A centroid side slip angle estimation method, system, electronic device and storage medium

Abstract

This invention discloses a method, system, electronic device, and storage medium for estimating the center of gravity sideslip angle, relating to the field of key information perception technology for vehicle systems. The method includes: acquiring sensor data of a target vehicle; determining whether the inertial data satisfies signal constraints; if satisfied, then: performing tire sideslip stiffness calculation on the target vehicle based on the sensor data to obtain a first time-varying tire sideslip stiffness; determining the center of gravity sideslip angle of the target vehicle based on the first time-varying tire sideslip stiffness and the inertial data; if not satisfied, then: performing signal compensation based on steering wheel angle and wheel speed to obtain virtual inertial data; performing tire sideslip stiffness calculation on the target vehicle based on steering wheel angle, virtual inertial data, and wheel speed to obtain a second time-varying tire sideslip stiffness; determining the center of gravity sideslip angle of the target vehicle based on the second time-varying tire sideslip stiffness and the virtual inertial data. This invention can improve the estimation accuracy of the vehicle's center of gravity sideslip angle.

Description

Technical Field

[0001] This invention relates to the field of key information perception technology for vehicle systems, and in particular to a method, system, electronic device, and storage medium for estimating the centroid sideslip angle. Background Technology

[0002] Accurate vehicle motion status information is a necessary prerequisite for the efficient operation of vehicle active safety systems, among which the vehicle's center of gravity sideslip angle information is particularly important. However, direct measurement of the vehicle's center of gravity sideslip angle requires expensive dedicated sensors with stringent installation requirements, making it suitable only for scientific research. In comparison, soft measurement methods based on information from ordinary onboard sensors are more feasible. However, when using soft measurement methods to estimate the vehicle's center of gravity sideslip angle, the time-varying nature of tire side stiffness and the reliability of onboard sensor signals are unavoidable issues.

[0003] Distributed drive vehicles possess significant advantages such as compact structure, low center of gravity, and high transmission efficiency. They can achieve various dynamic control functions by independently controlling the drive / braking torque of the electric motors, attracting continuous attention from industry and academia. In existing technologies based on distributed drive vehicles, one approach is to fit the time-varying tire lateral stiffness based on the tire's vertical force; however, the tire's vertical force on mass-produced vehicles is difficult to obtain directly. Another approach, assuming precise observation of the tire's lateral force using an observer, can inversely deduce the tire lateral stiffness from the tire's lateral force, but this approach requires high precision in system modeling. Furthermore, the unreliability of signals from ordinary vehicle sensors, such as signal loss or abrupt changes, may lead to erroneous intervention by active safety systems, potentially causing safety issues. Therefore, there is still significant room for improvement in the estimation accuracy of the vehicle's center of gravity sideslip angle in existing technologies. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, electronic device, and storage medium for estimating the center of gravity sideslip angle of a vehicle, which can improve the estimation accuracy of the center of gravity sideslip angle.

[0005] To achieve the above objectives, the present invention provides a method for estimating the centroid sideslip angle, comprising:

[0006] Acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed;

[0007] Determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the preset signal threshold, and the vehicle driving state;

[0008] If satisfied, then:

[0009] The tire lateral stiffness of the target vehicle is calculated based on the sensor data to obtain the first time-varying tire lateral stiffness.

[0010] The center of gravity offset angle of the target vehicle is determined based on the first time-varying tire lateral stiffness and the inertial data;

[0011] If not satisfied, then:

[0012] Signal compensation is performed based on the steering wheel angle and wheel speed to obtain virtual inertial data;

[0013] The tire lateral stiffness of the target vehicle is calculated based on the steering wheel angle, the virtual inertia data, and the wheel speed to obtain the second time-varying tire lateral stiffness.

[0014] The center of gravity offset angle of the target vehicle is determined based on the second time-varying tire lateral stiffness and the virtual inertial data.

[0015] Optionally, before determining whether the inertial data satisfies the signal constraint conditions, the method further includes:

[0016] The inertial data limit value is determined based on the steering wheel angle change rate threshold; the absolute value of the inertial data limit value is greater than the preset signal threshold.

[0017] The signal constraint conditions are determined based on the inertial data limit value, the preset signal threshold, and the vehicle driving state.

[0018] The signal constraint condition is either a first constraint condition or a second constraint condition; the first constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, the absolute value of the inertial data is less than the preset threshold of the signal, and the vehicle's driving state is straight-line driving; the second constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, and the absolute value of the inertial data is greater than or equal to the preset threshold of the signal.

[0019] Optionally, the step of calculating the tire lateral stiffness of the target vehicle based on the sensor data to obtain a first time-varying tire lateral stiffness specifically includes:

[0020] Obtain historical tire lateral stiffness curves;

[0021] The historical tire lateral stiffness curve is segmented and affined into several intervals and corresponding tire lateral stiffness constants.

[0022] Calculate the target tire yaw angle based on the sensor data;

[0023] Fuzzy logic operations are performed on the target tire side angle to determine the target interval corresponding to the target tire side angle, and the tire side stiffness corresponding to the target interval is used as the first time-varying tire side stiffness.

[0024] Optionally, determining the center-of-gravity yaw angle of the target vehicle based on the first time-varying tire lateral stiffness and the inertial data specifically includes:

[0025] The first time-varying tire lateral stiffness is estimated by lateral angle state estimation to obtain the first predicted value;

[0026] The yaw angle state is estimated from the inertial data to obtain the first measurement value;

[0027] The centroid offset angle of the target vehicle is obtained by performing unscented Kalman filtering on the first predicted value and the first measured value.

[0028] Optionally, the step of performing signal compensation based on the steering wheel angle and the wheel speed to obtain virtual inertial data specifically includes:

[0029] Determine the signal compensation model;

[0030] The steering wheel angle and wheel speed input signal compensation model is used to perform signal compensation to obtain virtual inertial data.

[0031] Optionally, determining the signal compensation model specifically includes:

[0032] Acquire training data; the training data includes simulation parameters and virtual inertial data corresponding to the simulation parameters; the simulation parameters include: multiple simulated steering wheel angles and multiple simulated wheel speeds used for simulation experiments;

[0033] Construct a multi-layer neural network model;

[0034] The training data is input into the multilayer neural network model for training, and the trained multilayer neural network model is determined as the signal compensation model.

[0035] Optionally, the steering wheel angle change rate threshold is 0.2.

[0036] The present invention also provides a centroid sideslip angle estimation system, comprising:

[0037] A data acquisition unit is used to acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed.

[0038] The signal determination unit is used to determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the preset signal threshold, and the vehicle driving state.

[0039] The first centroid deflection angle estimation unit is used to determine if the inertial data satisfies the signal constraint conditions, then:

[0040] The tire lateral stiffness of the target vehicle is calculated based on the sensor data to obtain the first time-varying tire lateral stiffness.

[0041] The center of gravity offset angle of the target vehicle is determined based on the first time-varying tire lateral stiffness and the inertial data;

[0042] The second centroid deflection angle estimation unit is used to: if the inertial data does not satisfy the signal constraint conditions, then:

[0043] Signal compensation is performed based on the steering wheel angle and wheel speed to obtain virtual inertial data;

[0044] The tire lateral stiffness of the target vehicle is calculated based on the steering wheel angle, the virtual inertia data, and the wheel speed to obtain the second time-varying tire lateral stiffness.

[0045] The center of gravity offset angle of the target vehicle is determined based on the second time-varying tire lateral stiffness and the virtual inertial data.

[0046] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described centroid side deflection angle estimation method.

[0047] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the centroid sideslip angle estimation method as described above.

[0048] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0049] This invention discloses a method, system, electronic device, and storage medium for estimating the center of gravity sideslip angle. The method includes acquiring sensor data of a target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed; determining whether the inertial data satisfies signal constraint conditions; the signal constraint conditions are determined based on a steering wheel angle change rate threshold, a preset signal threshold, and the vehicle's driving state; if satisfied, then: performing tire sideslip stiffness calculation on the target vehicle based on the sensor data to obtain a first time-varying tire sideslip stiffness; determining the center of gravity sideslip angle of the target vehicle based on the first time-varying tire sideslip stiffness and the inertial data; if not satisfied, then: performing signal compensation based on the steering wheel angle and the wheel speed to obtain virtual inertial data; performing tire sideslip stiffness calculation on the target vehicle based on the steering wheel angle, the virtual inertial data, and the wheel speed to obtain a second time-varying tire sideslip stiffness; determining the center of gravity sideslip angle of the target vehicle based on the second time-varying tire sideslip stiffness and the virtual inertial data. This invention uses signal constraints to determine the reliability of inertial data, thereby improving the accuracy of inertial data, which in turn improves the accuracy of time-varying tire lateral stiffness and, consequently, the estimation accuracy of the vehicle's center of gravity lateral slip angle. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of the centroid sideslip angle estimation method of the present invention;

[0052] Figure 2 This is a flowchart illustrating the centroid sideslip angle estimation method of the present invention.

[0053] Figure 3 This is a diagram of the wheel rotation dynamics model in this embodiment;

[0054] Figure 4 This is a segmented affine diagram of the tire lateral stiffness curve in this embodiment;

[0055] Figure 5 This is a schematic diagram of the membership function of the fuzzy logic system in this embodiment;

[0056] Figure 6 This is a schematic diagram illustrating the driving condition judgment based on the rate of change of steering wheel angle in this embodiment;

[0057] Figure 7 This is a schematic diagram illustrating the vehicle driving status determination in this embodiment;

[0058] Figure 8 This is a schematic diagram of the multi-layer neural network structure in this embodiment;

[0059] Figure 9 This is a block diagram of the centroid side slip angle estimation system of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] The purpose of this invention is to provide a method, system, electronic device, and storage medium for estimating the center of gravity sideslip angle of a vehicle, which can improve the estimation accuracy of the center of gravity sideslip angle.

[0062] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] like Figures 1-2 As shown, the present invention provides a method, system, electronic device, and storage medium for estimating the centroid sideslip angle, including:

[0064] Step 100: Acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed. The inertial data includes lateral acceleration and yaw rate.

[0065] Step 200: Determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the preset signal threshold, and the vehicle driving state. As a specific embodiment, the steering wheel angle change rate threshold is 0.2.

[0066] If satisfied, then:

[0067] Step 300: Calculate the tire lateral stiffness of the target vehicle based on the sensor data to obtain the first time-varying tire lateral stiffness. The specific process is as follows:

[0068] Obtain historical tire side stiffness curves; segment the historical tire side stiffness curves into several intervals and corresponding tire side stiffness constants; calculate the target tire side angle based on the sensor data; perform fuzzy logic operations on the target tire side angle to determine the target interval corresponding to the target tire side angle, and use the tire side stiffness corresponding to the target interval as the first time-varying tire side stiffness.

[0069] Step 400: Determine the center-of-gravity lateral angle of the target vehicle based on the first time-varying tire lateral stiffness and the inertial data. The specific process is as follows:

[0070] The first time-varying tire lateral stiffness is used to estimate the lateral angle state to obtain a first predicted value; the inertial data is used to estimate the lateral angle state to obtain a first measured value; and the first predicted value and the first measured value are subjected to unscented Kalman filtering to obtain the centroid lateral angle of the target vehicle.

[0071] If not satisfied, then:

[0072] Step 500: Perform signal compensation based on the steering wheel angle and wheel speed to obtain virtual inertial data. The specific process is as follows:

[0073] Determine the signal compensation model; perform signal compensation on the steering wheel angle and wheel speed input signal compensation model to obtain virtual inertial data.

[0074] The process of determining the signal compensation model is as follows:

[0075] Acquire training data; the training data includes simulation parameters and virtual inertial data corresponding to the simulation parameters; the simulation parameters include: multiple simulated steering wheel angles and multiple simulated wheel speeds used for simulation experiments; construct a multi-layer neural network model; input the training data into the multi-layer neural network model for training, and determine the trained multi-layer neural network model as the signal compensation model.

[0076] Step 600: Calculate the tire lateral stiffness of the target vehicle based on the steering wheel angle, the virtual inertia data, and the wheel speed to obtain the second time-varying tire lateral stiffness. The specific process is the same as the calculation method in step 300.

[0077] Step 700: Determine the center-of-gravity yaw angle of the target vehicle based on the second time-varying tire lateral stiffness and the virtual inertial data. The specific process is as follows:

[0078] The second time-varying tire lateral stiffness is used to estimate the lateral angle state to obtain a second predicted value; the virtual inertial data is used to estimate the lateral angle state to obtain a second measured value; and the second predicted value and the second measured value are subjected to unscented Kalman filtering to obtain the centroid lateral angle of the target vehicle.

[0079] The method further includes, before determining whether the inertial data satisfies the signal constraint conditions:

[0080] The first step is to determine the inertial data limit value based on the steering wheel angle change rate threshold; the absolute value of the inertial data limit value is greater than the preset signal threshold.

[0081] The second step is to determine the signal constraint conditions based on the inertial data limit value, the preset signal threshold, and the vehicle driving state.

[0082] The signal constraint condition is either a first constraint condition or a second constraint condition; the first constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, the absolute value of the inertial data is less than the preset threshold of the signal, and the vehicle's driving state is straight-line driving; the second constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, and the absolute value of the inertial data is greater than or equal to the preset threshold of the signal.

[0083] Based on the above method, taking a distributed drive vehicle as the target vehicle as an example, the specific execution process is as follows:

[0084] First, the dynamic structure of the vehicle system is determined, which is used for subsequent calculations of tire lateral stiffness.

[0085] Considering lateral and yaw motions, the two-degree-of-freedom model of the vehicle can be expressed as:

[0086]

[0087] Where m is the vehicle's curb weight, a y F is the lateral acceleration at the vehicle's center of gravity. xf F yf These are the longitudinal and lateral net forces acting on the front axle tires, respectively, F. xr F yr These are the longitudinal and lateral resultant forces acting on the rear axle tires, δ. f I is the front wheel steering angle. z Let be the vehicle's moment of inertia about the z-axis, r be the yaw rate, and a and b be the front and rear wheelbases, respectively.

[0088] Because it is impossible to directly measure the longitudinal resultant force F of the front wheel using ordinary sensors. xf In this embodiment, an unknown input observer (UIO) is used to observe it. The specific process is as follows:

[0089] consider Figure 3 The dynamic equilibrium equation of the wheel rotation model shown is:

[0090]

[0091] Where J is the moment of inertia of the wheel, ω i Let T be the rotational speed of the i-th wheel (i = 1, 2, 3, 4 representing the left front wheel, right front wheel, left rear wheel, and right rear wheel of the vehicle, respectively). di and T bi R represents the total driving torque and braking torque of the i-th wheel, respectively. eff This is the effective radius of the wheel.

[0092] The wheel rotation speed signal ω contains noise, and the angular acceleration signal obtained by directly differentiating it often exhibits severe fluctuations. Furthermore, actual vehicles are frequently subjected to various external disturbances during operation, such as uneven road surfaces. In the presence of disturbances, the following UIO structure can bring the system error close to zero:

[0093]

[0094] in, Let ρ be the observed longitudinal force of the i-th wheel. i χ² is a positive constant (the larger the value, the faster the observer response and the smaller the error). i As an intermediate variable. Observed. as well as The sum of these forces constitutes the longitudinal resultant force of the front wheels, i.e. Similarly, the longitudinal force on the left and right wheels of the rear axle The sum of these is the resultant longitudinal force of the rear axle, i.e.

[0095] The resultant lateral force F of the front and rear axles yf F yr This can be obtained using a tire model that incorporates time-varying lateral stiffness:

[0096]

[0097] Among them, C f C r The time-varying lateral stiffness of the front and rear wheels of the two-degree-of-freedom vehicle model, α, are respectively. f α r These are the real-time slip angles of the front and rear wheels, respectively, and they have the following relationship:

[0098]

[0099] Among them, v y v x These are the vehicle's lateral and longitudinal speeds, respectively.

[0100] Relation Substituting equations (1) and (2) into equation (3), and considering the process noise of the system, we obtain the following continuous system:

[0101]

[0102] Where, x = [v y ,r] T u = [δ f ,F xf ] T f c () is the state transition function, and w is the process noise vector of the system (its covariance matrix is ​​Q).

[0103] Secondly, the specific process for obtaining the time-varying tire lateral stiffness is as follows:

[0104] Tire lateral stiffness varies due to various factors (primarily tire vertical load, tire-road adhesion coefficient, and slip angle). Considering the low center of gravity of distributed drive vehicles, the change in tire vertical load during vehicle movement is small and its impact can be ignored. Furthermore, road conditions are generally good during normal vehicle operation. Therefore, this embodiment only considers the influence of the slip angle on tire lateral stiffness.

[0105] Based on the piecewise affine principle, a typical tire side slip stiffness curve is divided into five regions according to the size of the side slip angle, and each region is represented by a constant value C. y1 C y2 C y3 C y4 and C y5 Indicates the tire lateral stiffness in each region (e.g.) Figure 4 (As shown). The constant value C y1 C y2 C y3 C y4 and C y5 The lateral stiffness in each region can be approximately constant by fitting experimental data to form a lookup table.

[0106] Furthermore, based on the sensor data and equation (5), the target tire slip angle is calculated. Fuzzy logic is introduced to fuzzify the target tire slip angle, that is, the target tire slip angle is used as the input of the fuzzy logic. The target interval is determined by looking up the table based on the calculation result of the fuzzy logic, and then the tire slip stiffness corresponding to the target interval is obtained, thus obtaining the accurate tire slip stiffness. Figure 5 As shown, the membership functions of the fuzzy logic system are designed based on specific experimental data and empirical rules. Here, five trapezoidal functions are used to construct the corresponding membership functions. Finally, the weighted average decision method is used to determine the membership functions corresponding to the weight coefficients in the fuzzy logic system.

[0107] Then, the dynamic equations for the measurement signals and the vehicle's center of gravity sideslip angle estimation are determined, and these dynamic equations are used for subsequent calculations of the center of gravity sideslip angle.

[0108] Substituting equations (4) and (5) into equation (1), and considering the composition of the actual lateral acceleration signal and the yaw rate signal, we have the following discrete measurement relationship:

[0109] z k+1 =h(x k+1 ,u k )+v k+1 (7)

[0110] Where z = [a y ,r] T h() is the measurement function, and v represents the measurement noise vector of the system (whose covariance matrix is ​​R).

[0111] Discretizing equation (6), we get:

[0112] x k+1 =f(x) k ,u k )+w k (8)

[0113] Where f() is the discretized state transition function.

[0114] Within the UKF framework, the estimated value of the lateral vehicle speed can be obtained by iterative calculation using equations (7) and (8). Finally, according to the definition Determine the vehicle's center of gravity sideslip angle β.

[0115] Furthermore, when conducting actual measurements on a vehicle, it is also necessary to use signal constraints to assess the reliability of the inertial measurement unit (IMU) in the vehicle.

[0116] Steering wheel angle and longitudinal vehicle speed are the most important factors that change the vehicle's motion state. Simulation experiments have determined the following relationships: For sinusoidal input conditions with the same rate of change of steering wheel angle, both the vehicle's lateral acceleration and yaw rate signals tend to saturate with increasing longitudinal vehicle speed (at which point the road surface adhesion reaches its critical limit); while for non-sinusoidal input conditions with the same rate of change of steering wheel angle, both of these signals increase with increasing longitudinal vehicle speed.

[0117] Based on this, we can... Figure 6 The method shown, based on the rate of change of steering wheel angle, determines the confidence interval of the sensor measurement signal:

[0118] If the steering wheel angle is determined to be a sinusoidal input, the boundary is directly set to a critical constant value (the motion is relatively intense under this condition, and the road surface adhesion will reach its critical limit even at low speeds); if the steering wheel angle is determined to be a non-sinusoidal input, data from a typical double lane change condition is used to fit the boundary function. The lateral acceleration signal is then obtained. and yaw rate signal b r,k The upper boundary.

[0119] If the absolute value of the lateral acceleration signal If the sensor signal is within a certain threshold, it is considered reliable; otherwise, it is considered unreliable. The determination of the yaw rate signal follows the same principle. Furthermore, when the sensor signal is below a certain threshold, it should also be determined whether the vehicle is currently traveling in a straight line (e.g., ...). Figure 7 (As shown). If the vehicle is traveling in a straight line, no action is required; if the vehicle is not traveling in a straight line, the sensor is considered to be faulty.

[0120] When it is determined that the inertial data does not meet the signal constraint conditions, i.e., it is unreliable, virtual signal data needs to be generated to replace the unreliable data.

[0121] When the lateral acceleration a y When sensor measurements such as yaw rate r are unreliable (due to jumps or loss), a signal compensation module based on a multi-layer neural network is used to generate a virtual signal that matches the vehicle's operating conditions, replacing the actual sensor measurements to complete the update step in state estimation.

[0122] Training data for multilayer neural networks can be obtained through simulation experiments at different vehicle speeds (including longitudinal vehicle speed v). x Steering wheel angle δ, rate of change of steering wheel angle Its output consists of virtual sensor signals (including lateral acceleration and yaw rate). The multilayer neural network has a three-layer structure with 3 input layer nodes and 2 output layer nodes. The hidden layers are obtained through a grid search method, with 2 and 3 nodes respectively (specific structure as follows). Figure 8 (As shown).

[0123] Finally, the predicted value obtained by estimating the yaw angle state through time-varying tire yaw stiffness, and the measurement step obtained by estimating the yaw angle state through inertial data (or virtual inertial data), are input into the unscented Kalman filter for unscented Kalman filtering operation.

[0124] The UKF filtering result is typically a weighted sum of calculated values ​​based on the kinetic model and the measurement results. However, mismatches in the kinetic model parameters and anomalies in the sensor measurement signals (such as signal abrupt changes or loss) can cause the final estimation result to deviate from the true value.

[0125] To improve the robustness of the UKF algorithm, a two-layer structure is proposed: first, the validity of the propagation results of the sigma points in the prediction step is checked; second, the validity of the propagation results of the sigma points in the update step is checked. This aims to reduce errors caused by inaccurate sigma points, model parameter mismatch, etc.

[0126] Considering that the vehicle's motion state does not change abruptly during stable operation, the estimation results based on the dynamic model will also not change abruptly, meaning the rate of change of the state variables is bounded. Therefore, for the prior estimate at time k, if the rate of change of the state calculated from the state equation exceeds a certain threshold, the prediction result is considered unreliable, and the same applies to the update step. Furthermore, the sensor's measurement signals should also be bounded; when they exceed a threshold under a certain operating condition, they can be considered abnormal signals. The implementation process of the robust unscented Kalman filter based on a two-layer structure is as follows:

[0127] Considering the mean and variance of an n-dimensional random variable x, and using symmetric sampling to sample 2n+1 sigma points in the unscented transformation, we have the following sampling results:

[0128]

[0129] corresponding mean weights Sum of variance weights They are respectively:

[0130]

[0131] In the formula, κ is a parameter that adjusts the distance between the sigma point and the sampling center, which will affect the estimation effect of the algorithm; Let (n+κ)P be the matrix. x The i-th column of the square root matrix obtained by Cholesky decomposition.

[0132] At time k, the system state variables for the next time step are calculated using nonlinear state equations for each sigma point.

[0133]

[0134] Where ΔT is the sampling period. Simultaneously, the validity of the prediction results for each sigma point is verified:

[0135]

[0136] in, That is, under multiple working conditions at a specific vehicle speed The maximum value.

[0137] Furthermore, calculate the mean of the one-step prediction. and variance P k+1|k:

[0138]

[0139]

[0140] Based on the validity test of the sigma point measurements from the sensor measurement signals, an unscented transformation is performed on the predicted state variables to obtain 2n+1 new sigma points ζ. i Based on this, the measurement is updated to obtain the measurement value corresponding to each sigma point. and the expected measurement obtained by weighting it.

[0141]

[0142]

[0143]

[0144] in, B k =[B k1 B k2 ] T For a control system with a fixed sampling time, the sampling period ΔT is generally a constant, therefore the elements in e should be bounded. k The size is determined empirically, and we can let B... k It is a constant column vector.

[0145] The validity of the sigma points generated by the unscented transformation after measurement update typically follows a Bernoulli distribution, meaning there are only two possibilities: valid or invalid. At time k, if an element in e exceeds B... k If the size of the corresponding element in e is not specified, the data is considered invalid, and the measurement update at that point should be discarded, with the measurement from the previous time step used as a replacement; if all elements in e are less than B... k If the corresponding element is found, then the update result at that point is considered valid.

[0146] like Figure 9 As shown, the present invention also provides a centroid sideslip angle estimation system, comprising:

[0147] The data acquisition unit is used to acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed.

[0148] The signal determination unit is used to determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the signal preset threshold, and the vehicle driving state.

[0149] The first centroid offset angle estimation unit is used to: calculate the tire lateral stiffness of the target vehicle based on the sensor data to obtain a first time-varying tire lateral stiffness, and determine the centroid offset angle of the target vehicle based on the first time-varying tire lateral stiffness and the inertial data if the inertial data satisfies the signal constraint conditions.

[0150] The second centroid deflection angle estimation unit is used to: if the inertial data does not satisfy the signal constraint conditions, then:

[0151] Signal compensation is performed based on the steering wheel angle and wheel speed to obtain virtual inertial data; tire lateral stiffness is calculated for the target vehicle based on the steering wheel angle, the virtual inertial data, and the wheel speed to obtain a second time-varying tire lateral stiffness; the center of gravity offset angle of the target vehicle is determined based on the second time-varying tire lateral stiffness and the virtual inertial data.

[0152] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described centroid side deflection angle estimation method.

[0153] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the centroid sideslip angle estimation method as described above.

[0154] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0155] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for estimating the centroid sideslip angle, characterized in that, include: Acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed; Determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the preset signal threshold, and the vehicle driving state; If satisfied, then: The tire lateral stiffness of the target vehicle is calculated based on the sensor data to obtain the first time-varying tire lateral stiffness. The center of gravity offset angle of the target vehicle is determined based on the first time-varying tire lateral stiffness and the inertial data; If not satisfied, then: Signal compensation is performed based on the steering wheel angle and wheel speed to obtain virtual inertial data; The tire lateral stiffness of the target vehicle is calculated based on the steering wheel angle, the virtual inertia data, and the wheel speed to obtain the second time-varying tire lateral stiffness. The center of gravity offset angle of the target vehicle is determined based on the second time-varying tire lateral stiffness and the virtual inertial data.

2. The method for estimating the centroid sideslip angle according to claim 1, characterized in that, Before determining whether the inertial data satisfies the signal constraint conditions, the method further includes: The inertial data limit value is determined based on the steering wheel angle change rate threshold. The signal constraint condition is determined based on the inertial data limit value, the signal preset threshold, and the vehicle driving state; the absolute value of the inertial data limit value is greater than the signal preset threshold. The signal constraint condition is either a first constraint condition or a second constraint condition; the first constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, the absolute value of the inertial data is less than the preset threshold of the signal, and the vehicle's driving state is straight-line driving; the second constraint condition is that the absolute value of the inertial data is less than or equal to the absolute value of the inertial data limit value, and the absolute value of the inertial data is greater than or equal to the preset threshold of the signal.

3. The method for estimating the centroid sideslip angle according to claim 1, characterized in that, The step of calculating the tire lateral stiffness of the target vehicle based on the sensor data to obtain the first time-varying tire lateral stiffness specifically includes: Obtain historical tire lateral stiffness curves; The historical tire lateral stiffness curve is segmented and affined into several intervals and corresponding tire lateral stiffness constants. Calculate the target tire yaw angle based on the sensor data; Fuzzy logic operations are performed on the target tire side angle to determine the target interval corresponding to the target tire side angle, and the tire side stiffness corresponding to the target interval is used as the first time-varying tire side stiffness.

4. The method for estimating the centroid sideslip angle according to claim 1, characterized in that, The step of determining the center-of-gravity yaw angle of the target vehicle based on the first time-varying tire lateral stiffness and the inertia data specifically includes: The first time-varying tire lateral stiffness is estimated by lateral angle state estimation to obtain the first predicted value; The yaw angle state is estimated from the inertial data to obtain the first measurement value; The centroid offset angle of the target vehicle is obtained by performing unscented Kalman filtering on the first predicted value and the first measured value.

5. The method for estimating the centroid sideslip angle according to claim 1, characterized in that, The step of performing signal compensation based on the steering wheel angle and wheel speed to obtain virtual inertial data specifically includes: Determine the signal compensation model; The steering wheel angle and wheel speed input signal compensation model is used to perform signal compensation to obtain virtual inertial data.

6. The method for estimating the centroid sideslip angle according to claim 5, characterized in that, The determined signal compensation model specifically includes: Acquire training data; the training data includes simulation parameters and virtual inertial data corresponding to the simulation parameters; the simulation parameters include: multiple simulated steering wheel angles and multiple simulated wheel speeds used for simulation experiments; Construct a multi-layer neural network model; The training data is input into the multilayer neural network model for training, and the trained multilayer neural network model is determined as the signal compensation model.

7. The method for estimating the centroid sideslip angle according to claim 1, characterized in that, The threshold for the rate of change of steering wheel angle is 0.

2.

8. A centroid sideslip angle estimation system, characterized in that, include: A data acquisition unit is used to acquire sensor data of the target vehicle; the sensor data includes steering wheel angle, inertial data, and wheel speed. A signal determination unit is used to determine whether the inertial data meets the signal constraint conditions; the signal constraint conditions are determined based on the steering wheel angle change rate threshold, the preset signal threshold, and the vehicle driving state. The first centroid deflection angle estimation unit is used to determine if the inertial data satisfies the signal constraint conditions, then: The tire lateral stiffness of the target vehicle is calculated based on the sensor data to obtain the first time-varying tire lateral stiffness. The center of gravity offset angle of the target vehicle is determined based on the first time-varying tire lateral stiffness and the inertial data; The second centroid deflection angle estimation unit is used to: if the inertial data does not satisfy the signal constraint conditions, then: Signal compensation is performed based on the steering wheel angle and wheel speed to obtain virtual inertial data; The tire lateral stiffness of the target vehicle is calculated based on the steering wheel angle, the virtual inertia data, and the wheel speed to obtain the second time-varying tire lateral stiffness. The center of gravity offset angle of the target vehicle is determined based on the second time-varying tire lateral stiffness and the virtual inertial data.

9. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform the centroid sideslip angle estimation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the centroid sideslip angle estimation method as described in any one of claims 1 to 7.

Images