Road slope adaptive fusion estimation method and device, vehicle and storage medium

Abstract

The application relates to the technical field of parameter estimation, in particular to a road slope adaptive fusion estimation method and device, a vehicle and a storage medium, wherein the method comprises the following steps: identifying positioning information of a vehicle to obtain a speed vector and a pitch angle; constructing Kalman filters according to the speed vector and the pitch angle respectively; designing corresponding noise calculation methods; and realizing adaptive estimation of a first slope and a second slope of a road where the vehicle is located; based on a preset dynamic equation / motion equation and vehicle state information, taking a vehicle pitch angle speed as a slope change rate, establishing a slope state equation, reasonably modeling noise, realizing adaptive estimation of a third slope and a fourth slope of the road where the vehicle is located; and calculating optimal fusion weights to fuse multiple slopes from the first slope to the fourth slope, so as to obtain a final slope of the road where the vehicle is currently located. Therefore, the problem that related technologies cannot realize accurate and rapid slope estimation when the road condition is complex and the slope changes frequently is solved.

Description

Technical Field

[0001] This application relates to the field of parameter estimation technology, and in particular to an adaptive fusion estimation method, device, vehicle, and storage medium for road slope. Background Technology

[0002] Road gradient is one of the fundamental pieces of information about a road, playing a crucial role in the accurate and safe control of vehicles and the construction of road systems. For vehicles, road gradient is essential information for building high-precision vehicle dynamics models and is also key information for many safety systems in modern automobiles, such as hill start assist systems and start-up and shift control. Especially under complex and changing road conditions, rapid and accurate estimation of road gradient can improve vehicle control stability, fuel economy, and driving comfort.

[0003] In recent years, with the development of intelligent electric vehicles, new sensor systems not found in traditional vehicles have brought new solutions for slope estimation, a typical example being the Global Positioning System (GPS). GPS is independent of the vehicle system and unaffected by the vehicle's inherent characteristics, enabling high-precision estimation. However, GPS signals are susceptible to environmental interference, with signal quality degrading or even interrupted in some scenarios such as tunnels and between tall buildings in cities. Therefore, relying solely on GPS is not advisable. Furthermore, related technologies typically use dynamics and kinematics methods to achieve rapid road slope estimation. However, both methods have limitations. Dynamics-based methods require sufficient excitation and struggle to achieve accurate slope estimation at low speeds or when stationary. Kinematics-based methods are highly dependent on acceleration signals, easily leading to cumulative errors. Moreover, both methods largely assume a constant road slope and do not consider the rate of change of the road slope, making accurate and rapid slope estimation difficult under conditions of frequent slope changes. Summary of the Invention

[0004] This application provides a road slope adaptive fusion estimation method, device, vehicle, and storage medium to solve the problems of low accuracy and poor robustness in related technologies when the slope changes frequently.

[0005] The first aspect of this application provides an adaptive fusion estimation method for road slope, including the following steps:

[0006] The system acquires vehicle status information and positioning information from the positioning system; identifies the positioning information to obtain velocity vector and pitch angle; constructs first and second Kalman filters based on velocity vector and pitch angle respectively; designs corresponding noise calculation methods; and achieves adaptive estimation of the first and second slopes of the road where the vehicle is currently located.

[0007] Based on the preset dynamic equations and vehicle state information, the slope state equation is established with the vehicle pitch angular velocity as the slope change rate. Then, a third Kalman filter is constructed, and the influence of vehicle speed, wheel speed and steering is considered to model the noise. At the same time, the low-speed condition is processed to achieve adaptive estimation of the third slope of the road where the vehicle is currently located.

[0008] Based on the preset kinematic equations and vehicle state information, the vehicle pitch angular velocity is used as the slope change rate to establish the slope state equation, and then a fourth Kalman filter is constructed. The noise is modeled considering the curve conditions to achieve adaptive estimation of the fourth slope of the road where the vehicle is currently located.

[0009] Based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, the optimal fusion weight in the sense of linear minimum variance is calculated, and multiple slopes from the first to the fourth slope are fused to obtain the final slope of the road where the vehicle is currently located.

[0010] Optionally, in one embodiment of this application, the step of calculating the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain, error covariance, etc. of the Kalman filters from the first to the fourth slope, and fusing multiple slopes from the first to the fourth slope to obtain the final slope of the road where the vehicle is currently located, includes: if the signal strength is greater than a preset strength, then calculating the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain, error covariance, etc. of the Kalman filters from the first to the fourth slope, and performing optimal weighted fusion based on the first to the fourth slope and their respective weights to obtain the final slope; if the signal strength is less than or equal to the preset strength, then calculating the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain, error covariance, etc. of the Kalman filters from the third to the fourth slope, and performing weighted fusion based on the third slope, the fourth slope, and their respective weights to obtain the final slope.

[0011] Optionally, in one embodiment of this application, the formula for calculating the final slope is:

[0012]

[0013] Where, θ Fus For the final slope, λ i Let i be the weight of the i-th slope. Let i be the slope of the i-th gradient;

[0014] The weight λ of the i-th slope i The calculation formula is:

[0015] λ=[λ1, λ2, λ3, λ4]=Λ -1 e(e T Λ -1 e) -1

[0016] Among them, e=[I n , ..., I n ] T It is a 4n×n matrix, where n is the number of estimated state variables, and Λ is calculated by the following formula:

[0017]

[0018] Optionally, in one embodiment of this application, the step of calculating the first slope and the second slope of the road where the vehicle is currently located based on the velocity vector and the pitch angle respectively includes: constructing first and second Kalman filters based on the velocity vector and the pitch angle respectively, designing corresponding noise calculation methods, and realizing adaptive estimation of the first slope and the second slope of the road where the vehicle is currently located.

[0019] Optionally, in one embodiment of this application, the formula for calculating the third slope is:

[0020]

[0021] Where g is the acceleration due to gravity, F x V is the longitudinal force exerted on the tire by the ground, m is the total mass of the vehicle, and V is the total mass of the vehicle. x V represents the longitudinal vehicle speed. y Let F be the lateral speed, ψ be the yaw angle, and F be the lateral speed. roll For rolling resistance, F air This refers to air resistance.

[0022] The slope state equation is established by using the vehicle's pitch angular velocity as the rate of change of slope:

[0023]

[0024] in, denoted as vehicle pitch angular velocity, k as sampling time, x3 as third gradient, dT as sampling interval, and w3 as random noise.

[0025] The low-speed processing method is as follows:

[0026] θ3=θ3*tanh(|V x |)

[0027] The noise calculation method is as follows:

[0028]

[0029] Among them, RDyn This is the measurement noise of the third slope Kalman filter, where C1, C2, C3, and C4 are calibration constants, δ is the steering wheel angle, and W... ij , i∈{F,R}, j∈{L,R} are wheel speeds, and ε3 is a constant related to random noise.

[0030] Optionally, in one embodiment of this application, the formula for calculating the fourth slope is:

[0031]

[0032] Among them, a x ξ1 represents the measured value of the longitudinal acceleration of the entire vehicle, and ξ1 represents the noise.

[0033] The noise calculation method is as follows:

[0034] R Knt =P2*tanh(P1*(|δ|+ξ2))+ε4

[0035] Among them, R Knt The measurement noise of the fourth slope Kalman filter is given, δ is the steering wheel angle, P1 and P2 are calibration constants, and ξ2 and ε4 are constants set to ensure calculation stability.

[0036] The slope state equation is established by using the vehicle's pitch angular velocity as the rate of change of slope:

[0037]

[0038] Where x4 is the fourth slope and w4 is random noise.

[0039] A second aspect of this application provides a road slope adaptive fusion estimation device, comprising: an acquisition module for acquiring vehicle state information and positioning information from a positioning system; an identification module for identifying the positioning information to obtain a velocity vector and pitch angle, constructing first and second Kalman filters based on the velocity vector and pitch angle respectively, designing corresponding noise calculation methods, and realizing adaptive estimation of the first and second slopes of the road where the vehicle is currently located; and a calculation module for establishing a slope state equation based on a preset dynamic equation and vehicle state information, using the vehicle pitch angular velocity as the slope change rate, and then constructing a third Kalman filter, and considering the effects of vehicle speed, wheel speed, and steering to perform noise calculation. The system models and processes low-speed conditions to adaptively estimate the third slope of the road where the vehicle is currently located. Based on preset kinematic equations and vehicle state information, it establishes a slope state equation using the vehicle's pitch angular velocity as the slope change rate, and then constructs a fourth Kalman filter. It also considers the noise caused by curves to model and achieves adaptive estimation of the fourth slope of the road where the vehicle is currently located. The fusion module calculates the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, and fuses multiple slopes from the first to the fourth slope to obtain the final slope of the road where the vehicle is currently located.

[0040] Optionally, in one embodiment of this application, the fusion module is further configured to: if the signal strength is greater than a preset strength, calculate the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain, error covariance, etc. of the Kalman filters from the first slope to the fourth slope, and perform optimal weighted fusion based on the first slope to the fourth slope and their respective weights to obtain the final slope; if the signal strength is less than or equal to the preset strength, calculate the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain, error covariance, etc. of the Kalman filters from the third slope to the fourth slope, and perform weighted fusion based on the third slope, the fourth slope, and their respective weights to obtain the final slope.

[0041] Optionally, in one embodiment of this application, the formula for calculating the final slope is:

[0042]

[0043] Where, θ Fus For the final slope, λ i Let i be the weight of the i-th slope. Let λ be the weight of the i-th slope. i The calculation formula is:

[0044] λ=[λ1, λ2, λ3, λ4]=Λ -1e(e T Λ -1 e) -1

[0045] Among them, e=[I n , ..., I n ] T It is a 4n×n matrix, where n is the number of estimated state variables, and Λ is calculated by the following formula:

[0046]

[0047] Optionally, in one embodiment of this application, the calculation module is further configured to obtain velocity vectors in multiple directions and a total spatial vector; synthesize a horizontal velocity vector based on the velocity vectors in multiple directions; calculate the first slope based on the horizontal velocity vector and the total spatial vector; and use the pitch angle as the second slope.

[0048] The noise calculation method involves approximating the noise σ as ground velocity noise σ. D The ratio of the horizontal velocity to the horizontal velocity:

[0049]

[0050] Among them, V N V E These represent the northward and eastward velocities, respectively, with ε2 being a relatively small constant and σ being the standard deviation.

[0051] Optionally, in one embodiment of this application, the formula for calculating the third slope is:

[0052]

[0053] Where g is the acceleration due to gravity, F x V is the longitudinal force exerted on the tire by the ground, m is the total mass of the vehicle, and V is the total mass of the vehicle. x V represents the longitudinal vehicle speed. y Let F be the lateral speed, ψ be the yaw angle, and F be the lateral speed. roll For rolling resistance, F air This refers to air resistance.

[0054] The slope state equation is established by using the vehicle's pitch angular velocity as the rate of change of slope:

[0055]

[0056] in, denoted as vehicle pitch angular velocity, k as sampling time, x3 as third gradient, dT as sampling interval, and w3 as random noise.

[0057] The low-speed processing method is as follows:

[0058] θ3=θ3*tanh(|V x |)

[0059] The noise calculation method is as follows:

[0060]

[0061] Among them, R Dyn This is the measurement noise of the third slope Kalman filter, where C1, C2, C3, and C4 are calibration constants, δ is the steering wheel angle, and W... ij , i∈{F, R}, j∈{L, R} are wheel speeds, and ε3 is a constant related to random noise.

[0062] Optionally, in one embodiment of this application, the formula for calculating the fourth slope is:

[0063]

[0064] Among them, a x ξ1 represents the measured value of the longitudinal acceleration of the entire vehicle, and ξ1 represents the noise.

[0065] The noise calculation method is as follows:

[0066] R Knt =P2*tanh(P1*(|δ|+ξ2))+ε4

[0067] Among them, R Knt The measurement noise of the fourth slope Kalman filter is given, δ is the steering wheel angle, P1 and P2 are calibration constants, and ξ2 and ε4 are constants set to ensure calculation stability.

[0068] The slope state equation is established by using the vehicle's pitch angular velocity as the rate of change of slope:

[0069]

[0070] Where x4 is the fourth slope and w4 is random noise.

[0071] A third aspect of this application provides a vehicle, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the road slope adaptive fusion estimation method as described in the above embodiments.

[0072] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the road slope adaptive fusion estimation method as described in the above embodiments.

[0073] Therefore, this application has at least the following beneficial effects:

[0074] By identifying positioning information to obtain velocity vectors and pitch angles, and using these vectors and angles to calculate the first and second slopes of the road where the vehicle is currently located, the advantages of high accuracy and independence from vehicle system influence are effectively utilized. Furthermore, through dynamic and kinematic methods, the third and fourth slopes of the road are calculated, ensuring sufficient accuracy and uninterrupted slope estimation even under unstable positioning conditions. Based on the optimal fusion theory of multi-sensor Kalman filtering and the signal strength of the positioning system, multiple slopes from the first to the fourth slope are fused, effectively addressing conditions with varying slopes. This ensures high accuracy while achieving real-time and robust slope estimation. Therefore, this solves the problem of related technologies struggling to achieve accurate and rapid slope estimation in complex road conditions with frequent slope changes, fully leveraging the advantages of each method and expanding the overall algorithm's working range, thus providing reasonable slope estimation under complex conditions. Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. Attached Figure Description

[0075] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0076] Figure 1 This is a flowchart of an adaptive fusion estimation method for road slope provided according to an embodiment of this application;

[0077] Figure 2 This is a block diagram of a slope estimation system provided according to an embodiment of this application;

[0078] Figure 3 This is a block diagram of a road slope adaptive fusion estimation device according to an embodiment of this application;

[0079] Figure 4 This is a structural schematic diagram of a vehicle provided according to an embodiment of this application.

[0080] Explanation of reference numerals in the attached diagram: Acquisition module-100, Identification module-200, Calculation module-300, Fusion module-400, Memory-401, Processor-402, Communication interface-403. Detailed Implementation

[0081] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0082] Currently, road slope estimation methods are mainly divided into two categories: dynamic-based methods and kinematic-based methods.

[0083] Dynamics-based methods primarily rely on vehicle dynamics models and utilize techniques such as Kalman filtering, extended Kalman filtering, and recursive least squares to estimate road gradients. These methods can achieve high-accuracy estimates when the excitation is sufficient and the vehicle's state is relatively stable. However, due to their reliance on a high-precision dynamics model, they typically require numerous signals to construct the model, which introduces several problems. Firstly, the signals required by the model may not be directly obtainable from sensors, such as tire forces, necessitating additional estimation and increasing the computational burden. Secondly, the failure of any sensor signal can cause the entire algorithm to collapse; the introduction of numerous signals reduces robustness, making dynamics-based methods susceptible to external interference and resulting in poor reliability. Furthermore, dynamics-based methods require sufficient excitation, making accurate gradient estimation difficult to achieve at low vehicle speeds or when the vehicle is stationary.

[0084] Kinematic methods, based on kinematic equations, can be used for slope calculation by filtering signals. They typically require only a small amount of readily available signals, such as IMU (Inertial Measurement Unit) signals and vehicle speed signals. The limited signal count and simple model contribute to the robustness and low computational burden of kinematic methods, enabling a certain level of slope estimation accuracy. However, kinematic methods also have significant drawbacks. For example, methods based on acceleration equations are heavily reliant on acceleration signals, but accelerometers are often noisy, especially at low speeds due to vehicle vibrations and pitch motions. Methods based on pitch angular velocity integration are prone to cumulative errors due to pitch motion during vehicle acceleration and deceleration, random noise, and IMU sensor bias. Therefore, kinematic methods generally have lower accuracy.

[0085] The following description, with reference to the accompanying drawings, outlines an adaptive fusion estimation method, apparatus, vehicle, and storage medium for road slope according to embodiments of this application. Addressing the problems mentioned in the background section, this application provides an adaptive fusion estimation method for road slope. This method obtains the velocity vector and pitch angle by identifying positioning information. Using the velocity vector and pitch angle, it calculates the first and second slopes of the road where the vehicle is currently located, effectively leveraging its high accuracy and independence from vehicle system influences. Furthermore, through dynamics and kinematics methods, it calculates the third and fourth slopes of the road where the vehicle is currently located, ensuring sufficient accuracy and uninterrupted slope estimation even under unstable positioning conditions. Based on the optimal fusion theory of multi-sensor Kalman filtering and the signal strength of the positioning system, it fuses multiple slopes from the first to the fourth slope, effectively handling conditions with varying slopes. While maintaining high accuracy, it achieves high real-time performance and robustness in slope estimation. This solves the problems of related technologies, such as difficulty in achieving accurate and rapid slope estimation under conditions with frequent slope changes, and the existence of low accuracy and poor robustness.

[0086] Specifically, Figure 1 This is a flowchart illustrating an adaptive fusion estimation method for road slope provided in an embodiment of this application.

[0087] like Figure 1 As shown, the adaptive fusion estimation method for road slope includes the following steps:

[0088] In step S101, vehicle information and positioning information from the positioning system are obtained.

[0089] The positioning system can be selected according to the actual situation, such as GPS, Beidou, etc., without specific limitations. The following embodiments use GPS as an example for illustration.

[0090] In this embodiment, vehicle information can include steering wheel angle, wheel speed, vehicle speed, braking torque, driving torque, etc. On the one hand, the influence of vehicle speed, wheel speed, and steering on the dynamic equation is comprehensively considered, and the measurement noise is reasonably modeled; on the other hand, processing is performed for low-speed conditions. Through these processing steps, the normal operating range of the dynamic method is determined, effectively utilizing its high-precision range and facilitating subsequent fusion algorithm processing.

[0091] In step S102, the positioning information is identified to obtain the velocity vector and pitch angle. First and second Kalman filters are constructed based on the velocity vector and pitch angle, respectively. Corresponding noise calculation methods are designed to achieve adaptive estimation of the first and second slopes of the road where the vehicle is currently located.

[0092] It is understood that the embodiments of this application can introduce GPS and construct two methods: velocity vector method and pitch angle method, effectively leveraging its advantages of high precision and independence from vehicle system, filling the gaps where kinematic and dynamic methods cannot function properly under certain conditions.

[0093] In one embodiment of this application, the first and second slopes of the road where the vehicle is currently located are calculated based on the velocity vector and pitch angle, respectively. This includes: constructing first and second Kalman filters based on the velocity vector and pitch angle, designing corresponding noise calculation methods, and achieving adaptive estimation of the first and second slopes of the road where the vehicle is currently located. Taking GPS as an example, GPS is a system independent of the vehicle system, unaffected by the inherent characteristics and motion of the vehicle. Furthermore, due to its high speed accuracy, GPS can obtain relatively accurate slope estimates, and its calculation is simple and has low overhead. GPS outputs velocity vectors in three directions under a navigation coordinate system (also known as a "geographic coordinate system," typically including "Northeast Earth (NED)" and "Northeast Sky (ENU)"). Taking the NED navigation coordinate system as an example, the angle between the composite northeast velocity vector (i.e., the horizontal velocity vector) and the total spatial vector is the first slope, calculated using the following formula:

[0094]

[0095] Where θ1 represents the slope estimated by the GPS velocity vector method, V D V represents ground velocity. E and V N These represent the eastward and northward velocities, respectively. ε1 is a small constant set to ensure stability in numerical calculations. The negative sign is due to the fact that in the NED navigation coordinate system, V... D The positive direction is downward, while the positive direction of the ISO vehicle body vertical coordinate system is upward.

[0096] Considering the increased relative error of GPS signals under low-speed conditions, the accuracy of the first slope estimation also decreases, thus requiring noise modeling. In this embodiment, the noise σ (standard deviation) of the GPS velocity vector method can be approximated as the ground velocity noise σ. D The ratio of the horizontal velocity to the horizontal velocity:

[0097]

[0098] The first slope value is optimally filtered using the Kalman filter algorithm, and the GPS velocity vector method-calculated value is used as the sensor measurement value input to the Kalman filter. The constructed system can be represented as follows:

[0099] x1(k) = x1(k-1) + w1(k)

[0100] θ1 = x1(k) + v1(k)

[0101] Where w1 is the process noise and v1 is the measurement noise, and their variance is:

[0102] R = E[v1(k)v1(k)] T ]=σ 2

[0103] Furthermore, considering that GPS antennas are generally mounted on vehicle bodies, the pitch angle output by the GPS can be used as a reference for slope estimation. The output angle is the superposition of the road slope angle and the vehicle body pitch angle. Since the IMU is generally mounted on the sprung part of the vehicle, its measured acceleration signal is affected by both the slope and the vehicle body pitch. Therefore, the pitch angle output by the GPS is suitable as a compensation signal for acceleration. The second slope output by this method is:

[0104] θ2=θ GPS

[0105] Where θ2 is the slope estimated by the GPS pitch angle method, θ GPS This is the elevation angle signal output by GPS.

[0106] Similarly, in this embodiment, the second slope output by the GPS pitch angle method can be processed by a Kalman filter and then input into the fusion algorithm module, wherein the Kalman filter system equation is consistent with that of the GPS velocity vector method:

[0107] x2(k) = x2(k-1) + w2(k)

[0108] θ2=x2(k)+v2(k)

[0109] Here, w2 and v2 are the process noise and measurement noise, respectively. These two values ​​are small and not significantly related to other factors. Their variance can be set to a small fixed value. The filtered slope and other parameters are input to the fusion algorithm for fusion estimation.

[0110] In step S103, based on the preset dynamic equations and vehicle state information, a slope state equation is established using the vehicle pitch angular velocity as the slope change rate. Then, a third Kalman filter is constructed, and the influence of vehicle speed, wheel speed, and steering is considered to model the noise. At the same time, low-speed conditions are processed to achieve adaptive estimation of the third slope of the road where the vehicle is currently located. Based on the preset kinematic equations and vehicle state information, a slope state equation is established using the vehicle pitch angular velocity as the slope change rate. Then, a fourth Kalman filter is constructed, and the noise is modeled considering the curve conditions to achieve adaptive estimation of the fourth slope of the road where the vehicle is currently located.

[0111] Understandably, GPS signals are susceptible to interference and cannot guarantee all-weather signal coverage, making it necessary to use signals from traditional vehicles for slope estimation. Dynamic methods, as an important approach to slope estimation, offer high accuracy when the model is sufficiently precise. However, they require numerous signals, and the superposition of signal noise makes effective filtering difficult. For example, wheel acceleration noise is significant, potentially leading to inaccurate tire force calculations. Furthermore, they are unsuitable for estimation under conditions of insufficient excitation, such as when stationary or at extremely low speeds.

[0112] This application's embodiments address these shortcomings of dynamic methods by adding low-speed processing, comprehensively considering the influence of vehicle speed, wheel speed, and steering on the dynamic equations, rationally modeling noise, and introducing pitch angular velocity as the rate of gradient change, thereby achieving adaptive and highly real-time gradient estimation. The basic basis for the dynamic gradient estimation method is the vehicle's longitudinal dynamic equations (preset dynamic mode):

[0113]

[0114] Where m is the total vehicle mass, V x V represents the longitudinal vehicle speed. y F represents the lateral speed, and ψ represents the yaw angle. x The longitudinal force exerted on the tire by the ground can be calculated using a single-wheel model:

[0115]

[0116] Among them, T d For driving torque, T b It is the braking torque, I w R is the moment of inertia of the wheel. w That is the wheel radius.

[0117] F slope This is slope resistance, and also the basis for slope estimation:

[0118] F slope =mgsinθ D

[0119] Where, θ D This represents the slope estimated by the dynamic method.

[0120] F roll The rolling resistance is negligible due to its small value; η is the rolling resistance coefficient.

[0121] F roll =mgηcosθ D ≈mgη

[0122] F air Air resistance is a significant factor, especially at high speeds.D It is the air resistance coefficient, A is the vehicle's frontal area, ρ air For air density:

[0123]

[0124] Therefore, the third gradient of the road where the vehicle is currently located, calculated based on the dynamics method, can be expressed as:

[0125]

[0126] Furthermore, considering that the method of estimating the third slope of the road where the vehicle is currently located using dynamics relies on longitudinal excitation, it is very inaccurate and generates a lot of noise at low speeds or when stationary, requiring low-speed processing. Therefore, the embodiments of this application can use the hyperbolic tangent function for processing:

[0127] θ3=θ3*tanh(|V x |)

[0128] Since the tanh function value remains between 0 and 1 when the independent variable is greater than 0, and eventually tends to 1, the tanh function value is extremely small at low vehicle speeds. At this point, the dynamic gradient will also be constrained to a small value, thus reducing the impact on the subsequent fusion estimation algorithm and ensuring the continuity of the estimation, mitigating fluctuations in the estimation results. The calculated gradient value is used as a measurement value of a Kalman filter, and the filtered result is input into the fusion algorithm. Considering that the measurement value of this method includes vehicle pitch motion, the rate of change of the gradient should be equal to the negative of the pitch angular velocity output by the IMU. Therefore, the filtering system equation can be modeled as:

[0129]

[0130] θ3(k) = x3(k) + v3(k)

[0131] in, x1 is the vehicle pitch angular velocity output by the IMU; x2 is the state, i.e., the slope; v2 is the measurement noise.

[0132] Since this method involves numerous signals and is greatly affected by noise, it is necessary to model the measurement noise. The magnitude of the measurement noise will affect the magnitude of the filter covariance, and thus affect its weight in the fusion algorithm. Therefore, reasonable modeling is very important.

[0133] In this embodiment, the noise measured by the dynamic method is modeled as a function of vehicle speed, steering wheel angle, and wheel speed, depending on the vehicle's state under different operating conditions. The influence of vehicle speed and wheel speed is mainly reflected in the slip ratio. A larger slip ratio increases the error in tire force calculation and decreases the accuracy of the dynamic model. Furthermore, at lower vehicle speeds, the longitudinal excitation is smaller, resulting in insufficient representation of dynamic characteristics. The steering wheel angle is mainly used to measure whether the vehicle is traveling in a straight line. When the steering wheel angle is larger, the longitudinal dynamic equation is distorted, and the measurement noise should increase sharply to reduce the confidence of this method. Considering these factors, the measurement noise is set as follows:

[0134]

[0135] Where C1, C2, C3, and C4 are calibration constants, δ is the steering wheel angle, and W ij , i∈{F,R}, j∈{L,R} are wheel speeds, and ε3 is a constant related to random noise.

[0136] Furthermore, kinematic methods are also an important method for slope estimation. Their main advantages are simple models, easy signal acquisition, the ability to estimate slope even under static conditions, and good robustness. However, the acceleration or pitch angular velocity, which are the core signals, contain complex noise and sensor bias, and these problems are difficult to handle, thus limiting the accuracy of kinematic methods. In addition, longitudinal kinematics will fail for curve conditions.

[0137] Considering these shortcomings of kinematic methods, the embodiments of this application first use Kalman filtering for noise processing, and process the curve conditions based on the steering wheel angle. When the angle is large, a large measurement noise is assigned, thereby reducing the confidence of this method and reducing its weight in the fusion algorithm in the following text. In addition, pitch angular velocity is introduced as the slope change rate, thereby realizing adaptive and high real-time estimation of the slope.

[0138] The basic basis for estimating kinematic slope is the kinematic equation of longitudinal acceleration (preset kinematic mode):

[0139]

[0140] Among them, a x ξ1 represents the measured value of the longitudinal acceleration of the whole vehicle (i.e., the acceleration output by the IMU, including the slope component of the gravitational acceleration), θ4 represents the observed value of the kinematic slope.

[0141] Therefore, the fourth gradient of the road where the vehicle is currently located is calculated using the following formula based on kinematic methods:

[0142]

[0143] Similarly, the calculated fourth slope value is used as the measurement value for Kalman filtering. The filtered result also needs to be input into the fusion algorithm. In this embodiment, the method for estimating the fourth slope of the road where the vehicle is currently located using kinematics is the same as the dynamic method. The kinematic slope also includes the vehicle's pitch motion; therefore, its filtering system equation is consistent with the dynamic method.

[0144]

[0145] θ4(k) = x4(k) + v4(k)

[0146] Kinematic slope estimation also requires noise modeling to enhance the algorithm's robustness. Since the IMU is fixed to the vehicle, the measured longitudinal acceleration direction is always aligned with the positive X-axis of the vehicle coordinate system. This means that kinematic slope estimation has high requirements for the vehicle's straight-line stability. Therefore, in this embodiment, the measurement noise can be set to be related to the steering wheel angle, and its modeling is as follows:

[0147] R Knt =P2*tanh(P1*(|δ|+ξ2))+ε4

[0148] Where P1 and P2 are calibration constants, and ξ2 and ε4 are constants set to ensure calculation stability.

[0149] In step S104, based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters for the first to fourth slopes, the optimal fusion weight in the sense of linear minimum variance is calculated. Multiple slopes from the first to fourth slopes are then fused to obtain the final slope of the road where the vehicle is currently located. Since GPS signals are susceptible to interference and cannot guarantee all-weather signal coverage, this embodiment of the application optimizes the fusion of kinematic methods, dynamic methods, GPS velocity vector methods, and GPS pitch angle methods based on the signal strength of the positioning system and the vehicle's current road location. This achieves adaptive fusion estimation of road slope based on GPS, kinematics, and dynamics methods. On the one hand, the magnitude of measurement noise for each method affects its weight in the fusion algorithm. If the measurement noise of a method is high, its final weight in the fusion algorithm will be small; conversely, its weight will be large. This results in methods entering the distortion region acquiring significant measurement noise, which accounts for a small proportion of the fusion algorithm. Methods within the appropriate working range will receive an optimal weight, leading to a reasonable and optimal fusion estimation result. Furthermore, the final fusion estimation result is a weighted fusion of multiple methods, rather than a switching between different methods. This avoids the problem of estimation result jumps caused by frequent switching between different methods.

[0150] In one embodiment of this application, the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope. Multiple slopes from the first to the fourth slope are then fused to obtain the final slope of the road where the vehicle is currently located. This includes: if the signal strength is greater than a preset strength, the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope; optimal weighted fusion is performed based on the first to the fourth slope and their respective weights to obtain the final slope. If the signal strength is less than or equal to the preset strength, the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the third to the fourth slope; weighted fusion is performed based on the third and fourth slopes and their respective weights to obtain the final slope.

[0151] The preset strength can be set according to actual conditions and is not specifically limited. In this embodiment, GPS is determined to be available when the signal strength is greater than the preset strength, and GPS is determined to be unavailable when the signal strength is less than or equal to the preset strength. In actual implementation, this embodiment can integrate four methods—dynamics, kinematics, GPS velocity vector method, and pitch angle method—when GPS is available, with GPS as the primary method, to achieve high-precision fusion estimation of slope. Considering the vulnerability of GPS to environmental interference, only kinematics and dynamics methods are used for fusion estimation when GPS is unavailable, ensuring sufficient accuracy and uninterrupted slope estimation. By designing an access and shutdown switching strategy for road slope observations based on the GPS method, the advantages of each method are fully utilized, maximizing the overall algorithm's working range, thereby providing reasonable slope estimation under various complex working conditions.

[0152] Specifically, such as Figure 2 As shown, the fusion algorithm in this embodiment can employ a multi-sensor Kalman filter fusion algorithm, which can achieve optimal fusion estimation in the sense of linear minimum variance. When GPS is available, the four methods described above are treated as four different sensors, each processed using Kalman filtering. Optimal fusion is achieved using the estimated state, covariance, and Kalman gain output by the Kalman filter. This fusion has two characteristics: unbiasedness, meaning the expected value of the estimated slope equals the expected value of the actual slope; and optimality, meaning the variance of the fusion filter error is minimized. Corresponding Kalman filter equations are constructed for the GPS velocity vector method, GPS pitch angle method, kinematic method, and dynamic method, which can be expressed as:

[0153] θ(k+1)=Aθ(k)+Bu(k)+w(k)

[0154] yi (k)=H i θ(k)+v i (k), i = 1, 2, 3, 4

[0155] Where A, B, and H are the system matrix, input matrix, and measurement matrix, respectively, and y i Let v be the measurement value of the i-th sensor, w be the process noise with a mean of 0 and a covariance of Q. i The noise is measured with a mean of 0 and a covariance of R. i .

[0156] Let θ i Let i = 1, 2, 3, 4, representing the optimal Kalman filter values ​​for the GPS velocity vector method, GPS pitch angle method, kinematic method, and dynamic method, respectively. These values ​​can be calculated using the linear Kalman filter equations.

[0157] θ i (k|k-1)=Aθ i (k-1|k-1)+Bu(k)

[0158] P i (k|k-1)=AP i (k-1|k-1)A T +Q

[0159]

[0160] θ i (k|k)=θ i (k|k-1)+K i (k)(y i (k)-H i (k)x(k|k-1))

[0161] P i (k|k)=[IK i (k)H i ]P i (k|k-1)

[0162] Among them, K i P is the Kalman gain of the i-th sensor. i Let be the error covariance of the i-th sensor.

[0163] For ease of calculation, let's assume that w and v of the above four sensors... i If they are uncorrelated, the covariance of the i-th and j-th sensors can be calculated by the following formula:

[0164] P ij (k|k)=[IK i (k)H i [AP]ij (k-1|k-1)A+Q(k-1)Γ T ][IK j (k)H j ] T

[0165] The fusion weights can be derived from the above formula, let λ i Let λ be the weight of the i-th sensor in the fusion estimation. i It can be calculated using the following formula:

[0166] λ=[λ1, λ2, λ3, λ4]=Λ -1 e(e T Λ -1 e) -1

[0167] In the formula, e = [I n , ..., I n ] T It is a 4n×n matrix, where n is the number of estimated state variables, which is 1 here.

[0168] Λ can be calculated using the following formula:

[0169]

[0170] Finally, the slope estimated by fusion is:

[0171]

[0172] Furthermore, the error covariance of the fusion filter is:

[0173]

[0174] In summary, the slope fusion estimation in this application embodiment uses four sources of estimated slope: GPS velocity vector method, GPS pitch angle method, kinematic method, and dynamic method. Each method establishes a corresponding Kalman filter and outputs the estimated state, covariance, and Kalman gain, which are then used to calculate the fusion weights. For example, in curve conditions, the kinematic and dynamic methods cannot accurately calculate the slope. In this case, by assigning them significant measurement noise, their weights in the fusion algorithm are minimized, while the two GPS methods remain unaffected and can obtain larger weights, ultimately achieving a normal estimation of the slope in curve conditions.

[0175] Furthermore, when GPS signals are available, the GPS Kalman filter algorithm has relatively low measurement noise and therefore a larger weight in the fusion algorithm. In this case, the fusion estimation result is close to that of the GPS method and has higher accuracy. Considering that GPS signals are susceptible to interference and may be interrupted under certain operating conditions, only the kinematic and dynamic methods are fused when GPS is unavailable. In this case, the fusion of these two methods can ensure that the slope estimation is not interrupted and can also output an estimation result with a certain degree of accuracy.

[0176] The adaptive fusion estimation method for road slope proposed in this application obtains the velocity vector and pitch angle by identifying positioning information. Using the velocity vector and pitch angle, it calculates the first and second slopes of the road where the vehicle is currently located, effectively leveraging its advantages of high accuracy and independence from vehicle system influence. Furthermore, through dynamics and kinematics methods, it calculates the third and fourth slopes of the road where the vehicle is currently located, ensuring sufficient accuracy and uninterrupted slope estimation even under unstable positioning conditions. Based on the optimal fusion theory of multi-sensor Kalman filtering and the signal strength of the positioning system, it fuses multiple slopes from the first to the fourth slope, effectively addressing conditions with varying slopes. While maintaining high accuracy, it achieves high real-time performance and robustness in slope estimation. Therefore, it solves the problem that related technologies struggle to achieve accurate and rapid slope estimation in complex road conditions with frequent slope changes, fully utilizing the advantages of each method and expanding the overall algorithm's working range, thus providing reasonable slope estimation under complex conditions.

[0177] Next, referring to the accompanying drawings, an adaptive fusion estimation device for road slope proposed according to an embodiment of this application is described.

[0178] Figure 3 This is a block diagram of a road slope adaptive fusion estimation device according to an embodiment of this application.

[0179] like Figure 3 As shown, the road slope adaptive fusion estimation device 10 includes: an acquisition module 100, an identification module 200, a calculation module 300, and a fusion module 400.

[0180] The system comprises the following modules: Acquisition module 100 acquires vehicle information and positioning information from the positioning system; Identification module 200 identifies the positioning information to obtain the velocity vector and pitch angle, constructs first and second Kalman filters based on the velocity vector and pitch angle, designs corresponding noise calculation methods, and achieves adaptive estimation of the first and second slopes of the road where the vehicle is currently located; Calculation module 300 establishes a slope state equation based on preset dynamic equations and vehicle state information, using the vehicle's pitch angular velocity as the slope change rate, and then constructs a third Kalman filter. It also considers the effects of vehicle speed, wheel speed, and steering to reasonably model the noise, and performs calculations for low-speed conditions. To achieve adaptive estimation of the third slope of the road where the vehicle is currently located, the system processes the data to realize adaptive estimation of the fourth slope of the road. Based on the preset kinematic equations and vehicle state information, the slope state equation is established with the vehicle pitch angular velocity as the slope change rate. Then, a fourth Kalman filter is constructed, and the noise is reasonably modeled considering the curve conditions to realize adaptive estimation of the fourth slope of the road where the vehicle is currently located. The fusion module 400 is used to calculate the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first slope to the fourth slope. The fusion module ... In one embodiment of this application, the fusion module 400 further calculates the optimal fusion weights in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, and fuses multiple slopes from the first to the fourth slope to obtain the final slope of the road where the vehicle is currently located. This includes: if the signal strength is greater than a preset strength, then the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, and optimal weighted fusion is performed based on the first to the fourth slope and their respective weights to obtain the final slope; if the signal strength is less than or equal to the preset strength, then the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the third to the fourth slope, and weighted fusion is performed based on the third slope, the fourth slope and their respective weights to obtain the final slope.

[0181] In one embodiment of this application, the fusion module is further configured to: if the signal strength is greater than a preset strength, calculate the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters of the first to fourth slopes; and perform optimal weighted fusion based on the first to fourth slopes and their respective weights to obtain the final slope; if the signal strength is less than or equal to the preset strength, calculate the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters of the third to fourth slopes; and perform weighted fusion based on the third slope, the fourth slope and their respective weights to obtain the final slope.

[0182] In one embodiment of this application, the formula for calculating the final slope is:

[0183]

[0184] Where, θ Fus For the final slope, λ i Let i be the weight of the i-th slope. Let λ be the weight of the i-th slope. i The calculation formula is:

[0185] λ=[λ1, λ2, λ3, λ4]=Λ -1 e(e T Λ -1 e) -1

[0186] Among them, e=[I n , ..., I n ] T It is a 4n×n matrix, where n is the number of estimated state variables, and Λ is calculated by the following formula:

[0187]

[0188] In one embodiment of this application, the calculation module 300 is further configured to obtain velocity vectors in multiple directions and a total spatial vector; synthesize a horizontal velocity vector based on the velocity vectors in multiple directions; calculate a first slope based on the horizontal velocity vector and the total spatial vector; and use the pitch angle as a second slope; the noise calculation method is to approximate the noise σ as the ground velocity noise σ. D The ratio of the horizontal velocity to the horizontal velocity:

[0189]

[0190] Among them, V N V E These represent the northward and eastward velocities, respectively, with ε2 being a relatively small constant and σ being the standard deviation.

[0191] In one embodiment of this application, the formula for calculating the third slope is:

[0192]

[0193] Where g is the acceleration due to gravity, F x V is the longitudinal force exerted on the tire by the ground, m is the total mass of the vehicle, and V is the total mass of the vehicle. x V represents the longitudinal vehicle speed. y Let F be the lateral speed, ψ be the yaw angle, and F be the lateral speed. roll For rolling resistance, F air This refers to air resistance.

[0194] Using the vehicle's pitch angular velocity as the rate of change of slope, the slope state equation is established as follows:

[0195]

[0196] in, denoted as vehicle pitch angular velocity, k as sampling time, x3 as third gradient, dT as sampling interval, and w3 as random noise.

[0197] The low-speed processing method is as follows:

[0198] θ3=θ3*tanh(|V x |)

[0199] The noise calculation method is as follows:

[0200]

[0201] Among them, R Dyn This is the measurement noise of the third slope Kalman filter, where C1, C2, C3, and C4 are calibration constants, δ is the steering wheel angle, and W... ij , i∈{F,R}, j∈{L,R} are wheel speeds, and ε3 is a constant related to random noise.

[0202] In one embodiment of this application, the formula for calculating the fourth slope is:

[0203]

[0204] Among them, a x ξ1 represents the measured value of the longitudinal acceleration of the entire vehicle, and ξ1 represents the noise.

[0205] The noise calculation method is as follows:

[0206] R Knt =P2*tanh(P1*(|δ|+ξ2))+ε4

[0207] Among them, R KntThe measurement noise of the fourth slope Kalman filter is given, δ is the steering wheel angle, P1 and P2 are calibration constants, and ξ2 and ε4 are constants set to ensure calculation stability.

[0208] Using the vehicle's pitch angular velocity as the rate of change of slope, the slope state equation is established as follows:

[0209]

[0210] Where x4 is the fourth slope and w4 is random noise.

[0211] It should be noted that the foregoing explanation of the road slope adaptive fusion estimation method embodiment also applies to the road slope adaptive fusion estimation device of this embodiment, and will not be repeated here.

[0212] The road slope adaptive fusion estimation device proposed in this application obtains the velocity vector and pitch angle by identifying positioning information. It then uses the velocity vector and pitch angle to calculate the first and second slopes of the road where the vehicle is currently located, effectively leveraging its high accuracy and independence from vehicle system influence. Furthermore, it calculates the third and fourth slopes of the road using dynamics and kinematics methods, ensuring sufficient accuracy and uninterrupted slope estimation even under unstable positioning conditions. Based on the optimal fusion theory of multi-sensor Kalman filtering and the signal strength of the positioning system, it fuses multiple slopes from the first to the fourth slope, effectively addressing varying slope conditions. While maintaining high accuracy, it achieves high real-time performance and robustness in slope estimation. This solves the problem of related technologies struggling to achieve accurate and rapid slope estimation in complex road conditions with frequent slope changes, fully utilizing the advantages of each method and expanding the overall algorithm's working range, thus providing reasonable slope estimation under complex conditions.

[0213] Figure 4 A schematic diagram of the structure of a vehicle provided in an embodiment of this application. The vehicle may include:

[0214] The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0215] When the processor 402 executes the program, it implements the road slope adaptive fusion estimation method provided in the above embodiments.

[0216] Furthermore, the vehicle also includes:

[0217] Communication interface 403 is used for communication between memory 401 and processor 402.

[0218] The memory 401 is used to store computer programs that can run on the processor 402.

[0219] The memory 401 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0220] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0221] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0222] Processor 402 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of this application.

[0223] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described adaptive fusion estimation method for road slope.

[0224] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," 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 application. 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. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0225] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0226] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0227] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (FPGAs), field-programmable gate arrays (FPGAs), etc.

[0228] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. An adaptive fusion estimation method for road slope, characterized in that, Includes the following steps: Obtain vehicle status information and positioning information from the positioning system; The positioning information is identified to obtain the velocity vector and pitch angle. First and second Kalman filters are constructed based on the velocity vector and pitch angle, respectively. A corresponding noise calculation method is designed to achieve adaptive estimation of the first and second slopes of the road where the vehicle is currently located. Based on the preset dynamic equations and the vehicle's state information, the vehicle's pitch angular velocity is used as the slope change rate to establish the slope state equation. Then, a third Kalman filter is constructed, and the effects of vehicle speed, wheel speed, and steering are considered to model the noise. At the same time, low-speed conditions are processed to achieve adaptive estimation of the third slope of the road where the vehicle is currently located. Based on the preset kinematic equations and the vehicle's state information, the vehicle's pitch angular velocity is used as the slope change rate to establish the slope state equation, and then a fourth Kalman filter is constructed. The noise is modeled considering the curve conditions to achieve adaptive estimation of the fourth slope of the road where the vehicle is currently located. Based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, the optimal fusion weight in the sense of linear minimum variance is calculated, and multiple slopes from the first to the fourth slope are fused to obtain the final slope of the road where the vehicle is currently located. If the signal strength is greater than the preset strength, then the optimal fusion weight in the sense of linear minimum variance is calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first slope to the fourth slope. The optimal weighted fusion is then performed based on the first slope to the fourth slope and their respective weights to obtain the final slope. If the signal strength is less than or equal to the preset strength, then the optimal fusion weight in the sense of linear minimum variance is calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filter from the third slope to the fourth slope. The final slope is obtained by weighted fusion based on the third slope, the fourth slope and their respective weights. The formula for calculating the final slope is as follows: in, For the final slope, Let i be the weight of the i-th slope. Let i be the slope of the i-th gradient; The weight of the i-th slope The calculation formula is: in, for The matrix, where n is the number of estimated state variables. Calculated by the following formula: ,in, The covariance feature transfer matrix, The covariance matrix of state i and state j for The traces.

2. The method according to claim 1, characterized in that, The process of constructing first and second Kalman filters based on the velocity vector and the pitch angle, respectively, and designing corresponding noise calculation methods to achieve adaptive estimation of the first and second slopes of the road where the vehicle is currently located includes: Obtain velocity vectors in multiple directions and the total space vector; A horizontal velocity vector is synthesized based on the velocity vectors in the multiple directions, the first slope is calculated based on the horizontal velocity vector and the total spatial vector, and the pitch angle is used as the second slope; The noise calculation method is to calculate the noise... Approximating ground velocity noise The ratio of the horizontal velocity to the horizontal velocity: in, These are the northbound and eastbound speeds, respectively. It is a relatively small constant. The standard deviation is denoted as .

3. The method according to claim 2, characterized in that, The formula for calculating the third slope is: in, It is the acceleration due to gravity. This refers to the longitudinal force exerted on the tire by the ground. For the overall vehicle quality, For longitudinal vehicle speed, Lateral speed, For the horizontal swing angle, For rolling resistance, For air resistance; The slope state equation is established using the vehicle's pitch angular velocity as the rate of change of slope: in, Let k be the vehicle pitch angular velocity, and k be the sampling time. The third slope The sampling interval is... It is random noise; The low-speed processing method is as follows: The noise calculation method is as follows: in, It is the measurement noise of the third-slope Kalman filter. For calibration constants, For steering wheel angle, Let be the wheel speed of wheel ij. Let ij∈{FL, FR, RL, RR}, where FL is the left front wheel, FR is the right front wheel, RL is the left rear wheel, and RR is the right rear wheel. It is a constant related to random noise.

4. The method according to claim 3, characterized in that, The formula for calculating the fourth slope is: in, This is a measured value of the longitudinal acceleration of the entire vehicle. For noise; The noise calculation method is as follows: in, The measurement noise for the fourth slope Kalman filter, For steering wheel angle, For calibration constants, and These are constants set to ensure computational stability; The slope state equation is established using the vehicle's pitch angular velocity as the rate of change of slope: in, It is the fourth slope. It is random noise.

5. A road slope adaptive fusion estimation device, characterized in that, include: The acquisition module is used to acquire vehicle status information and positioning information from the positioning system; The identification module identifies the positioning information to obtain the velocity vector and pitch angle, constructs the first and second Kalman filters based on the velocity vector and pitch angle respectively, designs the corresponding noise calculation method, and realizes the adaptive estimation of the first and second slopes of the road where the vehicle is currently located. The calculation module, based on the preset dynamic equations and the vehicle's state information, establishes a slope state equation using the vehicle's pitch angular velocity as the slope change rate, and then constructs a third Kalman filter. It also considers the effects of vehicle speed, wheel speed, and steering to model noise, while processing low-speed conditions to achieve adaptive estimation of the third slope of the road where the vehicle is currently located. Based on the preset kinematic equations and the vehicle's state information, it establishes a slope state equation using the vehicle's pitch angular velocity as the slope change rate, and then constructs a fourth Kalman filter. It also considers curve conditions to model noise, achieving adaptive estimation of the fourth slope of the road where the vehicle is currently located. The fusion module calculates the optimal fusion weight in the sense of linear minimum variance based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first to the fourth slope, and fuses multiple slopes from the first to the fourth slope to obtain the final slope of the road where the vehicle is currently located. If the signal strength is greater than a preset strength, then the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the first slope to the fourth slope. Optimal weighted fusion is then performed based on the first slope to the fourth slope and their respective weights to obtain the final slope. If the signal strength is less than or equal to the preset strength, then the optimal fusion weights in the sense of linear minimum variance are calculated based on the signal strength of the positioning system and the Kalman gain and error covariance of the Kalman filters from the third slope to the fourth slope. Weighted fusion is then performed based on the third slope, the fourth slope, and their respective weights to obtain the final slope. The formula for calculating the final slope is as follows: in, For the final slope, Let i be the weight of the i-th slope. Let i be the slope of the i-th gradient; The weight of the i-th slope The calculation formula is: in, for The matrix, where n is the number of estimated state variables. Calculated by the following formula: ,,in, The covariance feature transfer matrix, The covariance matrix of state i and state j for The traces.

6. A vehicle, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the road slope adaptive fusion estimation method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the road slope adaptive fusion estimation method as described in any one of claims 1-4.

Images