A commercial vehicle mass and slope joint estimation method considering whole vehicle mass change

Abstract

The present application belongs to the field of intelligent transportation, and particularly relates to a commercial vehicle mass and slope joint estimation method considering whole vehicle mass change, comprising the following steps: step one, obtaining vehicle state data and vehicle inherent parameters, and calculating and obtaining the acceleration a and transmission ratio i at the kth moment g ; step two, analyzing the longitudinal dynamics of the commercial vehicle, and establishing a longitudinal dynamics mathematical model; step three, establishing an original vehicle mass model and a road slope model; step four, based on the original vehicle mass model and the road slope model, constructing a double forgetting factor least square mass estimation model and a slope estimation model; step five, specifically analyzing the problems involved in the variable mass estimation of the original algorithm; step six, designing a double forgetting factor least square improved algorithm, and building a MATLAB / Simulink improved algorithm model to jointly estimate the commercial vehicle mass and slope and effectively verify the same.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation, and in particular relates to a method for jointly estimating the mass and gradient of commercial vehicles that takes into account changes in the overall vehicle mass. Background Technology

[0002] Commercial vehicles are among the most capable, efficient, and wide-ranging road transport vehicles, holding significant economic value and thus occupying a crucial position in the transportation industry. In recent years, with the continuous growth in commercial vehicle sales and technological advancements, the demands of those working in the commercial vehicle sector have expanded beyond simple transportation. Higher requirements have been placed on these vehicles. For example, profit-driven transportation companies seek substantial reductions in fuel consumption to lower transportation costs, while private owners not only desire fuel efficiency but also enhanced driving safety and ease of operation. This has spurred research into automation and, consequently, extensive studies on vehicle control.

[0003] Modern automotive control systems comprise multiple subsystems. Vehicle stability control relies not only on the vehicle controller but also on acquired vehicle and environmental parameters. Real-time and accurate acquisition of changing or unknown structural parameters (such as vehicle mass) and environmental parameters (such as road gradient) can significantly improve vehicle stability control. The development, implementation, and improvement of control systems are all based on input variables. Given the wide range of mass variations in commercial vehicles and the variability of operating road conditions, vehicle mass and road gradient have become two crucial input variables in current research on the control of automatic transmissions and safety systems in commercial vehicles. Real-time estimation of vehicle dynamics model parameters is fundamental to vehicle control, and vehicle mass and road gradient are important parameters in these models, widely used in vehicle control systems such as auxiliary braking systems, ABS control, cruise control, and transmission shift pattern formulation. Accurate and real-time estimation of vehicle mass and road gradient can effectively improve vehicle power and fuel economy.

[0004] Adjusting the shift control strategy online based on vehicle mass not only makes the vehicle run more smoothly during automatic shifting but also achieves a more economical shift control strategy. Based on information such as road slope, speed, and acceleration, the vehicle power coefficient can be calculated, thereby estimating the engine power required to overcome air resistance and rolling resistance, thus achieving emission control. However, commercial vehicles not only experience a wide range of mass variations during transportation but also face issues such as the coupling between road slope and vehicle mass during identification. These problems pose a severe challenge to vehicle parameter estimation and control, seriously affecting the accuracy of perception of vehicle mass and road slope during actual driving. Therefore, exploring a joint estimation method for commercial vehicle mass and slope that considers changes in vehicle mass is of great significance for improving the adaptability of AMT control in commercial vehicles under the influence of varying mass and slope, and for improving power performance and fuel economy.

[0005] Currently, methods for obtaining vehicle weight and road gradient can be divided into three categories: One category is sensor-based methods. This method involves installing additional sensors on the vehicle, such as pendulum angular displacement sensors or inertial navigation systems, to directly measure the gradient angle and then calculate the vehicle weight. For example, collecting gradient signals obtained by adding accelerometers or GPS global positioning systems allows for the calculation of the road gradient angle, and then the vehicle's weight is derived based on longitudinal vehicle dynamics. This method is more suitable for static conditions, such as gradient detection during vehicle start-up. During normal vehicle operation, due to the influence of longitudinal acceleration, suspension deformation, and road bumps, sensors cannot obtain accurate values ​​for road gradient. Furthermore, these methods require high-quality GPS signals, making it difficult to achieve good results when using low-cost GPS. For more accurate detection, high-precision sensors are needed, which inevitably increases vehicle costs.

[0006] The second category is identification methods based on vehicle longitudinal dynamics or kinematics. These methods use parameter identification as their theoretical basis and only require acquiring relevant signal data (engine torque percentage, engine speed, vehicle speed, gear position, etc.) from the vehicle's CAN bus to estimate mass and gradient. The advantage of this approach is that it does not increase the complexity of vehicle wiring and is less susceptible to environmental interference. While there are many methods in this area, when vehicle parameters (weight, etc.) and external resistance (gradient) change simultaneously with driving time and operating conditions, a large amount of data needs to be acquired via the CAN bus to meet the higher accuracy requirements for estimating variable mass and time-varying gradients. This not only increases the complexity of the estimation process but also, in cases of excessive data acquisition, can cause data saturation, causing the algorithm to lose its corrective capabilities.

[0007] The third type of estimation method uses machine learning. This method acquires driving data of vehicles at different weights and road gradients, and then trains the vehicle on the acquired data. This not only achieves accurate estimation of weight and gradient, but also retains recognition performance during braking. However, this method requires a large amount of data to accurately estimate parameter values ​​of different sizes, and the workload of data acquisition is relatively large. Summary of the Invention

[0008] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is to provide a joint estimation method for the mass and slope of commercial vehicles that considers changes in the overall vehicle mass. This method aims to improve the problem that the double forgetting factor least squares method fails to estimate the variable mass of the vehicle due to data saturation, and that the slope estimation performance degrades when starting on an incline. The method proposes an improved algorithm based on the double forgetting factor recursive least squares method. During operation, the vehicle speed is used as the judgment parameter for stopping. After the vehicle stops, the influence of data saturation is eliminated by re-initializing the covariance matrix. Its advantages are that it has fast tracking performance for changing mass, and high mass recognition accuracy and fast convergence speed under both constant mass and changing mass conditions. At the same time, the improved algorithm improves the convergence speed and accuracy of slope estimation when starting on an incline after mass changes.

[0009] To achieve the above objectives, this invention provides a method for jointly estimating the mass and gradient of commercial vehicles, taking into account changes in overall vehicle mass. The method comprises the following steps:

[0010] Step 1: Obtain vehicle state data and inherent vehicle parameters, and calculate the acceleration 'a' and transmission ratio 'i' at time k. g ;

[0011] Step 2: Analyze the longitudinal dynamics of the commercial vehicle and establish a mathematical model of the longitudinal dynamics;

[0012] Step 3: Establish the original vehicle mass model and road slope model;

[0013] Step 4: Based on the original vehicle mass model and road slope model, construct a double-forgetting-factor least squares mass estimation model and a slope estimation model;

[0014] Step 5: Analyze in detail the problems involved in variable quality estimation in the original algorithm;

[0015] Step 6: Design an improved algorithm for the least squares method with double forgetting factors, build an improved algorithm model in Matlab / Simulink to jointly estimate the mass and gradient of commercial vehicles, and verify its effectiveness.

[0016] As an optimization, in step one, the vehicle status data includes at least the engine torque T. eCar speed v x The vehicle's inherent parameters include at least the tire rolling radius r, the final drive ratio i0, the road rolling resistance coefficient f, and the transmission system mechanical efficiency η. t Vehicle air drag coefficient C d And the vehicle's frontal windward area A.

[0017] As an optimization, in step two, the longitudinal dynamics model of the commercial vehicle is as follows:

[0018]

[0019] Among them, T e For engine torque, i g For the transmission ratio, i0 is the main reducer ratio, and η is the transmission ratio of the gearbox. t Where r is the mechanical efficiency of the transmission system, m is the tire rolling radius, g is the vehicle mass, f is the acceleration due to gravity, α is the road rolling resistance coefficient, and C is the road gradient. d Where A is the vehicle's drag coefficient, ρ is the vehicle's frontal area, and v is the air density. x Where is the vehicle speed, δ is the vehicle rotational mass conversion factor, and a x To accelerate the vehicle.

[0020] As an optimization, in step three, the original vehicle mass model is:

[0021]

[0022] Where m is the mass of the car, n is the engine speed, and η is the engine speed. t For the mechanical efficiency of the transmission system, v x For car speed, C d Where A is the vehicle's air resistance coefficient, ρ is the vehicle's frontal area, g is the acceleration due to gravity, f is the road rolling resistance coefficient, and a is the air density. x Let δ be the vehicle acceleration, δ be the vehicle rotational mass conversion factor, and i be the road gradient.

[0023] As an optimization, in step three, the original road slope model is:

[0024]

[0025] Where m is the mass of the car, n is the engine speed, and η is the engine speed. t For the mechanical efficiency of the transmission system, v x For car speed, C d Where A is the vehicle's air resistance coefficient, ρ is the vehicle's frontal area, g is the acceleration due to gravity, f is the road rolling resistance coefficient, δ is the vehicle's rotational mass conversion factor, and a is the air density.x To accelerate the vehicle.

[0026] As an optimization, in step four, the construction of the double-forgotten factor least squares mass estimation model and the slope estimation model can be obtained by the following method: transforming the vehicle longitudinal dynamics model into a least squares identification form to obtain the least squares identification model of vehicle mass; on this basis, establishing a double-forgotten factor recursive least squares algorithm model, and using engine torque, transmission ratio, vehicle speed and acceleration as inputs, building a MATLAB / Simulink algorithm model, with the algorithm output being the vehicle mass estimate and the road slope estimate.

[0027] As an optimization, in step five, the original algorithm addresses the issues involved in variable quality estimation, primarily those related to when the vehicle starts moving and when it stops.

[0028] As an optimization, in step six, the improved algorithm implementation model for commercial vehicle mass estimation and slope estimation can be obtained through the following method: For the calculation logic of the improved algorithm, the vehicle speed is first used as the stopping judgment condition. During the estimation process, the vehicle speed is monitored in real time to see if it is equal to 0. If the vehicle speed is equal to 0, the covariance matrix P(k) will be reassigned to make P(k) = P(0). This enables the algorithm to accurately track and effectively estimate the changing mass. Based on this calculation logic, the original algorithm is improved, and a Simulink model of the improved algorithm is built in Simulink under the MATLAB environment, so that the improved algorithm and simulation results can more accurately describe the joint estimation of commercial vehicle mass and slope.

[0029] As an optimization, in step six, the method of maintaining parameter output commonly used in current research is adopted to process the estimation results during gear shifting and braking. The basic logic is to detect whether there is a gear shifting or braking signal in real time during the estimation process. If there is a gear shifting or braking, the current parameter estimation value of the algorithm will not be updated, but the estimation value of the previous moment will be output.

[0030] In summary, the present invention has the following advantages: Based on the double-forgetting-factor recursive least squares method, the present invention proposes an improved algorithm, which, compared with existing technologies, has the following advantages:

[0031] (1) The improved algorithm has a fast tracking performance for changing quality, and has high recognition accuracy and good stability. When the quality changes, the average absolute error of the improved algorithm for quality estimation and slope estimation is smaller.

[0032] (2) In addition, the improved algorithm has good stability during gear shifting and braking, and high slope estimation accuracy. Under the condition that the mass remains unchanged, the improved algorithm is better than the original algorithm in terms of mass estimation and mean absolute error of slope estimation.

[0033] (3) In terms of slope start estimation after quality change, the improved algorithm considering quality change can quickly and accurately identify the changed quality. Therefore, compared with the original algorithm, the improved algorithm has a shorter slope estimation convergence time and higher estimation accuracy.

[0034] (4) Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from the practice of the invention. The objectives and other advantages of the invention can be realized and obtained from the following description. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating a method for jointly estimating the mass and gradient of a commercial vehicle that considers changes in overall vehicle mass, as described in this invention.

[0036] Figure 2 This is a schematic diagram illustrating the analysis of the problems involved in variable quality estimation using the original algorithm.

[0037] Figure 3 This is the logical flowchart of the improved algorithm.

[0038] Figure 4 This is a MATLAB / Simulink model diagram of the improved algorithm. Detailed Implementation

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, in the description of the present invention, those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0040] This invention provides a method for jointly estimating the mass and gradient of commercial vehicles while considering variations in overall vehicle mass; please refer to [link to relevant documentation]. Figures 1 to 4 It should be noted that the illustrations provided in this embodiment are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0041] This invention acquires vehicle driving status by collecting total data from the CAN bus using a CANape device provided by Vector GmbH (Germany). It then uses the Simulink module in MATLAB software from MathWorks (USA) to build an improved Simulink algorithm model based on the double forgetting factor recursive least squares method, achieving joint estimation of vehicle mass and road slope. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0042] like Figure 1 As shown, this embodiment provides a method for jointly estimating the mass and gradient of a commercial vehicle while considering changes in the overall vehicle mass. Specifically, it includes the following six steps:

[0043] The following is a detailed explanation of the six steps:

[0044] Step 1: Obtain vehicle status data through the vehicle's CAN bus data acquisition device, and calculate other relevant parameters of the model by combining the vehicle status data with the vehicle's inherent parameters.

[0045] This embodiment uses the CAN bus data acquisition equipment currently adopted by major automobile manufacturers. This equipment can connect to the controllers, actuators, and sensors inside the vehicle and transmit the data uniformly to the OBD-II interface. In this embodiment, the CANape device provided by Vector GmbH (Germany) is inserted into the OBD-II interface, and then real-time vehicle status data (engine torque T) is received through the CANape software on the PC. e The vehicle speed (v), engine speed (n), throttle opening (Th), brake signal (Br), steering wheel angle (Steer), and gear information (Ge) are stored in the software database.

[0046] The inherent parameters of a vehicle include at least the tire rolling radius r and the final drive ratio i. g Road rolling resistance coefficient f, transmission system mechanical efficiency η t Vehicle air drag coefficient C d The vehicle's frontal air area is A, and the air density is ρ.

[0047] The acceleration 'a' can be obtained by the velocity 'v' with respect to time difference, and the acceleration at time k can be expressed as:

[0048]

[0049] Where Δt is the sampling period of the data acquisition software module, and v(k) represents the vehicle speed at time k.

[0050] Transmission ratio i gThe ratio of i0 to i0 can be obtained as follows, where i0 is the transmission ratio of the main reducer:

[0051]

[0052] Step 2: Analyze the longitudinal dynamics of the commercial vehicle and establish a mathematical model of the longitudinal dynamics.

[0053] According to automotive theory, the longitudinal dynamics model of a car is as follows:

[0054] F t =F f +F i +F w +F j (Equation 3)

[0055] Among them, F t For vehicle driving force, F f For rolling resistance, F i For slope resistance, F w For air resistance, F j To increase resistance. The relationship between the vehicle's driving force and the acceleration resistance is:

[0056]

[0057] The formula for calculating rolling resistance is:

[0058] F f =mgf cosα (Equation 5)

[0059] According to road design standards, the slope in road design generally does not exceed 10 degrees. Therefore, for model simplification, cosα can be approximated as 1. The magnitude of slope resistance is related to gravity and the slope, as shown in the following formula:

[0060] F i =mg sinα (Equation 6)

[0061] Similarly, sinα can be approximated as 1. The formula for calculating air resistance is:

[0062]

[0063] The longitudinal dynamic mathematical model of the vehicle's driving state data can be obtained through the above calculation formula:

[0064]

[0065] Step 3: Establish the original vehicle mass model and road slope model.

[0066] After conversion, the original car mass model can be obtained as follows:

[0067]

[0068] Similarly, after transformation, the original road slope model can be obtained as follows:

[0069]

[0070] Step 4: Based on the original vehicle mass model and road slope model, construct a double forgetting factor least squares mass estimation model and a slope estimation model.

[0071] 1. Transform the longitudinal dynamics model of the vehicle into a least-squares identification form to obtain the least-squares identification model of the vehicle mass; according to parameter estimation theory, the least-squares form of the system can be expressed as follows:

[0072] y = h T θ (Equation 11)

[0073] Where y is the system output vector, h is the system input vector, and θ is the vector of parameters to be estimated. Let the rolling resistance coefficient f = tan(β) f ), β f This is the equivalent slope.

[0074] After transformation, the least squares form of the vehicle longitudinal dynamics mathematical model can be obtained:

[0075]

[0076] The corresponding parameters for each item are as follows:

[0077] y = a x (Equation 14)

[0078]

[0079]

[0080] (1) Establish a least squares identification model with a forgetting factor.

[0081] After the vehicle starts, the vehicle mass m remains almost constant and is a slow variable. However, since the dynamic system in which it is located is a time-varying system, old data will cause data saturation and affect the identification results. Therefore, a forgetting factor needs to be introduced to highlight the new data and thus enhance the tracking ability of least squares identification.

[0082] The recursive algorithm, after obtaining new input and output data each time, introduces a correction term to adjust the estimate based on the previous time step, so that the estimate meets the accuracy requirements. Its expression can be represented as:

[0083]

[0084] in, This is the estimated parameter vector at the current moment. Let Δ be the estimated parameter vector from the previous time step. k This is a correction term calculated based on the input and output data at the current moment.

[0085] Through derivation and transformation, the basic calculation formula for online identification using the recursive least squares method is as follows:

[0086]

[0087]

[0088] P(k)=[IK(k)φ T (k)]P(k-1) (Equation 20)

[0089] This represents the estimated value at the current moment. Let P(k) represent the estimated value at time k-1, K(k) represent the scaling factor, φ(k) represent the input and output information of the system at the current time, P(k) represent the covariance matrix at the current time, P(k-1) represent the covariance matrix at time k-1, Λ(k) represent the weight at time k, and I represent the identity matrix.

[0090] Basic recursive algorithms can cause data saturation when too much data is collected, thus losing their ability to correct errors. From the perspective of algorithmic calculation principles, the root cause is that the covariance matrix gradually approaches zero, leading to a similarly low proportionality coefficient. Consequently, the current estimate no longer updates with the system output. Therefore, a forgetting factor is needed to reduce the influence of historical data on the current data and improve the algorithm's ability to estimate time-varying parameters. Since the quality and slope identified in this invention have different time-varying characteristics, a forgetting factor recursive least squares method is used to estimate the two state parameters.

[0091] Through derivation and transformation, the loss function of the double forgetting factor least squares algorithm is as follows:

[0092]

[0093] λ1 and λ2 are the input and output information; λ1 and λ2 are the forgetting factors of the two estimated parameters θ1 and θ2, respectively.

[0094] Furthermore, by differentiating the loss function with respect to θ and setting the result to zero, the estimated values ​​of the parameters can be obtained. The recursive forms of the estimated parameters θ1 and θ2 are as follows:

[0095] The recursive form of the least squares quality estimation model with double forgetting factors is as follows:

[0096]

[0097]

[0098]

[0099] Wherein, λ1 is the forgetting factor corresponding to the vehicle mass m parameter to be estimated, and the value range is [0,1). Usually, the forgetting factor λ is not less than 0.9. The smaller λ is, the smaller the contribution of historical data to the current data and the faster the forgetting speed. By default, λ is 0.95.

[0100] The recursive form of the least squares slope estimation model with double forgetting factors is as follows:

[0101]

[0102]

[0103]

[0104] Wherein, λ2 is the forgetting factor corresponding to the road slope i to be estimated, and the value range is [0,1). Usually, the forgetting factor λ is not less than 0.9. The smaller λ is, the smaller the contribution of historical data to the current data and the faster the forgetting speed. By default, λ is 0.95.

[0105] Step 5: Analyze in detail the problems involved in variable quality estimation in the original algorithm.

[0106] The original algorithm addresses the problems involved in variable quality estimation, such as... Figure 2 As shown, the following problems exist:

[0107] (1) After the vehicle starts moving, the quality estimate of the original algorithm The mass is equal to the actual vehicle mass m1. However, due to the large mass parameter value, the error covariance matrix P(k) tends to infinitesimal, approaching 0, within a short period after the algorithm starts running. This causes the gain matrix K(k) to also tend to infinitesimal. Therefore, the algorithm affects the estimated parameters. Its ability to correct changes weakens until it loses its ability to track the quality of changes;

[0108] (2) At time t0, the vehicle briefly stops without turning off the engine, causing a change in its mass. When the vehicle resumes operation, the original algorithm will lose its ability to correct for the change and will be unable to accurately track and estimate the actual mass m2. Due to real-life needs, vehicles frequently experience brief stops and mass changes during operation, such as the rapid loading and unloading of containers by port trucks, and the picking up and dropping off of passengers on buses and taxis. Failure to accurately estimate the changed mass will severely impact the control performance of the vehicle control system.

[0109] Step 6: Design an improved algorithm for the least squares method with double forgetting factors, build an improved algorithm model in Matlab / Simulink to jointly estimate the mass and gradient of commercial vehicles, and verify its effectiveness.

[0110] The principle of the improved algorithm for the double forgetting factor least squares method is attached. Figure 3 The specific steps are as follows:

[0111] (1) Calculate the gain matrix K(k): Substitute the obtained original state parameters and inherent parameters of the vehicle into Equations 18 and 20 to obtain the covariance matrix. P(0), then calculate the gain matrix K(k), its expression is as follows:

[0112] K(k)=P(k-1)h(k)(hT(k)P(k-1)h(k)+λ) -1 (Equation 27)

[0113] (2) Calculate the parameters to be estimated Covariance matrix P(k): Determines whether the vehicle has engaged braking or gear shifting. If such engagements have occurred, the parameters to be estimated at the current time are retained. The algorithm outputs parameters; furthermore, during the estimation process, it monitors whether the vehicle speed is equal to 0 in real time. If the vehicle speed is equal to 0, the covariance matrix P(k) will be reassigned to make P(k) = P(0), thereby realizing the algorithm's accurate tracking and effective estimation of the changing quality. Otherwise, the parameters of the covariance matrix P(0) are maintained.

[0114] Among them, the parameters to be estimated at the current time The parameter output expression is as follows:

[0115]

[0116] The expression for initializing the covariance matrix P(k) is as follows:

[0117] P(k)=P(0) (Equation 29)

[0118] The expression for maintaining the output of the covariance matrix P(k) parameters is as follows:

[0119] P(k)=P(k-1) (Equation 30)

[0120] (3) Update the parameters to be estimated Covariance matrix P(k): In step 2), if the vehicle does not brake or shift gears, the parameters to be estimated are updated sequentially. The covariance matrix P(k) is calculated, and then the process jumps to step 1). Let k = k-1 and substitute it into Equation 27 to calculate the gain matrix K(k) at time k-1. This process is repeated iteratively to update the estimated value of the parameter to be estimated and gradually approach the actual value.

[0121] Among them, the parameters to be estimated are updated. The expression is as follows:

[0122]

[0123] The expression for updating the covariance matrix P(k) is as follows:

[0124] P(k)=(IK(k)h T (k))P(k-1)λ -1 (Equation 32)

[0125] Finally, as attached Figure 4 As shown, based on the improved algorithm of least squares with double forgetting factor, an improved algorithm model in Matlab / Simulink is built to jointly estimate the mass and gradient of commercial vehicles and to verify its effectiveness.

[0126] Verification has shown that the improved algorithm can effectively estimate different constant and varying masses, with better estimation accuracy than the original algorithm. Furthermore, the improved algorithm can accurately estimate the slope in real time under different constant and varying masses, with better estimation performance than the original algorithm.

[0127] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for combined estimation of mass and slope of a commercial vehicle considering the change of the mass of the vehicle, characterized in that, comprising the following steps, step one, obtaining vehicle state data and vehicle inherent parameters, and calculating the acceleration a and transmission ratio i at the kth moment g ; Step two, analyze the longitudinal dynamics of commercial vehicles and establish a longitudinal dynamics mathematical model; Step three, establish an original vehicle mass model and a road slope model; Step four, based on the original vehicle mass model and the road slope model, construct a double-forgetting factor least squares mass estimation model and a slope estimation model; Step five, analyze the problems involved in the original algorithm for variable mass estimation; The problems involved in the original algorithm for variable mass estimation include: (1) after the vehicle has started, the quality estimate of the original algorithm equals the actual vehicle mass ml, but since the quality parameter value is large, in a short time after the algorithm starts running, the error covariance matrix P(k) tends to be infinitesimal, approaching 0, so that the gain matrix K(k) also tends to be infinitesimal, thus weakening the correction ability of the algorithm to the estimated parameter and losing the ability to track the changing mass. (2) At time t0, the vehicle stops for a short time without turning off the engine, at which time the mass changes, and when the vehicle drives again, the original algorithm will not be able to accurately track and estimate the changed actual mass m2 due to the loss of correction ability of the algorithm; Due to the need of actual life, the vehicle stops for a short time and the mass changes frequently during driving, including the rapid loading and unloading of containers by port trucks, the boarding and alighting of passengers by buses and taxis; If the changed mass is not accurately estimated, the control performance of the vehicle control system will be seriously affected; Step six, design an improved algorithm based on double-forgetting factor least squares method, and build a MATLAB / Simulink improved algorithm model to jointly estimate the mass and slope of commercial vehicles and effectively verify them; The principle of the improved algorithm based on double-forgetting factor least squares method includes the following steps: (1) Calculate the gain matrix K(k): Substitute the acquired original state parameters and inherent parameters of the vehicle into Equation 18 and Equation 20 to obtain the covariance matrix P(0), and then calculate the gain matrix K(k), the expression of which is as follows: K(k) = P(k - 1)h(k) (h T (k)P(k - 1)h(k) + λ) -1 (Formula 27) (2) Calculate the parameters to be estimated Covariance matrix P(k): determine whether the vehicle has braking or shifting operation, if such operation, keep the current time parameter to be estimated The further algorithm monitors the vehicle speed in real time during the estimation process. If the vehicle speed is equal to 0, the covariance matrix P(k) will be re-assigned, making P(k) = P(0), so as to realize accurate tracking and effective estimation of the algorithm for changing quality. Otherwise, keep the parameter output of the covariance matrix P(0). The parameter output expression of the current time parameter to be estimated is as follows: The expression for initializing the covariance matrix P(k) is as follows: P(k) = P(0) (Equation 29) The expression for maintaining the covariance matrix P(k) parameter output is as follows: P(k) = P(k-1) (Equation 30) (3) updating the estimated parameters covariance matrix P(k) is updated in turn, and the process jumps to step 1) to calculate the gain matrix K(k) at time k-1 by substituting k=k-1 into equation 27, and the process iterates in this way to update the estimated values of the estimated parameters and gradually approaches the actual values; wherein the expression for updating the estimated parameters covariance matrix P(k) is updated in turn, and the process jumps to step 1) to calculate the gain matrix K(k) at time k-1 by substituting k=k-1 into equation 27, and the process iterates in this way to update the estimated values of the estimated parameters and gradually approaches the actual values; wherein the expression for updating the estimated parameters is as follows: The expression for updating the covariance matrix P(k) is as follows: P(k) = (I - K(k)h T (k))P(k - 1)λ -1 (32) Finally, based on the improved algorithm based on double-forgetting factor least squares method, a MATLAB / Simulink improved algorithm model is built to jointly estimate the mass and slope of commercial vehicles and effectively verify them; It is verified that the improved algorithm can effectively estimate different constant masses and changing masses, and the estimation accuracy is better than that of the original algorithm. The improved algorithm can accurately estimate the slope in real time under different constant masses and changing masses, and the estimation performance is better than that of the original algorithm.

2. The method of claim 1, wherein: In the step one, the vehicle state data at least includes engine torque T e , vehicle running speed v x and engine speed n; the vehicle inherent parameters at least include tire rolling radius r, main reducer transmission ratio i0, road rolling resistance coefficient f, drive train mechanical efficiency η t , vehicle air resistance coefficient C d and vehicle front area A.

3. The method of claim 2, wherein: In the step two, the longitudinal dynamics mathematical model is established as follows: where T e is the engine torque, i g is the transmission gear ratio, i0is the main reducer gear ratio, η t is the mechanical efficiency of the drive train, r is the tire rolling radius, m is the vehicle mass, g is the gravitational acceleration, f is the road rolling resistance coefficient, a is the road slope, C d is the vehicle air resistance coefficient, A is the vehicle frontal area, p is the air density, v x is the vehicle speed, d is the vehicle rotational mass conversion coefficient, a x is the vehicle acceleration.

4. The method of claim 3, wherein: In the step three, the original vehicle mass model is as follows: where m is the mass of the vehicle, n is the engine speed, and η t is the mechanical efficiency of the drive train, v x is the vehicle speed, C d is the vehicle air resistance coefficient, A is the vehicle frontal area, p is the air density, g is the gravitational acceleration, f is the road rolling resistance coefficient, a x is the vehicle acceleration, d is the vehicle rotational mass conversion coefficient, and i is the road slope.

5. The method of claim 3, wherein: In the step three, the original road slope model is as follows: where m is the mass of the vehicle, n is the engine speed, and η t is the mechanical efficiency of the drive train, v x is the speed of the vehicle, C d is the air resistance coefficient of the vehicle, A is the frontal area of the vehicle, p is the air density, g is the acceleration due to gravity, f is the road rolling resistance coefficient, d is the rotational mass conversion coefficient of the vehicle, a x is the acceleration of the vehicle.

6. The method of claim 1, wherein: In the step four, the least squares mass estimation model and the slope estimation model are obtained by the following method: the longitudinal dynamics model of the vehicle is converted into a least squares identification form, thereby obtaining a least squares identification model of the vehicle mass; On this basis, a double-forgetting factor recursive least squares algorithm model is established, and a MATLAB / Simulink algorithm model is built with the engine torque, transmission ratio, vehicle speed and acceleration as inputs, and the algorithm outputs are the vehicle mass estimation value and the road slope estimation value.

7. The method of claim 6, wherein: In the step five, the problems involved in the original algorithm for variable mass estimation mainly include the problems involved in the starting and stopping of the vehicle.

8. The method of claim 7, wherein: In the step six, the commercial vehicle mass estimation and slope estimation improved algorithm model can be obtained by the following method: for the calculation logic of the improved algorithm, first take the vehicle speed as the parking judgment condition, monitor the vehicle speed in real time during the estimation process, if the vehicle speed is equal to 0, the covariance matrix P(k) will be re-assigned once, so that P(k) = P(0), so as to realize the accurate tracking and effective estimation of the algorithm to the changing mass, improve the original algorithm according to the calculation logic, and build the Simulink model of the improved algorithm in the MATLAB environment, so that the improved algorithm and the simulation results can accurately describe the joint estimation of the commercial vehicle mass and the slope.

9. The method of claim 8, wherein: The estimation results during gear shifting and braking are processed by using the method commonly used in the present research to maintain the parameter output, the basic logic of which is to detect in real time during the estimation process whether there is a gear shifting or braking signal, if there is gear shifting or braking, the current parameter estimation value of the algorithm is no longer updated, but the estimation value of the last moment is output.

Images