Method, system and device for estimating motion state of distributed drive electric vehicle

Abstract

The application discloses a distributed drive electric vehicle motion state estimation method, system and device, and relates to the intelligent automobile active safety technical field.The application comprises the following steps: receiving the steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force and tire lateral force parameter information of a vehicle; and constructing a nonlinear vehicle dynamics model and a magic formula tire model.The application designs a vehicle state estimator containing random first-order measurement time delay and data packet loss, adopts independent random variables satisfying Bernoulli distribution to represent the randomness of the measurement time delay and the data packet loss, and compared with a traditional estimation algorithm, the method does not need to assume that the time delay is an integer multiple of the sampling interval, and does not need to assume that the sensor data only exists in two cases of complete normal transmission and complete failure, thereby effectively enhancing the robustness of the centroid side slip angle estimation under non-ideal conditions such as the measurement time delay and the data packet loss.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of active safety technology of intelligent vehicles, in particular to a motion state estimation method, system and device for a distributed drive electric vehicle. BACKGROUND

[0002] As the best carrier for realizing safe and efficient unmanned driving, the distributed drive electric vehicle has become the mainstream trend of future electric vehicle development. Compared with the existing chassis structure, the distributed drive electric vehicle directly installs the driving motor in or near the driving wheel, saves the complex transmission system, releases more motion degrees of freedom of the vehicle chassis, and can greatly improve the driving safety and control flexibility of the vehicle. Therefore, the accurate acquisition of the key motion state of the vehicle (such as the center of mass side slip angle and yaw angular velocity) as the premise of robust chassis control is crucial for the electric vehicle to realize high safety and high maneuverability.

[0003] At present, Kalman filter and its derivative algorithms have become effective methods for vehicle state estimation, mainly including Kalman filter, extended Kalman filter, unscented Kalman filter, etc. The traditional Kalman filter algorithm has high estimation accuracy under non-stationary random signals, but it is only suitable for linear systems. Considering that the established vehicle dynamics model is a high-dimensional nonlinear system, the traditional Kalman filter algorithm cannot effectively estimate the vehicle state. The extended Kalman filter improves the state estimation accuracy by Taylor series expansion of the nonlinear function, but this algorithm has high-order truncation error, and the Jacobian matrix can only be solved in the case of continuous differentiability of the function. The stability under high-dimensional strong nonlinear systems cannot be guaranteed. Compared with Kalman filter and extended Kalman filter, unscented Kalman filter has higher estimation accuracy under high-dimensional strong nonlinear systems. However, unscented Kalman filter can only achieve high-precision estimation of vehicle motion state under ideal conditions. In addition, due to the physical limitations of vehicle sensors and the congestion of communication networks, the measurement data obtained based on vehicle sensors is not completely correct, which can easily lead to a decrease in estimation accuracy or even filter divergence. SUMMARY

[0004] The purpose of the present application is to provide a motion state estimation method, system and device for a distributed drive electric vehicle, which can effectively solve the problem of decreased estimation accuracy caused by random time-varying time delay and data packet loss.

[0005] According to the first aspect of the present application, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a motion state estimation method for a distributed drive electric vehicle, comprising the following steps:

[0006] receiving the steering wheel angle, lateral acceleration, longitudinal acceleration, yaw angular velocity, tire longitudinal force and tire lateral force parameter information of the vehicle;

[0007] A nonlinear vehicle dynamics model and a magic formula tire model are constructed, and a nonlinear discrete state-space equation for estimating vehicle motion state is constructed based on the nonlinear vehicle dynamics model and the magic formula tire model.

[0008] Considering the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation, independent random variables satisfying Bernoulli distribution are used to describe the data measurement delay and packet loss phenomenon, and a discrete state-space model containing random first-order measurement delay and data packet loss is constructed.

[0009] Based on the unscented Kalman filter as the basic framework, and based on the orthogonal projection theory, the gain matrix and the covariance matrix of the estimation error are updated in real time to construct a centroid side deflection estimator for high-dimensional strongly nonlinear systems that considers sensor measurement time delay and data packet loss.

[0010] The acquired vehicle status information is input into the centroid sideslip angle estimator established above, which takes into account random first-order measurement delay and data packet loss, to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions.

[0011] Furthermore, the vehicle's steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters are collected by an angle sensor, a gyroscope, and a tire six-component force tester.

[0012] Furthermore, the nonlinear discrete state-space equations are constructed, and the specific steps are as follows:

[0013] A nonlinear seven-degree-of-freedom dynamic model is established to describe the dynamic characteristics of the vehicle. The dynamic equations are expressed as follows:

[0014]

[0015]

[0016] a y =[sinδ(F) x,fl +F x,fr )+cosδ(F y,fl +F y,fr )+F y,rl +F y,rr ] / m

[0017] In the formula, the subscripts fl, fr, rl, and rr represent the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively; v m β is the center of mass velocity, γ is the sideslip angle, γ is the yaw rate, δ is the steering wheel angle, m is the vehicle mass, and I is the center of mass velocity. z Let T be the moment of inertia about the vertical axis. f T is the front axle track. r The rear axle track, ay For lateral acceleration, F x,ij For the longitudinal force of the tire, F y,ij This refers to the lateral force of the tire.

[0018] Furthermore, a magic formula tire model is constructed to characterize the longitudinal and lateral forces of the tire. The specific steps are as follows:

[0019] F(x)=Dsin{Carctan[Earctan(Bx)+(1-E)Bx]}

[0020] In the formula, x is the tire slip angle or longitudinal slip ratio. When x is the tire slip angle α, ij When x is the slip ratio λ, F(x) characterizes the longitudinal force of the tire; when x is the slip ratio λ, F(x) represents the longitudinal force of the tire. ij In this context, F(x) represents the lateral force of the tire; B, C, D, and E are the fitting coefficients of the tire model, whose values ​​are jointly determined by the vertical load, tire camber angle, and road conditions. Among them, B is the stiffness factor, whose value determines the slope at the origin of the curve; C is the shape factor, whose value has a significant impact on the shape of the sine function; the peak value of the curve is mainly determined by the peak factor D; and E is the curvature factor, whose value determines the horizontal position and curvature of the peak point.

[0021] Tire slip angle and slip ratio are expressed as follows:

[0022]

[0023]

[0024] In the formula, w ij R is the angular velocity. ij Let be the tire radius, a be the distance from the center of mass to the front axle, and b be the distance from the center of mass to the rear axle.

[0025] Based on the tire force coupling relationship, the longitudinal force and lateral force of the tire can be transformed into the following form:

[0026]

[0027] In the formula λ ij For tire slip ratio, F x,ij For the longitudinal force of the tire, F y,ij For the lateral force of the tire, α ij This refers to the tire slip angle.

[0028] Furthermore, based on the nonlinear vehicle dynamics model and the magic formula tire model, a nonlinear discrete state-space model is constructed as follows:

[0029]

[0030] In the formula, x θ =[vm ,β] T Let z(θ) be the state vector, and z(θ) = [a y ] T Let u be the measurement vector. θ =[δ,γ,F xij ,F yij ] T f is the control vector. θ (.) represents the discrete nonlinear function of the state equation, Γ θ (.) represents the discrete nonlinear function of the measurement equation; w θ Let the covariance matrix be Q θ Process noise, v θ Let the covariance matrix be R θ Measurement noise;

[0031] w θ and v θ They are unrelated and satisfy the following relationship:

[0032]

[0033] In the formula, (.) T For the matrix transpose, E(.) denotes the expected value, Ω θ-j Let Kronecker function be used.

[0034] Furthermore, a nonlinear discrete state-space model incorporating random first-order measurement delays and data loss is constructed, as follows:

[0035] The discrete state equations and measurement equations for a strongly nonlinear system, considering the effects of measurement delay and data loss, are expressed as follows:

[0036]

[0037] In the formula, Ξ θ Let z(θ) be the actual measured value, and z(θ) be the ideal measured value without time delay and packet loss. Let λ be the first-order predicted value of the measurement vector. θ and ξ θ Let be independent random variables describing measurement delay and data loss, and satisfy the following Bernoulli distribution:

[0038] Prob(λ θ =1)=α; Prob(λ) θ =0)=1-α

[0039] Prob(ζ θ =1)=β; Prob(ζ) θ =0)=1-β

[0040] Where 0≤α≤1, 0≤β≤1, it should be noted that when λθ =1 and ξ θ When z = 0, data loss occurs in the system. In this case, the actual measurement value is the ideal measurement value z, which does not contain time delay or packet loss. θ Substitute; when λ θ =0 and ξ θ When z = 1, a measurement time lag occurs. At this time, the actual measurement value is determined by the measurement vector z from the previous time step. θ-1 Substitute; when λ θ =0 and ξ θ When the value is 0, measurement delay and data loss occur simultaneously. In this case, the actual measurement value is the first-order predicted value of the measurement vector. Update.

[0041] Furthermore, a centroid side slip angle estimator considering sensor measurement delay and data loss is constructed for a high-dimensional strongly nonlinear system. The specific steps are as follows:

[0042] (1) Initialization of the state vector and estimation error covariance matrix:

[0043]

[0044] (2) Based on prior state vector and the prior error covariance matrix P θ|θ The Sigma sampling points were obtained using symmetrical sampling as follows:

[0045]

[0046] In the formula, Here are the prior Sigma sampling points, and n is the dimension of the state vector. Let Φ be the i-th column of the square root matrix based on Cholesky decomposition; Φ = σ 2 (n+κ)-n is used to correct prediction bias; σ determines the Sigma point. The distribution range in the vicinity; κ = 3 - n;

[0047] Furthermore, the weights of the state vector mean and covariance matrix are expressed as follows:

[0048]

[0049] In the formula, The weights are the mean of the state vector. For the weights of the covariance matrix,

[0050] (3) Obtain the Sigma point by performing a nonlinear transformation based on the state equation:

[0051]

[0052]

[0053] In the formula Let f be the Sigma point after the nonlinear transformation of the state equation. θ (.) represents the discrete nonlinear function of the state equation;

[0054] The expressions for the first-order predicted values ​​of the state vector and the estimation error covariance matrix at time θ are:

[0055]

[0056] In the formula Q θ The process noise covariance matrix;

[0057] (4) Based on the first-order prediction value of the obtained state vector And the first-order predicted value P of the estimation error covariance matrix θ+1|θ The Sigma sampling points are then updated. The updated Sigma sampling points are represented as follows:

[0058]

[0059] In the formula For the updated Sigma sampling points;

[0060] (5) Substitute the updated Sigma sampling points into the observation equation Γ θ+1 (.), the expression for the expected observation at time θ is:

[0061]

[0062] The expression for the first-order predicted value of the observation vector at time θ is:

[0063]

[0064] Similarly, based on the state equation, a nonlinear transformation of the Sigma point is obtained...

[0065]

[0066] The first-order predicted values ​​of the state vector and the estimation error covariance matrix at time θ-1 are calculated as follows:

[0067]

[0068] (6) Based on the first-order prediction value of the obtained state vector And the first-order predicted value P of the estimation error covariance matrix θ|θ-1 The Sigma sampling points are reselected, and the expression for the reselected Sigma sampling points is as follows:

[0069]

[0070] In the formula These are the Sigma sampling points after reselection;

[0071] Substitute the newly selected Sigma sampling points into the observation equation Γ θ (.), the expression for the expected observation value at time θ-1 is obtained as follows:

[0072]

[0073] The first-order predicted value of the observation vector at time θ-1 is calculated as follows:

[0074]

[0075] The algorithm gain matrix is ​​updated as follows:

[0076]

[0077] The expressions for each variable are as follows:

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087] Therefore, the posterior state vector at time θ+1 can be expressed as:

[0088]

[0089] The estimated error covariance matrix at time θ+1 is updated as follows:

[0090]

[0091] According to a second aspect of the present invention, a distributed drive electric vehicle motion state estimation system is provided, comprising:

[0092] The receiving module is used to receive vehicle steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force and tire lateral force parameter information;

[0093] The first building module is used to construct a nonlinear vehicle dynamics model and a magic formula tire model, and to construct a nonlinear discrete state-space equation for estimating the vehicle's motion state based on the nonlinear vehicle dynamics model and the magic formula tire model.

[0094] The second building module is used to consider the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation. It uses independent random variables that satisfy Bernoulli distribution to describe the data measurement delay and packet loss phenomenon, and builds a discrete state-space model containing random first-order measurement delay and data packet loss.

[0095] The third module is used to construct a centroid side deflection estimator for high-dimensional strongly nonlinear systems that considers sensor measurement delay and data loss, based on the unscented Kalman filter as the basic framework and the orthogonal projection theory to dynamically update the gain matrix and the covariance matrix of the estimation error in real time.

[0096] The high-precision estimation module is used to input the acquired vehicle state information into the centroid sideslip angle estimator established above, which takes into account random first-order measurement delay and data packet loss, to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions.

[0097] According to a third aspect of the present invention, the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores the computer program capable of running on the processor, and when the processor loads and executes the computer program, it employs the above-described distributed drive electric vehicle motion state estimation method.

[0098] According to a fourth aspect of the present invention, the present invention provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the above-described distributed drive electric vehicle motion state estimation method.

[0099] This invention has at least the following beneficial effects:

[0100] This invention addresses the problem of poor vehicle state estimation accuracy caused by measurement delays and data packet loss in high-dimensional nonlinear systems. It designs a vehicle state estimator incorporating random first-order measurement delays and data packet loss, using independent random variables that satisfy a Bernoulli distribution to characterize the randomness of measurement delays and data packet loss. Compared with traditional estimation algorithms, the proposed method does not require the assumption that the delay is an integer multiple of the sampling interval, nor does it require the assumption that sensor data can only be transmitted normally or completely fail. This effectively enhances the robustness of centroid sideslip angle estimation under non-ideal conditions such as measurement delays and data packet loss.

[0101] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0102] Fig. 1 This is a flowchart illustrating the method described in this invention;

[0103] Fig. 2 This is a schematic diagram of the nonlinear vehicle dynamics model in this invention;

[0104] Fig. 3 This is a comparison chart of the centroid side slip angle estimation results of the algorithm proposed in this invention and the traditional unscented Kalman filter algorithm. Detailed Implementation

[0105] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0106] Please see Figs. 1-3 This invention provides a technical solution: a method for estimating the motion state of a distributed-drive electric vehicle, specifically including the following steps:

[0107] S1. Receives vehicle steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters;

[0108] It should be noted that the vehicle's steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters are collected by an angle sensor, gyroscope, and tire six-component force tester.

[0109] S2. Construct a nonlinear vehicle dynamics model and a magic formula tire model, and based on the nonlinear vehicle dynamics model and the magic formula tire model, construct the nonlinear discrete state-space equations for estimating the vehicle's motion state, as follows:

[0110] S2.1 Constructing the nonlinear discrete state-space equations, the specific steps are as follows:

[0111] A nonlinear seven-degree-of-freedom dynamic model is established to describe the dynamic characteristics of the vehicle. The dynamic equations are expressed as follows:

[0112]

[0113]

[0114] a y =[sinδ(F) x,fl +F x,fr )+cosδ(F y,fl +F y,fr )+F y,rl +F y,rr ] / m

[0115] In the formula, the subscripts fl, fr, rl, and rr represent the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively; v m β is the center of mass velocity, γ is the sideslip angle, γ is the yaw rate, δ is the steering wheel angle, m is the vehicle mass, and I is the center of mass velocity. z Let T be the moment of inertia about the vertical axis. f T is the front axle track. r The rear axle track, a y For lateral acceleration, F x,ij For the longitudinal force of the tire, F y,ij This refers to the lateral force of the tire;

[0116] S2.2 Constructing a Magic Formula tire model to characterize the longitudinal and lateral forces of the tire, the specific steps are as follows:

[0117] F(x)=Dsin{Carctan[Earctan(Bx)+(1-E)Bx]}

[0118] In the formula, x is the tire slip angle or longitudinal slip ratio. When x is the tire slip angle α, ij When x is the slip ratio λ, F(x) characterizes the longitudinal force of the tire; when x is the slip ratio λ, F(x) represents the longitudinal force of the tire. ij In this context, F(x) represents the lateral force of the tire; B, C, D, and E are the fitting coefficients of the tire model, whose values ​​are jointly determined by the vertical load, tire camber angle, and road conditions. Among them, B is the stiffness factor, whose value determines the slope at the origin of the curve; C is the shape factor, whose value has a significant impact on the shape of the sine function; the peak value of the curve is mainly determined by the peak factor D; and E is the curvature factor, whose value determines the horizontal position and curvature of the peak point.

[0119] Furthermore, the tire slip angle and slip ratio are expressed as follows:

[0120]

[0121]

[0122] In the formula, w ij R is the angular velocity. ij Let be the tire radius, a be the distance from the center of mass to the front axle, and b be the distance from the center of mass to the rear axle.

[0123] Based on the tire force coupling relationship, the longitudinal force and lateral force of the tire can be transformed into the following form:

[0124]

[0125] In the formula λ ij For tire slip ratio, F x,ij For the longitudinal force of the tire, F y,ij For the lateral force of the tire, α ij This refers to the tire slip angle;

[0126] S2.3 constructs a nonlinear discrete state-space model based on the nonlinear vehicle dynamics model and the magic formula tire model, as detailed below:

[0127]

[0128] In the formula, x θ =[v m ,β] T Let z(θ) be the state vector, and z(θ) = [a y ] T Let u be the measurement vector. θ =[δ,γ,F xij ,F yij ] T f is the control vector. θ (.) represents the discrete nonlinear function of the state equation, Γ θ (.) represents the discrete nonlinear function of the measurement equation; w θ Let the covariance matrix be Q θ Process noise, v θ Let the covariance matrix be R θ Measurement noise;

[0129] w θ and v θ They are unrelated and satisfy the following relationship:

[0130]

[0131] In the formula, (.) T For the matrix transpose, E(.) denotes the expected value, Ω θ-j The Kronecker function;

[0132] S3. Considering the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation, independent random variables satisfying Bernoulli distribution are used to describe the data measurement delay and packet loss phenomenon. A discrete state-space model containing random first-order measurement delay and data packet loss is constructed as follows:

[0133] A nonlinear discrete state-space model incorporating random first-order measurement delays and data loss is constructed as follows:

[0134] The discrete state equations and measurement equations for a strongly nonlinear system, considering the effects of measurement delay and data loss, are expressed as follows:

[0135]

[0136] In the formula, Ξ θ Let z(θ) be the actual measured value, and z(θ) be the ideal measured value without time delay and packet loss. Let λ be the first-order predicted value of the measurement vector. θ and ξ θ Let be independent random variables describing measurement delay and data loss, and satisfy the following Bernoulli distribution:

[0137] Prob(λ θ =1)=α; Prob(λ) θ =0)=1-α

[0138] Prob(ζ θ =1)=β; Prob(ζ) θ =0)=1-β

[0139] Where 0≤α≤1, 0≤β≤1, it should be noted that when λ θ =1 and ξ θ When z = 0, data loss occurs in the system. In this case, the actual measurement value is the ideal measurement value z, which does not contain time delay or packet loss. θ Substitute; when λ θ =0 and ξ θ When z = 1, a measurement time lag occurs. At this time, the actual measurement value is determined by the measurement vector z from the previous time step. θ-1 Substitute; when λ θ =0 and ξ θ When the value is 0, measurement delay and data loss occur simultaneously. In this case, the actual measurement value is the first-order predicted value of the measurement vector. The update shows that when neither measurement delay nor data loss occurs, the proposed algorithm is the unscented Kalman filter algorithm.

[0140] S4. Based on the unscented Kalman filter as the basic framework, and using orthogonal projection theory to dynamically update the gain matrix and the covariance matrix of the estimation error in real time, a centroid side slip angle estimator considering sensor measurement time delay and data packet loss for high-dimensional strongly nonlinear systems is constructed as follows:

[0141] (S4.1) Initialization of state vector and estimation error covariance matrix:

[0142]

[0143] In the formula P0 is the initial state vector estimate, P0 is the initial value of the estimation error covariance matrix, and E(.) is the expected value.

[0144] (S4.2) Based on prior state vector and the prior error covariance matrix P θ|θ The Sigma sampling points were obtained using symmetrical sampling as follows:

[0145]

[0146] In the formula, Here are the prior Sigma sampling points, and n is the dimension of the state vector. Let Φ be the i-th column of the square root matrix based on Cholesky decomposition; Φ = σ 2 (n+κ)-n is used to correct prediction bias; σ determines the Sigma point. The distribution range in the vicinity; κ = 3 - n;

[0147] Furthermore, the weights of the state vector mean and covariance matrix are expressed as follows:

[0148]

[0149] In the formula, The weights are the mean of the state vector. The weights are the covariance matrix weights. Since the system follows a Gaussian distribution, therefore...

[0150] (S4.3) Based on the state equation, the Sigma point is obtained by performing a nonlinear transformation:

[0151]

[0152] In the formula Let f be the Sigma point after the nonlinear transformation of the state equation. θ (.) represents the discrete nonlinear function of the state equation;

[0153] The expressions for the first-order predicted values ​​of the state vector and the estimation error covariance matrix at time θ are:

[0154]

[0155] In the formula Q θ The process noise covariance matrix;

[0156] (S4.4) First-order prediction based on the obtained state vector And the first-order predicted value P of the estimation error covariance matrix θ+1|θ The Sigma sampling points are then updated. The updated Sigma sampling points are represented as follows:

[0157]

[0158] In the formula For the updated Sigma sampling points;

[0159] (S4.5) Substitute the updated Sigma sampling points into the observation equation Γ θ+1 (.), the expression for the expected observation at time θ is:

[0160]

[0161] The expression for the first-order predicted value of the observation vector at time θ is:

[0162]

[0163] Similarly, based on the state equation, a nonlinear transformation of the Sigma point is obtained...

[0164]

[0165] The first-order predicted values ​​of the state vector and the estimation error covariance matrix at time θ-1 are calculated as follows:

[0166]

[0167] (S4.6) First-order prediction based on the obtained state vector And the first-order predicted value P of the estimation error covariance matrix θ|θ-1 The Sigma sampling points are reselected, and the expression for the reselected Sigma sampling points is as follows:

[0168]

[0169] Substitute the newly selected Sigma sampling points into the observation equation Γ θ (.), the expression for the expected observation value at time θ-1 is obtained as follows:

[0170]

[0171] The first-order predicted value of the observation vector at time θ-1 is calculated as follows:

[0172]

[0173] The algorithm gain matrix is ​​updated as follows:

[0174]

[0175] The expressions for each variable are as follows:

[0176]

[0177]

[0178]

[0179]

[0180]

[0181]

[0182]

[0183]

[0184]

[0185] Therefore, the posterior state vector at time θ+1 can be expressed as:

[0186]

[0187] The estimated error covariance matrix at time θ+1 is updated as follows:

[0188]

[0189] S5. Input the acquired vehicle status information into the centroid sideslip angle estimator established above, which takes into account random first-order measurement delay and data packet loss, to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions.

[0190] In summary, Fig. 3 This is a comparison of the centroid side-slip angle estimation results of the method proposed in this embodiment and the traditional unscented Kalman filter algorithm. It can be seen that the centroid side-slip angle estimation accuracy of the method proposed in this embodiment is better than that of the traditional unscented Kalman filter algorithm, and it has strong robustness to measurement delay and non-ideal conditions of data loss.

[0191] Example 2:

[0192] This invention provides a distributed drive electric vehicle motion state estimation system, comprising:

[0193] The receiving module is used to receive vehicle steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force and tire lateral force parameter information;

[0194] The first building module is used to construct a nonlinear vehicle dynamics model and a magic formula tire model, and to construct a nonlinear discrete state-space equation for estimating the vehicle's motion state based on the nonlinear vehicle dynamics model and the magic formula tire model.

[0195] The second building module is used to consider the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation. It uses independent random variables that satisfy Bernoulli distribution to describe the data measurement delay and packet loss phenomenon, and builds a discrete state-space model containing random first-order measurement delay and data packet loss.

[0196] The third module is used to construct a centroid side deflection estimator for high-dimensional strongly nonlinear systems that considers sensor measurement delay and data loss, based on the unscented Kalman filter as the basic framework and the orthogonal projection theory to dynamically update the gain matrix and the covariance matrix of the estimation error in real time.

[0197] The high-precision estimation module is used to input the acquired vehicle state information into the centroid sideslip angle estimator established above, which takes into account random first-order measurement delay and data packet loss, to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions.

[0198] Specifically, the aforementioned receiving module, first construction module, second construction module, third construction module, and high-precision estimation module can be embedded into a computer processing system. The computer, based on the distributed drive electric vehicle motion state estimation method provided above, calls the aforementioned modules to complete the task of equivalent modeling. The aforementioned receiving module, first construction module, second construction module, third construction module, and high-precision estimation module can perform operations according to the specific steps given by the distributed drive electric vehicle motion state estimation method.

[0199] It should be noted that the division of the various modules in the above system is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. These modules can be implemented entirely in software through processing element calls; they can be fully implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. For example, the receiving module can be a separate processing element or integrated into a chip in the aforementioned device. Alternatively, it can be stored as program code in the memory of the aforementioned device, and called and executed by a processing element of the device. The implementation of other modules is similar. Furthermore, these modules can be fully or partially integrated together or implemented independently. The processing element mentioned here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0200] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to form a system-on-a-chip (SOC).

[0201] Example 3:

[0202] The present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores the computer program capable of running on the processor. When the processor loads and executes the computer program, it employs the above-mentioned distributed drive electric vehicle motion state estimation method.

[0203] It should be noted that the terminal device can be a computer device such as a desktop computer, a laptop computer, or a cloud server, and the terminal device includes, but is not limited to, a processor and a memory. For example, the terminal device may also include input / output devices, network access devices, and buses.

[0204] Furthermore, the processor can be a central processing unit (CPU). Of course, depending on the actual use, other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. can also be used. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it in this regard.

[0205] Example 4:

[0206] The present invention provides a storage medium containing computer-executable instructions, characterized in that the computer-executable instructions, when executed by a computer processor, are used to perform the above-described method for estimating the motion state of a distributed drive electric vehicle.

[0207] The computer program can be stored in a computer-readable medium. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or certain middleware. The computer-readable medium includes any entity or device capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the computer-readable medium includes, but is not limited to, the above-mentioned components.

[0208] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0209] For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances. When an element is referred to as being "assembled on," "mounted on," "fixed to," or "set on" another element, it may be directly on the other element or there may be an intermediate element present. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be an intermediate element present. The terms "vertical," "horizontal," "upper," "lower," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible embodiments.

[0210] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0211] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. A method for estimating the motion state of a distributed-drive electric vehicle, characterized in that, Includes the following steps: Receives vehicle steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters; A nonlinear vehicle dynamics model and a magic formula tire model are constructed, and a nonlinear discrete state-space equation for estimating vehicle motion state is constructed based on the nonlinear vehicle dynamics model and the magic formula tire model. Considering the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation, independent random variables satisfying Bernoulli distribution are used to describe the data measurement delay and packet loss phenomenon, and a discrete state-space model containing random first-order measurement delay and data packet loss is constructed. Based on the unscented Kalman filter as the basic framework, and based on the orthogonal projection theory, the gain matrix and the covariance matrix of the estimation error are updated in real time to construct a centroid side deflection estimator for high-dimensional strongly nonlinear systems that considers sensor measurement time delay and data packet loss. The acquired vehicle status information is input into a centroid sideslip angle estimator that takes into account sensor measurement delay and data packet loss, so as to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions. A nonlinear discrete state-space model incorporating random first-order measurement delays and data loss is constructed as follows: The discrete state equations and measurement equations for a strongly nonlinear system, considering the effects of measurement delay and data loss, are expressed as follows: In the formula, These are actual measured values. This represents the ideal measurement value, free from time delay and packet loss. The first-order predicted value of the measurement vector. and Let be independent random variables describing measurement delay and data loss, and satisfy the following Bernoulli distribution: in, It should be pointed out that when and When data packet loss occurs in the system, the actual measurement value is replaced by the ideal measurement value that does not contain time delays or packet loss. Instead; when and At this time, a measurement time lag occurs, and the actual measurement value is changed from the measurement vector at the previous time step. Instead; when and At this time, measurement delay and data packet loss occur simultaneously. In this case, the actual measurement value is used as the first-order predicted value of the measurement vector. Update.

2. The method for estimating the motion state of a distributed-drive electric vehicle according to claim 1, characterized in that: The vehicle's steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters are collected by an angle sensor, a gyroscope, and a tire six-component force tester.

3. The method for estimating the motion state of a distributed-drive electric vehicle according to claim 2, characterized in that, The specific steps for constructing nonlinear discrete state-space equations are as follows: A nonlinear seven-degree-of-freedom dynamic model is established to describe the vehicle's dynamic characteristics. The dynamic equations are expressed as follows: In the formula, the subscript These represent the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively. For the velocity of the center of mass, The sideslip angle is the angle of the center of mass. The yaw rate is angular velocity. For steering wheel angle, For vehicle quality, Let the moment of inertia be about the vertical axis. The front axle track. The rear axle track. It is lateral acceleration. For the longitudinal force of the tire, This refers to the lateral force of the tire.

4. The method for estimating the motion state of a distributed-drive electric vehicle according to claim 3, characterized in that, The Magic Formula tire model is constructed to characterize the longitudinal and lateral forces of the tire. The specific steps are as follows: In the formula, The tire slip angle or longitudinal slip ratio, when Tire slip angle hour, Characterizing the longitudinal force of the tire; when slip ratio hour, Characterizes the lateral force of the tire; , , , The fitting coefficients for the tire model are determined by the vertical load, tire camber angle, and road conditions. This is the stiffness factor, and its value determines the slope at the origin of the curve. This is the shape factor, and its value has a significant impact on the shape of the sine function; The peak value of the curve is mainly composed of the peak factor. Decide; The curvature factor determines the horizontal position and curvature of the peak point; Tire slip angle and slip ratio are expressed as follows: In the formula, Angular velocity, For the tire radius, This is the distance from the center of gravity to the front axle. This is the distance from the center of mass to the rear axle; Based on the tire force coupling relationship, the tire longitudinal force and lateral force are transformed into the following forms: In the formula For tire slip ratio, For the longitudinal force of the tire, This refers to the lateral force of the tire. This refers to the tire slip angle.

5. The method for estimating the motion state of a distributed-drive electric vehicle according to claim 4, characterized in that, Based on the nonlinear vehicle dynamics model and the magic formula tire model, a nonlinear discrete state-space model is constructed as follows: In the formula, For state vectors, For measurement vectors, For control vectors, The state equation is a discrete nonlinear function. The measurement equation is a discrete nonlinear function; The covariance matrix is Process noise, The covariance matrix is Measurement noise; and They are unrelated and satisfy the following relationship: In the formula, For matrix transpose, Represents the mathematical expectation. Let Kronecker function be used.

6. The method for estimating the motion state of a distributed-drive electric vehicle according to claim 5, characterized in that, The specific steps for constructing a centroid side slip angle estimator for a high-dimensional strongly nonlinear system, considering sensor measurement delays and data loss, are as follows: (1) Initialization of state vector and estimation error covariance matrix: In the formula This is the estimated value of the initial state vector. To estimate the initial value of the error covariance matrix, This is the expected value; (2) Based on prior state vector and prior error covariance matrix The Sigma sampling points were obtained using symmetrical sampling as follows: In the formula, Here are the prior Sigma sampling points, and n is the dimension of the state vector. The square root matrix based on Cholesky decomposition is the first... List; Used to correct prediction bias; The decision of the Sigma point is in The distribution range in the vicinity; ; Furthermore, the weights of the state vector mean and covariance matrix are expressed as follows: In the formula, The weights are the mean of the state vector. For the weights of the covariance matrix, ; (3) Based on the state equation, the Sigma point is obtained by performing a nonlinear transformation: In the formula For the Sigma point after the nonlinear transformation of the state equation, The state equation is a discrete nonlinear function; The expressions for the first-order predicted values ​​of the time-state vector and the estimation error covariance matrix are: In the formula The process noise covariance matrix; (4) First-order prediction based on the obtained state vector And the first-order predicted value of the estimation error covariance matrix The Sigma sampling points are then updated. The updated Sigma sampling points are represented as follows: In the formula For the updated Sigma sampling points; (5) Substitute the updated Sigma sampling points into the observation equation. ,get The expression for the expected observation at time step is: The expression for the first-order prediction value of the observation vector at time step is: Similarly, based on the state equation, a nonlinear transformation of the Sigma point is obtained... The first-order predicted values ​​of the time-state vector and the estimation error covariance matrix are calculated as follows: (6) First-order prediction based on the obtained state vector And the first-order predicted value of the estimation error covariance matrix The Sigma sampling points are reselected, and the expression for the reselected Sigma sampling points is as follows: In the formula These are the Sigma sampling points after reselection; Substitute the newly selected Sigma sampling points into the observation equation. ,get The expression for the expected observation at time is: The first-order prediction value of the observation vector at time step is calculated as follows: The algorithm gain matrix is ​​updated as follows: The expressions for each variable are as follows: therefore The posterior state vector at time t can be represented as: The estimated error covariance matrix at time step 1 is updated as follows: 。 7. A distributed drive electric vehicle motion state estimation system, used to implement the distributed drive electric vehicle motion state estimation method according to any one of claims 1 to 6, characterized in that, include: The receiving module is used to receive information on vehicle steering wheel angle, lateral acceleration, longitudinal acceleration, yaw rate, tire longitudinal force, and tire lateral force parameters. The first building module is used to construct a nonlinear vehicle dynamics model and a magic formula tire model, and to construct a nonlinear discrete state-space equation for estimating the vehicle's motion state based on the nonlinear vehicle dynamics model and the magic formula tire model. The second building module is used to consider the impact of measurement vector delay and non-ideal data packet loss on vehicle state estimation. It uses independent random variables that satisfy Bernoulli distribution to describe the data measurement delay and packet loss phenomenon, and builds a discrete state-space model containing random first-order measurement delay and data packet loss. The third module is used to construct a centroid side deflection estimator for high-dimensional strongly nonlinear systems that considers sensor measurement delay and data loss, based on the unscented Kalman filter as the basic framework and the orthogonal projection theory to dynamically update the gain matrix and the covariance matrix of the estimation error in real time. The high-precision estimation module is used to input the acquired vehicle state information into the centroid sideslip angle estimator established above, which takes into account random first-order measurement delay and data packet loss, to achieve high-precision estimation of the vehicle centroid sideslip angle under non-ideal conditions.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, The memory stores a computer program that can run on a processor. When the processor loads and executes the computer program, it employs the distributed drive electric vehicle motion state estimation method as described in any one of claims 1 to 6.

9. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the distributed drive electric vehicle motion state estimation method as described in any one of claims 1 to 6.

Images