitude value of the i-th row grid, N is the number of grid divisions, λ j is the longitude value of the j-th column grid.

[0034] Further, S2 comprises the following steps:

[0035] S2-1. Establish a PVDF smart tire finite element model, verify the accuracy of the modeling method through real vehicle tests, and build an array PVDF smart tire finite element model on this basis;

[0036] S2-2, determine the influencing factors and value range of the output voltage of the smart tire PVDF sensor array according to daily driving conditions, and generate a sample matrix S in combination with the Sobol sequence method;

[0037] The sample matrix S is decomposed into 2+D sub-sample matrices, where D represents the number of factors affecting the output voltage of the array piezoelectric film sensor;

[0038] S2-3, quantify the z installation positions of the circumferential sensor array on the inner liner layer using the tire transverse pattern, and install z PVDF sensor arrays to form a new intelligent tire finite element model. On this basis, use the new intelligent tire finite element model to calculate 2+D sub-sample matrices, and extract the output voltages of the z sensor arrays and their corresponding tire lateral forces to form a voltage data set Data_VT1 and a lateral force data set Data_Fy;

[0039] S2-4. The average value of the first-order main effect index of the output voltage values ​​of all sensors in the g-th PVDF sensor array is taken as the first-order main effect index of the g-th installation position, denoted as Arrange the g first-order main effect indicators in descending order, and select the first t installation positions as the optimal installation positions of the PVDF sensor array;

[0040] S2-5, based on the determined t sensor array installation positions, the g-th PVDF sensor array data Data_VT1 gFurther extraction is performed, and l specific voltage values ​​of each sensor array voltage data are selected as input features; the selection logic is: taking the maximum voltage in the sensor array under each working condition as a reference, c voltage value points are selected along the direction of voltage decrease on both sides of the point, and finally l voltage value points are selected as input features; the processed data is recorded as Data_VT2 g ;

[0041] The extracted data sets Data_Fy and Data_VT2 g Divided into training set and test set, the training set is denoted as Data_Fy x and Data_VT2 g_x , the test set is denoted as Data_Fy c and Data_VT2g c ;

[0042] S2-6, according to Data_Fy and Data_VT2 g Based on the data characteristics, the mean function m(x) and covariance function k(x,x') types are determined by empirical methods.

[0043] The voltage data and tire lateral force of the training data set are expressed in the form of Gaussian process distribution, that is, the lateral force Data_Fy x The function f(Data_VT2 g_x ) is subject to the mean value m(Data_VT2 g_x ), covariance is k(Data_VT2 g_x ,Data_VT2 g_x '); on this basis, the initial values ​​of the hyperparameters and the number of iterations in the Gaussian process regression model are determined, and the hyperparameters σ and F in the covariance function are determined by minimizing the negative logarithm of the marginal likelihood function by the gradient descent method. The negative logarithm of the marginal likelihood function is specifically expressed as:

[0044]

[0045] Where -logp(f|x) represents the negative logarithm of the marginal likelihood function, f is the Gaussian process regression function, and x is the voltage input feature. For fitting the training data, n is the dimension of the training data set, k n is the n-dimensional covariance function, is the n-dimensional variance vector, I n is an n-dimensional unit vector, is a penalty item, is the normalization constant;

[0046] According to the prior definition, the joint prior distribution can be obtained by combining the training set and test set collected in S2-1:

[0047]

[0048] Where, f(Data_VT2 g_x ) is the tire lateral force corresponding to the training set, f(Data_VT2 g_c ) is the tire lateral force corresponding to the test number, and N is a normal distribution;

[0049] S2-7, calculate the estimated value f(Data_VT2 g_c ) and substitute it into S2-6 to obtain the hyperparameter solution for the real-time lateral force of the tire.

[0050] Furthermore, the expression of the first-order main effect index is:

[0051]

[0052] in, Represents the first-order main effect index of the Δth influencing factor on the voltage output of the gth sensor array, 1≤Δ≤D, Δ=5 represents the influencing factor is the tire side slip angle, D is the number of influencing factors; Mean(Data_VT1(A) i ,Data_VT1(B) i ) represents the average value of the numerical vector formed by combining the i-th column of the sensor array output voltage matrix corresponding to matrix A and the i-th column of the sensor array output voltage matrix corresponding to matrix B, represents the jth value of the i-th column of the sensor array output voltage matrix corresponding to matrix B, Represents the matrix AB Δ The corresponding sensor array output voltage matrix has the jth value in the i-th column, Represents the jth value in the i-th column of the sensor array output voltage matrix corresponding to matrix A.

[0053] Further, step S3 includes the following steps:

[0054] S3-1. Obtain the front and rear wheelbase l of the vehicle f and l r , calculate the vertical loads on the four wheels when the vehicle is stationary, including the static vertical loads Ld of the left front, right front, left rear and right rear tires fl , Ld fr , Ld rl and Ld rr ;

[0055] The tire load is Ld fl and Ld rlThe cornering characteristic test under the conditions is used to obtain the cornering angle-side force data, and the fixed value K of the front and rear wheel cornering stiffness is obtained by combining the magic tire formula of the side force and the curve fitting method. f and K r ;

[0056] S3-2, define the two front wheel steering angles to be the same and equal to δ, on this basis, obtain the vehicle sensor signal to calculate δ; define the two front axle wheels’ side slip angles to be the same and equal to α f , the side slip angles of the two wheels on the rear axle are the same and equal to α r ;

[0057] S3-3, using the difference between the linear tire force and the intelligent tire lateral force estimation value to adaptively adjust the tire cornering stiffness; the adaptive values ​​of the front and rear tire cornering stiffness can be expressed as:

[0058]

[0059] in, and are the adaptive values ​​of the cornering stiffness of the front and rear wheels, and are the sum of the estimated lateral forces of the two front wheels and the two rear wheels respectively; F yf and F yr are the sum of the linear tire forces of the two front wheels and the two rear wheels respectively.

[0060] Furthermore, the fixed values ​​of the front and rear tire cornering stiffness can be expressed as:

[0061]

[0062] Among them, B f , C f and D f is the load Ld fl or Ld fr The corresponding parameters obtained by fitting the side deviation characteristic data, B r , C r and D r is the load Ld rl or Ld rr Parameters obtained by fitting the corresponding cornering characteristic data.

[0063] Further, S4 specifically includes the following steps:

[0064] S4-1, obtain vehicle parameters, including vehicle mass, vehicle yaw moment of inertia, distance from vehicle center of mass to front axle and distance to rear axle; obtain real-time signals from vehicle sensors, including vehicle longitudinal acceleration and lateral acceleration, yaw angular velocity and steering wheel angle;

[0065] S4-2, construct a vehicle dynamics model based on the real-time adaptive values ​​of the front and rear wheel cornering stiffness, vehicle parameters and real-time signals of the on-board sensors, which is specifically expressed as follows:

[0066]

[0067] Among them, a x is the longitudinal acceleration of the vehicle, I z is the vehicle yaw moment of inertia, is the longitudinal acceleration, v x is the longitudinal velocity of the vehicle, γ is the yaw rate, is the yaw angular acceleration, β is the sideslip angle of the vehicle center of mass, m is the vehicle mass, and are the adaptive values ​​of the cornering stiffness of the front and rear wheels, respectively, f and l r are the front and rear wheelbases of the vehicle, I z is the moment of inertia of the vehicle around the z-axis.

[0068] Further, step S5 includes the following steps:

[0069] S5-1. Define the state vector J, input vector U and observation vector Z of the vehicle key driving state estimation system, which are expressed as follows:

[0070]

[0071] Among them, a x is the longitudinal acceleration of the vehicle, a y is the lateral acceleration of the vehicle, γ is the yaw rate, v x is the longitudinal velocity of the vehicle, β is the sideslip angle of the vehicle center of mass, and γ is the yaw rate;

[0072] S5-2. Define the state prior estimation equation P(·) and the observation estimation equation H(·), which are expressed as follows:

[0073]

[0074] Among them, P1, P2 and P3 represent the prior estimates of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively, and v x0 , β0 and γ0 represent the initial values ​​of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively. s is the system sampling time; and are the sum of the estimated values ​​of the lateral forces of the two front wheels and the two rear wheels respectively;

[0075] S5-3. Establish a noise covariance matrix adaptive adjustment strategy based on sliding window self-adjustment to obtain the noise covariance adaptive matrix. First, define the covariance matrix of the new information as:

[0076] C k =E[v k v k T ]

[0077] Among them, v k is the error between the actual value and the predicted value of the measured variable in the kth step, k>0; E is the mathematical expectation;

[0078] The actual covariance matrix of the innovation is defined as:

[0079]

[0080] Among them, M represents the length of the sliding window, expressed as:

[0081]

[0082] Among them, int(·) is the rounding function, is a constant;

[0083] Noise Covariance Adaptive Matrix The expression is written as:

[0084]

[0085] Where, trace(·) is the trace function, and R is the noise covariance matrix;

[0086] S5-4, based on the extended Kalman filter principle and the first-order Taylor series expansion theorem, the calculation process parameters, including the Jacobian matrix Pm and Hm, and the prior estimation error covariance matrix Pf and the Kalman gain Km, are expressed as:

[0087]

[0088] Pf k =Pm k ·Pl k-1 ·Pm k T +Q

[0089]

[0090] Among them, Pm k is the Jacobian matrix of the k-th step state transfer equation, Hm k is the Jacobian matrix of the k-th step measurement equation, Pl k-1is the error covariance matrix of the k-1th step, Q is the process noise covariance matrix, and its parameters can be manually calibrated by trial and error;

[0091] S5-5, based on the state prior estimation information, the measurement estimation information and the sensor measurement information, the key driving state of the vehicle is estimated, and the error covariance matrix Pl is updated. The formula is specifically expressed as follows:

[0092]

[0093] Pl k =(I3-Km k ·Hm k )·Pl k

[0094] Where I is the identity matrix, J k Km is the estimated value of the key driving state of the vehicle in the kth step, including the vehicle longitudinal speed, center of mass sideslip angle and yaw rate; k is the Kalman gain at the kth step.

[0095] Beneficial effects:

[0096] (1) The present invention establishes a finite element model of an array-type PVDF intelligent tire based on finite element software, which greatly reduces the time and financial costs required for the scientific selection of the installation position of the PVDF sensor array in the tire and the extraction of training data for the tire lateral force estimation algorithm.

[0097] (2) For the established array-type PVDF intelligent tire, the present invention also measures the sensitivity of the lateral force change during tire operation to the output voltage of the PVDF sensor array based on the Sobol sensitivity analysis method, and determines the optimal installation position of the sensor array based on this, which can improve the signal acquisition capability of the intelligent tire system.

[0098] (3) Based on the array PVDF type intelligent tire technology, the present invention combines the Gaussian process regression algorithm to propose a tire lateral force estimation method with low cost, high accuracy, good stability and strong generalization.

[0099] (4) Based on the tire lateral force prior information provided by the array PVDF type intelligent tire system, the present invention combines the on-board sensor information and vehicle dynamics methods to design a tire lateral stiffness adaptive algorithm, which is beneficial to improving the accuracy of the vehicle model built based on this.

[0100] (5) Based on the vehicle state information measured by low-orbit satellites and on-board sensors, as well as vehicle dynamics methods, the present invention combines the principle of new information adaptive extended Kalman filter to design a high-precision and high-reliability vehicle key driving state estimation method.

[0101] (6) Based on the tire lateral force estimation value provided by the intelligent tire system and the tire lateral force calculated by the vehicle dynamics method, combined with the Kalman filter innovation adaptive principle, a sliding window self-adjusting noise covariance matrix adaptive adjustment strategy is designed, which is beneficial to improving the estimation accuracy of the vehicle's key driving state. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 The present invention is a flowchart of an embodiment of a method for estimating a key driving state of a vehicle based on multi-source information fusion of low-orbit satellites and vehicle-mounted sensors in the present invention.

[0103] Figure 2 This is the modeling process of the piezoelectric smart tire finite element model in the patent of this invention.

[0104] Figure 3 This is a schematic diagram of the finite element model of the array PVDF smart tire in the patent of this invention.

[0105] Figure 4 Schematic diagram of all potential installation locations for the PVDF sensor array in the tire inner liner.

[0106] Figure 5 The calculation results of the first-order main effect index of different PVDF sensor array installation positions when the lateral force changes.

[0107] Figure 6 Schematic diagram of the three-freedom vehicle dynamics model.

[0108] Figure 7 This is the lateral force estimation result of the left front wheel intelligent tire when the vehicle passes through a double lane change test road at high speed on a high-adhesion road.

[0109] Figure 8 It is the estimated vehicle speed when the vehicle passes through the double lane change test road at high speed on a high-adhesion road surface.

[0110] Fig. 9 It is the estimated sideslip angle of the center of mass when the vehicle passes through a double lane change test road at high speed on a high-adhesion road.

[0111] Fig.10 It is the estimated yaw rate of the vehicle on a high-adhesion road when it passes through a double-lane test road at high speed.

[0112] Fig.11 It is the lateral force estimation result of the left front wheel intelligent tire when the vehicle passes through the double lane change test road at medium speed on a low-adhesion road.

[0113] Fig.12 It is the estimated speed of a vehicle passing a double lane change test road at a medium speed on a low-adhesion road surface.

[0114] Fig.13It is the estimated sideslip angle of the center of mass when the vehicle passes through a double lane change test road at medium speed on a low-adhesion road.

[0115] Fig.14 It is the estimated yaw rate of the vehicle on a low-adhesion road when passing through a double-lane test road at medium speed. DETAILED DESCRIPTION

[0116] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0117] Combined with Figure 1-14 As shown, the present invention proposes a method for estimating a vehicle's key driving state based on multi-source information fusion of low-orbit satellites and vehicle-mounted sensors, comprising the following steps:

[0118] S1. Based on the low-orbit satellite measurement signal and the Doppler frequency shift formula, the real-time longitudinal velocity of the vehicle is obtained;

[0119] First, based on the low-orbit satellite ephemeris data received by the vehicle-mounted receiver, the position of each satellite in the geodetic coordinate system is calculated. The acquired ephemeris data includes the three-dimensional coordinates P of each satellite. s =(X s ,Y s ,Z s ), movement speed v s and the transmission signal rate f0, where X s , Y s , Z s are the X, Y and Z coordinates of satellite s respectively.

[0120] Furthermore, the three-dimensional coordinates of the low-orbit satellite are converted into the earth coordinate system, which can be specifically expressed as:

[0121]

[0122] Among them, Φ s , s and h s are the latitude, longitude and altitude of satellite s, and R is the radius of the earth.

[0123] After obtaining the position of each LEO satellite, it is necessary to determine the visible range of each LEO satellite, calculate the intersection of the visible range, and divide the grid to select candidate initial value points to narrow the solution range.

[0124] First, taking the subsatellite point of each low-orbit satellite as the center, calculate its visible radius r on the groundv :

[0125]

[0126] Furthermore, the visible projection range of each low-orbit satellite in the longitude and latitude directions is:

[0127]

[0128] Among them, L lon is the vertical boundary of the satellite visible projection range, L lat It is the lateral boundary of the satellite visible projection range.

[0129] Furthermore, the visible ranges of all low-orbit satellites are taken, and the intersection area of ​​the visible ranges is calculated, which is represented as an intersection rectangle:

[0130]

[0131] Among them, ΔΦ is the span of the rectangular area in the latitude direction, Φ min is the southernmost boundary of the intersection area, that is, the lowest latitude value within the rectangular range, Φ max is the northernmost boundary of the intersection area, that is, the highest latitude value within the rectangular range, Δλ is the span of the rectangular area in the longitude direction, and λ min is the westernmost boundary of the intersection area, that is, the lowest longitude value within the rectangular range, λ max It is the easternmost boundary of the intersection area, that is, the highest longitude value within the rectangular range.

[0132] Furthermore, the intersection matrix is ​​divided into 200×200 grids, and the center point of each grid is used as a candidate initial value point:

[0133]

[0134] Among them, Φ i is the latitude value of the i-th row grid, N is the number of grid divisions, λ j is the longitude value of the j-th column grid.

[0135] The Doppler frequency shift equation is established by using the relative motion between the vehicle and the satellite, and the vehicle's real-time position is solved using the least squares method. The constructed Doppler frequency shift equation is:

[0136]

[0137] Where Δf i is the ith frequency shift observation value, f0 is the satellite’s signal transmission frequency, P u represents the desired position of the vehicle, P s is the satellite position vector, v urepresents the desired speed of the vehicle, c represents the signal propagation speed, τ i Represents the frequency shift measurement error. On this basis, the Doppler frequency shift equation is linearized and the candidate initial value point is substituted. At this time, the Doppler observation equation can be approximated as:

[0138]

[0139] Where x is the variable to be determined; Δx is the vehicle position and speed correction to be determined, denoted as Δx=(Δx u ,Δy u ,Δz u ,Δv x , u ,Δv y,u ,Δv z,u ), Δx u , Δy u , Δz u , Δv x,u , Δv y,u , Δv z,u They are respectively the vehicle longitudinal position correction, lateral position correction, vertical position correction, longitudinal speed correction, lateral speed correction and vertical speed correction; Represents the Jacobian matrix of the observation equation for the variable to be determined.

[0140] Furthermore, based on the vehicle position and velocity correction Δx to be solved, the least squares method is used to solve the vehicle's real-time longitudinal velocity. First, for the observations of multiple low-orbit satellites, an overdetermined set of equations is constructed:

[0141] d=H·Δx+τ (8)

[0142] Among them, d=[Δf1,Δf2,...,Δf n ] T , represents the frequency shift observation value vector, H represents the observation matrix, which contains the unit direction vectors of the satellite and the vehicle and the satellite velocity, Δx is the vehicle position and velocity correction to be solved, and τ is the observation noise.

[0143] Furthermore, the least squares method is used to solve:

[0144] Δx=(H T H) -1 H T d (9)

[0145] Furthermore, the vehicle position and speed are updated iteratively in real time:

[0146]

[0147] Finally, until the residual ||d|| meets the set threshold (the threshold in this embodiment is 0.001), extract v u The speed in the x direction is taken as the real-time longitudinal speed of the vehicle.

[0148] S2. Based on the intelligent tire measurement signal and combined with the Gaussian process regression algorithm, the real-time lateral force of the tire is obtained.

[0149] First, based on Figure 2 The modeling process shown in the figure establishes a finite element model of a smart tire with a piezoelectric film sensor installed on the inner liner, and verifies the accuracy of the modeling method through a real vehicle test platform. Figure 3 As shown, an array piezoelectric film type intelligent tire finite element model is constructed by using a circumferential array of single piezoelectric film sensors, wherein a single circumferential sensing array contains a total of 35 piezoelectric film sensors.

[0150] Furthermore, in order to make the final selected PVDF sensor array installation position have better adaptability to working conditions, it is necessary to consider as many factors affecting the output voltage as possible when designing the first-order main effect index calculation simulation scheme. In view of the purpose of the present invention is to construct a tire lateral force estimation model, the influencing factors considered include tire slip rate, side slip angle, vertical load, road adhesion coefficient, vehicle speed and tire pressure, and the value range shown in Table 1 is determined according to the daily vehicle driving state.

[0151] Table 1: Value ranges of various influencing factors

[0152]

[0153]

[0154] Furthermore, the sequence generation method is combined to generate a 200×12-dimensional sample matrix S, written as:

[0155]

[0156] Among them, S ij It is the calculated value of the upper and lower limits of the six influencing factors combined with the Sobol sequence generation algorithm.

[0157] Furthermore, the sample matrix S is decomposed into 8 sub-sample matrices, written as:

[0158]

[0159] Among them, change(B i ,A i ) represents replacing the i-th column of matrix B with the i-th column of matrix A to form a new matrix AB i, a total of 6 matrix replacement operations are performed to form 6 sub-sample matrices, where each row of each sub-sample matrix represents a set of test conditions.

[0160] In fact, the position of the PVDF sensor array on the circumference of the tire does not need to be determined, because it is always symmetrical with the longitudinal center symmetry plane of the tire. Figure 4 As shown, the present invention quantifies and names the potential installation positions of the sensor arrays based on the transverse pattern on a single tire pitch of the finite element tire model, and installs 62 PVDF sensor arrays to form a new intelligent tire finite element model. On this basis, the new model is used to calculate 8 sub-sample matrices, and the output voltages of the 62 sensor arrays and their corresponding tire lateral forces are extracted to form a voltage data set Data_VT1 and a lateral force data set Data_Fy, which are written as:

[0161]

[0162] Among them, Data_VT1 g represents the g-th PVDF sensor array, represents the output voltage of the jth sensor on the gth PVDF sensor array under the i-th test condition, fy i Represents the tire lateral force under the i-th test condition. The average value of the first-order main effect index (FMEI) of the output voltage values ​​of all sensors in the g-th PVDF sensor array is taken as the first-order main effect index of the g-th installation position, denoted as MFMEI, which can be specifically expressed as:

[0163]

[0164] in, Mean(Data_VT1(A) represents the first-order main effect index of the Δth (1≤Δ≤6, Δ=1 represents tire pressure, Δ=2 represents vertical load, Δ=3 represents vehicle speed, Δ=4 represents adhesion coefficient, Δ=5 represents sideslip angle, Δ=6 represents slip rate) factor on the voltage output of the gth sensor array. i ,Data_VT1(B) i ) represents the average value of the numerical vector formed by combining the i-th column of the sensor array output voltage matrix corresponding to matrix A and the i-th column of the sensor array output voltage matrix corresponding to matrix B, represents the jth value of the i-th column of the sensor array output voltage matrix corresponding to matrix B, Represents the matrix AB Δ The corresponding sensor array output voltage matrix has the jth value in the i-th column, Represents the jth value in the i-th column of the sensor array output voltage matrix corresponding to matrix A.

[0165] Furthermore, considering that the tire force cannot be directly controlled, the present invention achieves the purpose of indirectly controlling the tire lateral force by controlling the tire slip angle. Figure 5 The results of the top 10 MFMEIs in descending order when the side slip angle changes are shown. Considering the limitation of the installation space in the tire in practical application, the present invention selects the top 4 positions, namely 24, 37, 44 and 17, as the optimal installation positions of the PVDF sensor array. On this basis, the finite element model of the array PVDF smart tire is reconstructed to obtain the finite element model of the smart tire with the optimal configuration of the tire sensor.

[0166] After determining the installation positions of the four sensor arrays, the voltage data Data_VT1 is obtained. g Further extraction is performed, and 5 specific voltage values ​​of each sensor array voltage data are selected as input features; the specific selection logic is: taking the maximum voltage in the sensor array under each working condition as a reference, 4 voltage value points are selected along the direction of voltage decrease on both sides of the point, and finally 5 voltage value points are selected as input features. The re-selected data is recorded as Data_VT2 g , which can be written as:

[0167]

[0168] Further, extract the obtained data sets Data_Fy and Data_VT2 g 80% of the total data is used as the training set, denoted as Data_Fy x and Data_VT2 g_x , 20% as test set, denoted as Data_Fy c and Data_VT2g c On this basis, according to Data_Fy and Data_VT2 g The data characteristics of are combined with the empirical method to determine the mean function m(x) and covariance function k(x,x') types, written as:

[0169] m(x)=E[f(x)] (16)

[0170]

[0171] Where f(x) is the latent function, which represents Data_VT2 g_x enter The corresponding tire lateral force fy, E is the mathematical expectation, σ 2 is the signal variance, F i represents the variance scale of the i-th dimension, x i is the known data of the i-th dimension, x i ' is x i, where d is the dimension of the variable x.

[0172] Furthermore, the voltage data and tire lateral force of the training data set are expressed in the form of Gaussian process distribution, that is, the lateral force Data_Fy x The function f(Data_VT2 g_x ) is subject to the mean value m(Data_VT2 g_x ), covariance is k(Data_VT2 g_x ,Data_VT2 g_x '). On this basis, the initial values ​​of the hyperparameters and the number of iterations in the Gaussian process regression model are determined, and the hyperparameters σ and F in the covariance function are determined by minimizing the negative logarithm of the marginal likelihood function by the gradient descent method. The negative logarithm of the marginal likelihood function can be specifically expressed as:

[0173]

[0174] Where -logp(f|x) represents the negative logarithm of the marginal likelihood function, f is the Gaussian process regression function, and x is the voltage input feature. For fitting the training data, n is the dimension of the training data set, k n is the n-dimensional covariance function, is the n-dimensional variance vector, I n is an n-dimensional unit vector, is a penalty item, is the normalization constant.

[0175] Furthermore, according to the prior definition, the joint prior distribution can be obtained by combining the collected training set and test set:

[0176]

[0177] Where, f(Data_VT2 g_x ) is the tire lateral force corresponding to the training set, f(Data_VT2 g_c ) is the tire lateral force corresponding to the test number, and N is a normal distribution.

[0178] Further, the estimated value f(Data_VT2 g_c ), written as:

[0179]

[0180] in, represents the estimated mean. Substituting the solved hyperparameters into the real-time tire lateral force f(Data_VT2g_c) can be obtained. g_x)) is the variance corresponding to f(Data_VT2g_x).

[0181] S3, based on the real-time tire lateral force obtained in S2, combined with the vehicle-mounted sensor information, design a tire cornering stiffness adaptive algorithm to obtain the real-time cornering stiffness adaptive value;

[0182] Get the front and rear wheelbase l of the vehicle f (1.015 m in this embodiment) and l r (1.895m in this embodiment), calculate the vertical load of the four wheels when the vehicle is stationary, which can be specifically written as:

[0183]

[0184] Among them, Ld fl , Ld fr , Ld rl and Ld rr are the static vertical loads of the left front, right front, left rear and right rear tires of the vehicle, m represents the total mass of the vehicle body (1412 kg in this embodiment), g represents the acceleration of gravity (9.81 m / s in this embodiment), 2 );l f is the distance from the front axle to the center of mass of the vehicle, l r is the distance from the rear axle to the center of mass of the vehicle.

[0185] Furthermore, the cornering characteristic tests were carried out under the conditions of tire loads of 4510.1N and 2415.7N to obtain the cornering angle-lateral force data, and the fixed values ​​of the front and rear wheel cornering stiffness K were obtained by combining the magic tire formula for lateral force and the curve fitting method. f and K r The Magic Tire formula for lateral force used is:

[0186] Y=Dsin(Carctan(Bx-E(Bx-arctan(Bx)))) (22)

[0187] Among them, Y is the tire lateral force output item, x is the tire slip angle input item, and B, C, D and E are the parameters to be fitted.

[0188] Furthermore, the fixed values ​​of the front and rear tire cornering stiffness can be expressed as:

[0189]

[0190] Among them, B f , C f and D f is the load Ld fl (or Ld fr ) corresponding to the side deviation characteristic data fitting parameters, B r , Cr and D r is the load Ld rl (or Ld rr ) is the parameter obtained by fitting the corresponding side slip characteristic data. It is defined that the two front wheel steering angles are consistent and equal to δ. On this basis, the vehicle sensor signal is obtained to calculate δ, which can be expressed as:

[0191]

[0192] Among them, δ sw represents the steering wheel angle, i w represents the angular transmission ratio of the vehicle steering system, which is 17.6 in this embodiment. Define the two front axle side slip angles to be the same and equal to α f , the side slip angles of the two wheels on the rear axle are the same and equal to α r At this time, the side slip angles of the front and rear wheels can be expressed as:

[0193]

[0194] Among them, β, γ and v x are the vehicle's center of mass slip angle, yaw rate and longitudinal velocity, respectively, which can be obtained by the vehicle's key driving state estimator. The tire's cornering stiffness is adaptively adjusted using the difference between the linear tire force and the intelligent tire lateral force estimation value. At this time, the adaptive values ​​of the front and rear tire cornering stiffness can be expressed as:

[0195]

[0196] in, and are the adaptive values ​​of the cornering stiffness of the front and rear wheels, and are the sum of the estimated values ​​of the lateral forces of the two front wheels and the two rear wheels, respectively, which can be obtained by the established tire lateral force estimation model, F yf and F yr It is the sum of the linear tire forces of the two front wheels and the two rear wheels respectively, which can be obtained by multiplying the fixed value of the cornering stiffness and the tire slip angle.

[0197] S4. Acquire vehicle parameters and real-time sensor signals, combine with real-time cornering stiffness adaptive values, and establish a vehicle dynamics model.

[0198] Obtain vehicle parameters, including vehicle mass (1412 kg in this embodiment), vehicle yaw moment of inertia (1536.7 kg / m in this embodiment) 2 ), the distance from the center of mass of the vehicle to the front axle (1.015m in this embodiment) and the distance to the rear axle (1.895m in this embodiment); obtaining real-time signals from on-board sensors, including the longitudinal acceleration and lateral acceleration of the vehicle, yaw angular velocity and steering wheel angle;

[0199] Establish as Figure 6 The three-degree-of-freedom vehicle dynamics model is shown, and the following assumptions are made:

[0200] (1) Ignore the aerodynamics of the vehicle;

[0201] (2) Ignore the roll, pitch and vertical motion of the vehicle and only consider the longitudinal, lateral and yaw motions;

[0202] (3) The road surface slope angle and inclination angle are zero;

[0203] (4) The center of mass of the vehicle is fixed and coincides with the origin of the vehicle coordinate system;

[0204] (5) The front wheel angle of the car is linearly related to the steering wheel angle, and the rear wheel angle is fixed;

[0205] Furthermore, the dynamic equations of the vehicle in the longitudinal, lateral and yaw directions can be obtained:

[0206]

[0207] Among them, a x is the longitudinal acceleration of the vehicle, I z is the vehicle yaw moment of inertia, is the longitudinal acceleration, .

[0208] S5. Establish a vehicle key driving state estimation algorithm based on the vehicle dynamics model and the new information adaptive extended Kalman filter to obtain the real-time vehicle longitudinal speed, center of mass sideslip angle and yaw angular velocity.

[0209] First, the state vector J, input vector U and observation vector Z of the vehicle key driving state estimation system are defined as follows:

[0210] J=(v x , β, γ) T =(J1, J2, J3) T (28)

[0211] U=(a x , δ) T =(U1, U2) T (29)

[0212] Z=(a y , v x , γ) T =(Z1, Z2, Z3) T (30)

[0213] Among them, a x is the longitudinal acceleration of the vehicle, a yis the lateral acceleration of the vehicle, which can be obtained by the on-board sensor. Similarly, the yaw rate γ can also be easily obtained by the on-board sensor. The longitudinal velocity v x It can be obtained by measuring with a low-orbit satellite. The vehicle's center of mass sideslip angle β is in a state to be estimated, and δ is the front wheel turning angle.

[0214] Then, the state prior estimation equation P(·) and the observation estimation equation H(·) are defined as follows:

[0215]

[0216] Among them, P1, P2 and P3 represent the prior estimates of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively, and v x0 , β0 and γ0 represent the initial values ​​of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively, and are all set to 0 in this embodiment. s is the system sampling time, which is set to 0.001s in this embodiment. is the adaptive value of the front wheel cornering stiffness, is the adaptive value of the rear wheel cornering stiffness, and m is the vehicle mass.

[0217] Furthermore, a noise covariance matrix adaptive adjustment strategy based on sliding window self-adjustment is established to obtain the noise covariance adaptive matrix. First, the covariance matrix of the new information is defined as:

[0218] C k =E[v k v k T ] (33)

[0219] Among them, v k is the error between the actual value and the predicted value of the measured variable in the kth step (k>0), and E is the mathematical expectation.

[0220] Furthermore, the actual covariance matrix of the new information is defined as:

[0221]

[0222] Where M represents the length of the sliding window, which can be further expressed as:

[0223]

[0224] Among them, int(·) is the rounding function, is a constant, which is 10 in this embodiment; at this time, the noise covariance adaptive matrix The expression can be written as:

[0225]

[0226] Wherein, trace(·) is the trace function, and R is the noise covariance matrix, which is:

[0227]

[0228] Furthermore, based on the extended Kalman filter principle and the first-order Taylor series expansion theorem, the calculation process parameters, including the Jacobian matrix Pm and Hm, as well as the prior estimation error covariance matrix Pf and the Kalman gain Km, are calculated, and their expressions are:

[0229]

[0230] Pf k =Pm k ·Pl k-1 ·Pm k T +Q (40)

[0231]

[0232] Among them, Pm k is the Jacobian matrix of the k-th step state transfer equation, Hm k is the Jacobian matrix of the k-th step measurement equation, Pl k-1 is the error covariance matrix of the k-1th step, Pl0 refers to the first step, that is, the value of pl when k=1, that is, the initial value of pl, which is obtained by manual parameter adjustment by trial and error, and Q is the process noise covariance matrix, the parameters of which can be manually calibrated by trial and error. In this embodiment, they are:

[0233]

[0234] Furthermore, based on the state prior estimation information, the measurement estimation information and the sensor measurement information, the key driving state of the vehicle is estimated, and the error covariance matrix Pl is updated. Their formulas can be specifically expressed as follows:

[0235]

[0236] Pl k =(I3-Km k ·Hm k )·Pl k (45)

[0237] Where I is the identity matrix, J k is the estimated value of the key driving state of the vehicle at the kth step, including the vehicle longitudinal velocity, center of mass sideslip angle and yaw angular velocity.

[0238] Figures 7 to 10The key driving state estimation method of the vehicle proposed in the present invention is tested at 80km / h on a high adhesion (adhesion coefficient of 0.8) road surface under double lane change conditions, where C0 represents the measured values ​​of the center of mass slip angle and yaw rate, and C1 represents the measured values ​​of the longitudinal vehicle speed provided by the low-orbit satellite, the voltage collected by the tire PVDF array sensor, the center of mass slip angle and yaw rate. The results show that the estimation accuracy of the lateral force of the left front wheel is relatively high, and the estimated absolute error is within ±40N. The maximum absolute errors of the estimated values ​​of the longitudinal speed, center of mass slip angle and yaw rate of the vehicle under the C0 condition are 0.0265m / s, 0.004rad and 0.0235rad / s respectively. The maximum absolute errors of the estimated values ​​of the longitudinal speed, center of mass slip angle and yaw rate of the vehicle under the C1 condition are 0.0011m / s, 0.0017rad and 0.0006rad / s respectively. Compared with only the vehicle's yaw rate and lateral acceleration as measurements, with the addition of low-orbit satellite and in-tire sensor information, the maximum absolute errors of vehicle longitudinal velocity, center of mass sideslip angle and yaw rate estimation decreased by 95.84%, 57.54% and 97.45%, respectively.

[0239] Figures 11 to 14 The vehicle key driving state estimation method proposed in the present invention is tested at 50km / h on a high adhesion (adhesion coefficient of 0.3) road surface under double lane change conditions. The results show that the left front wheel lateral force estimation curve has a very high degree of overlap with the true curve, and the estimated absolute error is within ±24N. The maximum absolute errors of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate estimation values ​​under the C0 condition are 0.0178m / s, 0.0039rad and 0.0114rad / s, respectively. The maximum absolute errors of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate estimation values ​​under the C1 condition are 0.004m / s, 0.0029rad and 0.0022rad / s, respectively. Compared with only the vehicle yaw rate and lateral acceleration as the measurement quantity, with the addition of low-orbit satellite and tire in-tire sensor information, the maximum absolute errors of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate estimation are reduced by 77.52%, 25.64% and 80.7%, respectively. In summary, compared with the vehicle driving state estimation method that uses traditional on-board sensor information as the measurement quantity, the vehicle key driving state estimation method based on the fusion of multi-source information of low-orbit satellite and on-board sensors proposed in the present invention shows a better estimation performance.

[0240] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A vehicle key driving state estimation method based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors, characterized in that: The following steps are involved: S1. Based on the low-orbit satellite measurement signal and the Doppler frequency shift formula, the real-time longitudinal velocity of the vehicle is obtained; S2, based on the intelligent tire measurement signal, combined with the Gaussian process regression algorithm, obtain the real-time tire lateral force; S3. Based on the real-time lateral force of the tire and combined with the vehicle-mounted sensor information, a tire cornering stiffness adaptive algorithm is designed to obtain a real-time cornering stiffness adaptive value; S4, acquiring vehicle parameters and real-time sensor signals, and establishing a vehicle dynamics model in combination with the real-time cornering stiffness adaptive value; S5. Establish a vehicle key driving state estimation algorithm based on the vehicle dynamics model and the new information adaptive extended Kalman filter to obtain the real-time vehicle longitudinal speed, center of mass sideslip angle and yaw angular velocity.

2. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: S1 includes the following steps: S1-1, based on the low-orbit satellite ephemeris data received by the vehicle-mounted receiver, calculate the position of each satellite in the geodetic coordinate system; S1-2, based on the position of each low-orbit satellite, determine the visible range of each low-orbit satellite to obtain the intersection of the visible range, and divide the grid to select candidate initial value points; S1-3. Use the relative motion between the vehicle and the satellite to establish the Doppler frequency shift equation, linearize the Doppler frequency shift equation, and substitute the candidate initial value points; use the least squares method to solve the real-time position of the vehicle and calculate the real-time longitudinal velocity of the vehicle.

3. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: The constructed Doppler frequency shift equation is expressed as: Where Δf i is the ith frequency shift observation value, f0 is the satellite’s signal transmission frequency, P u represents the desired position of the vehicle, P s is the satellite position vector, v u represents the desired speed of the vehicle, v s represents the speed of the satellite, c represents the signal propagation speed, τ i represents the frequency shift measurement error; Linearize the Doppler frequency shift equation and substitute the candidate initial value points into it. The Doppler observation equation is approximated as: Where x is the variable to be determined; Δx is the vehicle position and speed correction to be determined, denoted as Δx=(Δx u ,Δy u ,Δz u ,Δv x , u ,Δv y,u ,Δv z,u ), Δx u , Δy u , Δz u , Δv x,u , Δv y,u , Δv z,u They are respectively the vehicle longitudinal position correction, lateral position correction, vertical position correction, longitudinal speed correction, lateral speed correction and vertical speed correction; Represents the Jacobian matrix of the observation equation for the variable to be determined; v u (0) is the initial velocity vector of the ground vehicle.

4. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: Taking the subsatellite point of each low-orbit satellite as the center, calculate its visible radius r on the ground v : The visible projection range of each low-orbit satellite in the longitude and latitude directions is: Take the visible range of all low-orbit satellites and calculate their intersection area, which is represented as a rectangle: Divide the intersection matrix into N×N grids and use the center point of each grid as a candidate initial value point: Where R is the radius of the Earth, h s is the altitude of satellite s, L lon is the vertical boundary of the satellite visible projection range, L lat is the lateral boundary of the satellite visible projection range, ΔΦ is the span of the rectangular area in the latitude direction, Φ min is the southernmost boundary of the intersection area, that is, the lowest latitude value within the rectangular range, Φ max is the northernmost boundary of the intersection area, that is, the highest latitude value within the rectangular range, Δλ is the span of the rectangular area in the longitude direction, and λ min is the westernmost boundary of the intersection area, that is, the lowest longitude value within the rectangular range, λ max is the easternmost boundary of the intersection area, that is, the highest longitude value within the rectangular range, Φ i is the latitude value of the i-th row grid, N is the number of grid divisions, λ j is the longitude value of the j-th column grid.

5. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: S2 includes the following steps: S2-1. Establish a PVDF smart tire finite element model, verify the accuracy of the modeling method through real vehicle tests, and build an array PVDF smart tire finite element model on this basis; S2-2, determine the influencing factors and value range of the output voltage of the smart tire PVDF sensor array according to daily driving conditions, and generate a sample matrix S in combination with the Sobol sequence method; The sample matrix S is decomposed into 2+D sub-sample matrices, where D represents the number of factors affecting the output voltage of the array piezoelectric film sensor; S2-3, quantify the z installation positions of the circumferential sensor array on the inner liner layer using the tire transverse pattern, and install z PVDF sensor arrays to form a new intelligent tire finite element model. On this basis, use the new intelligent tire finite element model to calculate 2+D sub-sample matrices, and extract the output voltages of the z sensor arrays and their corresponding tire lateral forces to form a voltage data set Data_VT1 and a lateral force data set Data_Fy; S2-4. The average value of the first-order main effect index of the output voltage values ​​of all sensors in the g-th PVDF sensor array is taken as the first-order main effect index of the g-th installation position, denoted as Arrange the g first-order main effect indicators in descending order, and select the first t installation positions as the optimal installation positions of the PVDF sensor array; S2-5, based on the determined t sensor array installation positions, the g-th PVDF sensor array data Data_VT1 g Further extraction is performed, and l specific voltage values ​​of each sensor array voltage data are selected as input features; the selection logic is: taking the maximum voltage in the sensor array under each working condition as a reference, c voltage value points are selected along the direction of voltage decrease on both sides of the point, and finally l voltage value points are selected as input features; the processed data is recorded as Data_VT2 g ; The extracted data sets Data_Fy and Data_VT2 g Divided into training set and test set, the training set is denoted as Data_Fy x and Data_VT2 g_x , the test set is denoted as Data_Fy c and Data_VT2g c ; S2-6, according to Data_Fy and Data_VT2 g Based on the data characteristics, the mean function m(x) and covariance function k(x,x') types are determined by empirical methods. The voltage data and tire lateral force of the training data set are expressed in the form of Gaussian process distribution, that is, the lateral force Data_Fy x The function f(Data_VT2 g_x ) is subject to the mean value m(Data_VT2 g_x ), covariance is k(Data_VT2 g_x ,Data_VT2 g_x '); on this basis, the initial values ​​of the hyperparameters and the number of iterations in the Gaussian process regression model are determined, and the hyperparameters σ and F in the covariance function are determined by minimizing the negative logarithm of the marginal likelihood function by the gradient descent method. The negative logarithm of the marginal likelihood function is specifically expressed as: Where -logp(f|x) represents the negative logarithm of the marginal likelihood function, f is the Gaussian process regression function, and x is the voltage input feature. For fitting the training data, n is the dimension of the training data set, k n is the n-dimensional covariance function, is the n-dimensional variance vector, I n is an n-dimensional unit vector, is a penalty item, is the normalization constant; According to the prior definition, the joint prior distribution can be obtained by combining the training set and test set collected in S2-1: Where, f(Data_VT2 g_x ) is the tire lateral force corresponding to the training set, f(Data_VT2 g_c ) is the tire lateral force corresponding to the test number, and N is a normal distribution; S2-7, calculate the estimated value f(Data_VT2 g_c ) and substitute it into S2-6 to obtain the hyperparameter solution for the real-time lateral force of the tire.

6. The method for estimating critical driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 5 is characterized in that: The expression of the first-order main effect index is: in, Represents the first-order main effect index of the Δth influencing factor on the voltage output of the gth sensor array, 1≤Δ≤D, Δ=5 represents the influencing factor is the tire side slip angle, D is the number of influencing factors; Mean(Data_VT1(A) i ,Data_VT1(B) i ) represents the average value of the numerical vector formed by combining the i-th column of the sensor array output voltage matrix corresponding to matrix A and the i-th column of the sensor array output voltage matrix corresponding to matrix B, represents the jth value of the i-th column of the sensor array output voltage matrix corresponding to matrix B, Represents the matrix AB Δ The corresponding sensor array output voltage matrix has the jth value in the i-th column, Represents the jth value in the i-th column of the sensor array output voltage matrix corresponding to matrix A.

7. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: Step S3 includes the following steps: S3-1. Obtain the front and rear wheelbase l of the vehicle f and l r , calculate the vertical loads on the four wheels when the vehicle is stationary, including the static vertical loads Ld of the left front, right front, left rear and right rear tires fl , Ld fr , Ld rl and Ld rr ; The tire load is Ld fl and Ld rl The cornering characteristic test under the conditions is used to obtain the cornering angle-side force data, and the fixed value K of the front and rear wheel cornering stiffness is obtained by combining the magic tire formula of the side force and the curve fitting method. f and K r ; S3-2, define the two front wheel steering angles to be the same and equal to δ, on this basis, obtain the vehicle sensor signal to calculate δ; define the two front axle wheels’ side slip angles to be the same and equal to α f , the side slip angles of the two wheels on the rear axle are the same and equal to α r ; S3-3, using the difference between the linear tire force and the intelligent tire lateral force estimation value to adaptively adjust the tire cornering stiffness; the adaptive values ​​of the front and rear tire cornering stiffness can be expressed as: in, and are the adaptive values ​​of the cornering stiffness of the front and rear wheels, and are the sum of the estimated lateral forces of the two front wheels and the two rear wheels respectively; F yf and F yr are the sum of the linear tire forces of the two front wheels and the two rear wheels respectively.

8. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 7 is characterized in that: The fixed values ​​of the front and rear tire cornering stiffness can be expressed as: Among them, B f , C f and D f is the load Ld fl or Ld fr The corresponding parameters obtained by fitting the side deviation characteristic data, B r , C r and D r is the load Ld rl or Ld rr Parameters obtained by fitting the corresponding cornering characteristic data.

9. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1 is characterized in that: S4 specifically includes the following steps: S4-1, obtain vehicle parameters, including vehicle mass, vehicle yaw moment of inertia, distance from vehicle center of mass to front axle and distance to rear axle; obtain real-time signals from vehicle sensors, including vehicle longitudinal acceleration and lateral acceleration, yaw angular velocity and steering wheel angle; S4-2, construct a vehicle dynamics model based on the real-time adaptive values ​​of the front and rear wheel cornering stiffness, vehicle parameters and real-time signals of the on-board sensors, which is specifically expressed as follows: Among them, a x is the longitudinal acceleration of the vehicle, I z is the vehicle yaw moment of inertia, is the longitudinal acceleration, v x is the longitudinal velocity of the vehicle, γ is the yaw rate, is the yaw angular acceleration, β is the sideslip angle of the vehicle center of mass, m is the vehicle mass, and are the adaptive values ​​of the cornering stiffness of the front and rear wheels, respectively, f and l r are the front and rear wheelbases of the vehicle, I z is the moment of inertia of the vehicle around the z-axis.

10. The method for estimating key driving states of a vehicle based on multi-source information fusion of low-orbit satellite and vehicle-mounted sensors according to claim 1, characterized in that: Step S5 includes the following steps: S5-1. Define the state vector J, input vector U and observation vector Z of the vehicle key driving state estimation system, which are expressed as follows: J=(v x ,b,c) T =(J1,J2,J3) T U=(a x ,d) T =(U1,U2) T Z=(a y ,v x ,γ) T =(Z1,Z2,Z3) T Among them, a x is the longitudinal acceleration of the vehicle, a y is the lateral acceleration of the vehicle, γ is the yaw rate, v x is the longitudinal velocity of the vehicle, β is the sideslip angle of the vehicle center of mass, and γ is the yaw rate; S5-2. Define the state prior estimation equation P(·) and the observation estimation equation H(·), which are expressed as follows: Among them, P1, P2 and P3 represent the prior estimates of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively, and v x0 , β0 and γ0 represent the initial values ​​of the vehicle longitudinal velocity, center of mass sideslip angle and yaw rate, respectively. s is the system sampling time; and are the sum of the estimated values ​​of the lateral forces of the two front wheels and the two rear wheels respectively; S5-3. Establish a noise covariance matrix adaptive adjustment strategy based on sliding window self-adjustment to obtain the noise covariance adaptive matrix. First, define the covariance matrix of the new information as: C k =E[v k v k T ] Among them, v k is the error between the actual value and the predicted value of the measured variable in the kth step, k>0; E is the mathematical expectation; The actual covariance matrix of the innovation is defined as: Among them, M represents the length of the sliding window, expressed as: Among them, int(·) is the rounding function, is a constant; Noise Covariance Adaptive Matrix The expression is written as: Where, trace(·) is the trace function, and R is the noise covariance matrix; S5-4, based on the extended Kalman filter principle and the first-order Taylor series expansion theorem, the calculation process parameters, including the Jacobian matrix Pm and Hm, and the prior estimation error covariance matrix Pf and the Kalman gain Km, are expressed as: Pf k =Pm k ·Pl k-1 ·Pm k T +Q Among them, Pm k is the Jacobian matrix of the k-th step state transfer equation, Hm k is the Jacobian matrix of the k-th step measurement equation, Pl k-1 is the error covariance matrix of the k-1th step, Q is the process noise covariance matrix, and its parameters can be manually calibrated by trial and error; S5-5, based on the state prior estimation information, the measurement estimation information and the sensor measurement information, the key driving state of the vehicle is estimated, and the error covariance matrix Pl is updated. The formula is specifically expressed as follows: Pl k =(I3-Km k ·Hm k )·Pl k Where I is the identity matrix, J k Km is the estimated value of the key driving state of the vehicle in the kth step, including the vehicle longitudinal speed, center of mass sideslip angle and yaw rate; k is the Kalman gain at the kth step.

Images