CN117195515A - Nonlinear data dictionary quality estimation method for automatic driving vehicle - Google Patents

Abstract

The invention discloses a nonlinear data dictionary quality estimation method of an automatic driving vehicle, which comprises the following steps: friction coefficient calculation, namely, under the influence of weather and geographical conditions, the friction factor is a dynamic change value, and the measurement is carried out by establishing a dynamic friction calculation model; wind resistance estimation; establishing a ramp kinematics equation, and converting the original problem into a nonlinear problem; solving a kinematic equation on the ramp using nonlinear Newton iterations; constructing a high-dimension digital dictionary, and preparing for quality estimation query of a nonlinear dictionary; the method for searching the adjacent points determines the actual motion condition of the vehicle at a certain moment according to the input variable and searches out the theoretical mass value closest to the actual motion condition; and outputting a quality estimation result and calculating an error range of the quality estimation. The nonlinear digital dictionary method can simultaneously estimate two road conditions of road leveling estimation and quality estimation with gradient, has higher solving precision combined with the digital dictionary, and has wide coverage, easy popularization and smaller restriction.

Description

Nonlinear data dictionary quality estimation method for automatic driving vehicle

Technical Field

The invention relates to the field of automatic calculation of weight of a driving vehicle, in particular to a nonlinear data dictionary quality estimation method of the driving vehicle.

Background

Currently, most countries and regions often require that the weight of a commercial vehicle be measured within a prescribed time period to ensure compliance with relevant regulatory and safety standards. This typically involves measuring the weight of a government approved agency during its annual or semi-annual inspection or yearly test, and some countries and regions also require that the commercial operating vehicle use a weigh station or other measuring station for periodic inspection to ensure that the weight of the truck is within legal limits and meets regulatory requirements. Where an electronic scale or other weight measuring device may be used. However, in some situations, such as when there is a change in truck load capacity, such as after unloading or reloading, a re-check of weight is required, and it is difficult to find a professional truck load measuring company around the perimeter due to the relatively large tonnage of the vehicle. The value of the real-time quality estimation model is easily highlighted. In summary, periodic checking of vehicle weight is important for all commercial vehicles in terms of compliance and safety, and an electronic scale for measuring truck weight is one of the commonly used measuring tools.

In addition, luxury vehicles are typically equipped with a vehicle load mass estimation system that allows for viewing of the current load mass conditions of the vehicle in a multi-function display dashboard or in-vehicle extension system of the vehicle. The system generally collects data related to the load mass of the vehicle through various sensors such as a vehicle suspension sensor, a vehicle speed sensor and the like, processes the data, and then displays related load mass information on an instrument panel. With the help of the system, a driver can more accurately know the current load condition of the vehicle so as to adjust and plan according to the actual condition. This is particularly useful for driving over bumpy, bumpy roads or when cargo of different weight loads is to be transported, and can effectively improve the safety awareness and driving efficiency of the driver. However, the calculation process is relatively complicated, a plurality of sensors are needed, the cost reduction and efficiency improvement strategies are difficult to achieve, and particularly, for common commercial vehicles, the mass production is realized, so that the cost saving is also the key of the consideration of a host factory.

The current common methods for estimating the quality of the automobile through extensive investigation include the following methods:

(1): weight measurement method: the method evaluates the mass of the vehicle by measuring the weight of the vehicle by means of an instrument and comparing it under different loading conditions. The method has the advantages of high accuracy, large restriction on places and instruments, low flexibility, requirement on vehicle standstill and limited maximum tonnage.

(2): VIN code checking method: the VIN code is an abbreviation for vehicle identification number, and each vehicle has a unique corresponding VIN code. By checking the VIN code, important information of a manufacturer, a production date, a vehicle type and the like of the vehicle can be obtained, so that the quality of the vehicle can be accurately estimated.

(3): mechanical inspection method: this method evaluates the quality and reliability of an automobile by inspecting mechanical components of the vehicle, such as the frame, engine, suspension system, steering system, etc. The method has lower accuracy and larger deviation.

(4): historical archive inspection: for the second-hand vehicle, the quality of the vehicle can be evaluated by checking the history file of the second-hand market of the vehicle, knowing related information such as past maintenance records of the vehicle and the like. Also, the method is only suitable for no-load vehicle and lacks real-time performance.

(5): travel test method: the performance and quality of the automobile are evaluated by testing the running performance of the automobile, such as acceleration, braking, turning, suspension, etc.

The invention belongs to a 5 th method running test method, which can dynamically calculate the quality of different road conditions and loading degrees in real time, can meet the real-time requirement, and is flexible and practical. The above methods have advantages and disadvantages, some of which require measurement and inspection by specialized equipment, and some of which require consideration of the effects of human factors. The quality of the automobile can be estimated more comprehensively and accurately by comprehensively using various methods.

Therefore, a rational real-time quality estimation model is found on the premise of not loading a high-end sensor as much as possible, and the problem to be solved by a plurality of vehicle mass production enterprises is solved. In view of the above description of the problem background, the problem to be solved by this patent is again given a branch of economic benefit and engineering land. On the premise of limited cost, an effective commercial vehicle load estimation model is explored.

Disclosure of Invention

The invention aims to provide a nonlinear data dictionary quality estimation method for an automatic driving vehicle, so as to solve the problems in the background technology.

In order to achieve the above purpose, the present invention provides the following technical solutions: a nonlinear data dictionary quality estimation method of an automatic driving vehicle comprises the following steps:

s1: friction coefficient calculation, namely, under the influence of weather and geographical conditions, the friction factor is a dynamic change value, and the measurement is carried out by establishing a dynamic friction calculation model;

s2: wind resistance estimation;

s3: establishing a ramp kinematics equation, and converting the original problem into a nonlinear problem;

s4: solving a kinematic equation on the ramp using nonlinear Newton iterations;

s5: constructing a high-dimension digital dictionary, and preparing for quality estimation query of a nonlinear dictionary;

s6: the method for searching the adjacent points determines the actual motion condition of the vehicle at a certain moment according to the input variable and searches out the theoretical mass value closest to the actual motion condition;

s7: and outputting a quality estimation result and calculating an error range of the quality estimation.

Preferably, in S1, the friction coefficient calculation adopts two strategies to calculate different road sections respectively, and the specific steps include:

s1a: according to the friction coefficient estimation strategy I, approximate equivalent application of friction coefficients of different road segments is realized by adopting a default value and a manual input method;

s1b: adopting a deep learning automatic identification method to identify a corresponding label according to the roughness of a video pavement by adopting a level road friction coefficient estimation strategy II;

s1c: the road surface type label of the deep learning test sample corresponds to the theoretical friction coefficient, namely the road surface type label can be obtained, namely the corresponding friction coefficient is obtained.

Preferably, S2 Stroke resistance F _air The evaluation steps of the values are as follows:

s2a: calculating the wind shielding area of the front section of the commercial vehicle, and obtaining the current running speed of the vehicle;

s2b: giving a piecewise function of wind resistance coefficient calculation, and finally outputting wind resistance F _air Is used for the evaluation of the (c).

Preferably, the specific steps in S4 include:

s4a: establishing a kinematic equation of a slope, and converting the kinematic equation into a nonlinear problem containing unknown parameters m and theta;

s4b: solving a nonlinear motion equation by using a Newton nonlinear iteration method, and setting the iteration times k and an error threshold epsilon;

s4c: outputting an approximate solution through multiple iterationsAnd +.>

Preferably, the specific steps in S5 include:

s5a: performing spatial dispersion on important variables in the equation according to a known range, and establishing a spatial mapping relation between the multidimensional variable and the traction mass ratio R, namely constructing a high-dimensional nonlinear digital dictionary;

s5b: the mass of the commercial vehicle at this moment in theory can be estimated using the ratio of the current actual tractive effort F to the tractive effort mass ratio R.

Preferably, the space of the adjacent point search in S6 is based on the space in the high-dimensional digital dictionary constructed in S4, and the specific steps include:

s6a: searching a theoretical quality reference value closest to the actual vehicle running state in a high-dimensional digital dictionary by using a neighbor point searching method, and preparing for quality estimation output;

s6b: the search is calculated using a neighbor point search method, including a minimum distance error search method and a Ball Tree search algorithm.

Preferably, the specific steps in S5a include:

respectively carrying out equidistant discrete on each variable of (F, theta, v, a) to obtain high-dimensional discrete coordinates, and preparing for subsequent dictionary inquiry;

the vehicle weight estimate is queried using a non-linear dictionary and the result is output, depending on the input parameters (F, θ, v, a), where θ may be iteratively obtained by Newton. And (F, v, a) may be acquired by the vehicle sensor.

Compared with the prior art, the invention has the beneficial effects that:

1. the method is different from the traditional calculation method in the calculation process, and the nonlinear digital dictionary estimation method does not depend on additional electronic sensors such as gyroscopes, accelerometers, three-dimensional digital elevation maps and the like. The cost can be controlled for mass production vehicles, and the method belongs to a better solution.

2. The mass estimation model of the traditional part of low-end vehicles is very simple, is only suitable for road leveling estimation, and cannot meet the mass estimation with gradient. The nonlinear digital dictionary method can simultaneously consider the two road conditions.

3. The vehicle mass estimation problem with the gradient is solved by a separate Newton iteration method, the stability is slightly poor, the vehicle mass estimation problem is easy to sink into local optimum, the error range of the gradient estimation value is less affected, and the deviation of the mass estimation result is larger. The nonlinear digital dictionary quality estimation method just solves the problem, only uses the angle estimated by Newton iteration method, and then combines the higher solving precision of the digital dictionary, and the error of the quality estimation result is in a controllable range.

4. The three-dimensional digital elevation map is used for solving the quality estimation problem, and has the defect that the three-dimensional map is extremely high in cost maintenance, and a certain time difference exists between the real-time updating of a road section and the synchronization of software. Therefore, the quality estimation method of the nonlinear digital dictionary has wide coverage, easy popularization and small limitation.

Drawings

FIG. 1 is a solution flow chart of a nonlinear digital dictionary quality estimation method in an embodiment of the present invention;

FIG. 2 is a graph of the spatial dispersion of torque, theoretical mass estimate, and tractive effort Fx in an embodiment of the invention;

FIG. 3 is a spatial dispersion diagram of acceleration, theoretical mass estimation, and traction mass ratio R in an embodiment of the present invention;

FIG. 4 is a spatial dispersion diagram of an angle, theoretical mass estimate, and traction mass ratio R in an embodiment of the present invention;

fig. 5 is a graph showing the torque of the engine rotating shaft and the torque of the wheels as the speed increases in the embodiment of the invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

In the description of the present invention, it should be noted that the azimuth or positional relationship indicated by the terms "vertical", "upper", "lower", "horizontal", etc. are based on the azimuth or positional relationship shown in the drawings, and are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the apparatus or element referred to must have a specific azimuth, be constructed and operated in a specific azimuth, and thus should not be construed as limiting the present invention.

In the description of the present invention, it should also be noted that, unless explicitly specified and limited otherwise, the terms "disposed," "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.

Referring to fig. 1-5, the present invention provides a technical solution: the nonlinear data dictionary quality estimation method of the automatic driving vehicle mainly aims at a slope stage and is a real-time weight estimation method of a moving vehicle, and the method comprises the following steps of:

s1: pretreatment condition one: the friction coefficient is calculated, and is influenced by weather and geographical conditions, the friction factor is a dynamic change value, and a dynamic friction calculation model is established;

s2: pretreatment condition II: when the vehicle runs at a high speed, the wind resistance is not negligible, and a wind resistance estimation formula is required to be established;

s3: establishes a ramp kinematics equation to convert the original problem into a nonlinear problem,

s4: nonlinear Newton iterations are used to solve the kinematic equations on the ramp,

s5: constructing a high-dimension digital dictionary, preparing for quality estimation query of the nonlinear dictionary,

s6: the method for searching the adjacent points determines the actual motion condition of the vehicle at a certain moment according to the input variable and searches the theoretical mass value closest to the actual motion condition,

s7: outputting the quality estimation resultAnd calculates an error range of the quality estimate.

In this embodiment, the main solving step of the preprocessing condition S1 includes:

s1a: adopting a default value and a manual input method to realize the approximate equivalent application of friction coefficients of different road segments; the model is divided into three conditions of high, medium and low according to the road surface roughness, and the default value is usually set as a medium level u=0.02; when rainy and snowy weather is met, human auxiliary input u=0.01 is needed; when a sandy road or a condition of larger resistance is met, a rough value u=0.03 is required to be input;

s1b: adopting a deep learning automatic identification method to identify a corresponding label according to the roughness of a video pavement by adopting a level road friction coefficient estimation strategy II;

s1c: the road surface type label of the deep learning test sample corresponds to the theoretical friction coefficient, and the obtained road surface type label is equal to the obtained corresponding friction coefficient.

The road label category and the coefficient of sliding friction and the coefficient of rolling friction are detailed in the following table;

in this embodiment, the main solving steps of the preprocessing condition two S2 are as follows:

s2a: the kinematic equation of the slope road also comprises wind resistance F _air The evaluation of the values is calculated as follows:

s2b: the above formula S _car Representing the front end area, k of the commercial vehicle _air The air resistance coefficient is shown, and the range of the automobile resistance coefficient is about: k is more than or equal to 0.3 _air K is less than or equal to 0.7 for different vehicle speeds _air The following relation is satisfied:

s2c: establishing a kinematic equation on a level road (S1 mode) according to Newton' S second law, and solving and outputting estimated vehicle weight values estimated in a level road stage

In this embodiment, S3 establishes a ramp kinematics equation, and the specific steps are described as follows:

s3a: and establishing a kinematic equation of the slope S2, and converting the kinematic equation into a nonlinear problem containing unknown parameters m and theta.

S3b: in combination with Newton's second law, the kinematic equation of the vehicle running on the ramp is shown in formula (4):

F _x ＝umgsinθ-umgcosθ-F _air ＝ma _x (4)

s3c: solving the nonlinear equation established in S2c by Newton nonlinear iteration method to obtain an approximate solutionAnd

in this embodiment, the Newton iteration method described in S4, the specific solving step is described as follows:

s4a: given an initial value of iteration x ₀ Epsilon omega, epsilon is selected>0, then, the iteration is repeated. When the gradient satisfying the function f converges to a small value rangeOr the maximum iteration number n is reached, which can be used as a stop condition;

s4b: calculation of Newton iteration gradient valuesAnd a Hessian matrix of second order partial derivatives>The calculation direction comprises iteration step change amount +.> Updating iteration point x _t+1 ＝x _t -d _t Corresponding momentThe array expression is:

x _t+1 ＝x _t +H ^-1 T (5)

s4c: approximation solution of output model by multiple iterationsThe Newton iteration result is not directly used as a final result because Newton iteration quality estimation errors are large, and angle errors are relatively small, and the Newton iteration quality estimation errors can be used as reference values of nonlinear dictionary queries.

The following table is a table of apparent vehicle dimensions and right ahead cross-sectional area references for a common commercial vehicle:

vehicle model	Long (m)	Wide (m)	High (m)	Area (m) 2 )
					Light truck	5	2	2.6	5.20
Medium-sized truck	6	2.3	2.6	5.98
					Heavy truck	9	2.5	3.5	8.75
Container truck	6-12	2.4-2.6	2.6-3.5	6.24-9.1
					Tank truck	8-14	2.5-3.5	3-4	7.5-14

In this embodiment, S5 is a construction of a nonlinear digital dictionary, and the specific solving steps are described as follows:

s5a: the real-time quality estimation of the vehicle in a moving state is jointly determined by the action of multiple parameters, wherein the parameters affecting the quality change are (F, theta, v, a), F represents the traction force of the vehicle, theta represents the slope inclination angle, v represents the running speed at the current moment, and a is the acceleration of the vehicle;

s5b: discretizing a continuous space, e.g. x, according to the actual range of each variable _i ＝(F _xi ，θ _k ，v _l ，a _j ) ^T Indicating that at time i, the vehicle is traveling on a slope having an inclination angle θ, and the corresponding acceleration is a _j The speed is v _l Longitudinal traction is F _xi The method comprises the steps of carrying out a first treatment on the surface of the Traction matrix F in discrete space _n×n×n×n ＝(F _{x_ijkl} ) _n×n×n×n Each spatial classification may be equally divided by any integer.

In a specific embodiment, the theoretical tractive effort may be derived from engine torque, detailed steps of:

s5b1: the torque T on the engine shaft needs to be calculated first _e Conversion to wheel torque T _w The transmission coefficient is recorded as K _d Transmission gear ratio is rate _gear ，η _axis Is the efficiency of the transmission shaft, the range of the value is generally 0.9-0.95, and the calculation formula can be referred to as follows:

T _w ＝T _e ·K _d ·rate _gear ·η _axis (6)

s5b2: the transmission ratio in the actual use process is closely related to the speed of the vehicle, the speed v of the vehicle and the rotation speed N of the engine _e And wheel speed N _w Reference is made to fig. 5 for theoretical trends; finally, converting the wheel torque into traction force, wherein the traction force meets the ratio of the wheel torque to the wheel radius;

s5c: the ratio of traction force to mass at any moment is marked as R, and the nonlinear relationship between the traction force and the mass is shown in figure 2;

s5d: reasonable interval of acceleration, a _xj ∈[0，10]The equation for equally dividing n into equal discrete values is:

s5e: the discrete effect of the acceleration a and the mass m is shown in figure 3, wherein the dark color is the maximum curved surface function of the mass ratio R, and the light color is the minimum curved surface function.

S5f: the relation between the traction force and the angle directly influences the speed of the vehicle and the friction force on the slope, and the angle range is not more than 7 degrees according to the highway construction standard; considering rough mountain roads, viaducts, culvert road conditions and the like, the theoretical angle range of the model is adjusted toThe mathematical formula of the angle dispersion is as follows:

s5f: the discrete effect of the angle theta and the mass m is shown in figure 4, wherein the dark color is the maximum curved surface function of the mass ratio R, and the light color is the minimum curved surface function;

s5g: the air resistance calculation formula can be considered as a quadratic function with respect to the speed variation, if the speed is in the range ofThe discrete formula of air resistance is:

s5h: the mass m is equally divided into k discrete points in the no-load and full-load maximum limit, and the speed v epsilon (0, 100) is equally divided into q equal parts;

s5i: the same vehicle loads different mass cargoes, moves along slopes with different inclination angles respectively, runs at different acceleration and speed, and generates N=i.j.k running permutation and combination, and each discrete point in the dictionary space corresponds to a vehicle with different actual loads;

s5j: by the above discrete method, a digital dictionary can be constructed, and the method is called a nonlinear digital dictionary quality estimation model in this embodiment because of nonlinear relations among variables.

In this embodiment, S6 is a proximity search algorithm, and the specific solving steps are described as follows:

s6a: using the approach to search, this patent proposes 2 methods, method one: searching for a point with the minimum distance difference value of adjacent points as an adjacent reference value by using a minimum distance error searching method;

s6b: when the discrete grid is thicker, and the number of discrete points in the dictionary is small, a simple calculation formula can be adopted:

s6c: finding the actual tractive effort F in the dictionary _actual Closest theoretical traction force F _theory Outputting a quality index k=index_k, obtaining a unique traction force quality ratio R (i, j, k) by using a table look-up method, and then obtaining a theoretical estimated value corresponding to a digital dictionary method;

s6d: in the second approach point searching method, the patent recommends using a Ball Tree searching algorithm, has a very fast searching speed aiming at high-dimensional data, and can quickly find out the traction force F matched with an actual value _theory The following briefly describes the solving steps of the Ball Tree algorithm:

s6d1: establishing a Ball Tree, and selecting a data point as a root node;

s6d2: selecting a radius to divide the data point into two spherical regions;

s6d3: steps S6d1-S6d2 are recursively performed for each sub-region separately until a stop condition is met (e.g., a specified tree depth is reached or the number of data points in the region is below a threshold);

s6d4: searching nearest neighbor, starting from a root node, and calculating the distance from a target point to a current node;

s6d5: narrowing the search range to a sub-region or sub-tree containing the sphere center according to the type of the current node (sphere region or centroid);

s6d6: recursively repeating step 1-2 until a leaf node is reached;

s6d7: after reaching the leaf node, finding the nearest neighbor point by calculating the distance between the target point and the data point in the leaf node;

s6d8: during the backtracking process it is checked if there may be more recent points in other subtrees and if so, the search is continued.

In this embodiment, S7 outputs a quality estimation result and a calculation error, and specifically includes the following steps:

s7a: dictionary query needs to be completed before outputting resultsStep, also in practice, is a reduced-dimension search, a process of successive approximation of an exact solution, a variable a is known in 4-dimensional space _j Becomes a 3-dimensional space along the extension direction of the acceleration axis), known as v _l The solving space is reduced to a plane, and then the variable theta is known _k Solving for the space reduction into a line, and knowing the actual traction force F _xi The solution is an intersection point, which is the ingenious point of the digital dictionary method;

s7b: only one set of known parameters x need to be input at a time _i ＝(F _xi ，θ _k ，v _l ，a _j ) ^T A closest search method of the Ball Tree is used, and a theoretical value of quality estimation closest to the current motion state can be found;

S7c：x _i ＝(F _xi ，θ _k ，v _l ，a _j ) ^T the high-dimensional discrete solution points constituting the quality estimation are input with a set of actual values as long as the grid is sufficiently denseThe closest traction force quality ratio coefficient Rate can be searched in the discrete solution space, and the actual value +.>Compared with the traction force quality ratio coefficient Rate, a digital dictionary quality estimated value can be obtained;

s7d: the key of the digital dictionary method is to find a multiple relation R of traction force to mass ratio, namely R is traction force mass ratio, then the mass of the commercial vehicle at the moment in theory can be estimated by using the ratio of the current actual traction force F to the traction force mass ratio R, and a specific calculation formula is as follows:

s7e: the relative error of the nonlinear mass estimation is calculated as follows:

wherein m is _ture Is the actual weight of the vehicle and,is the result of the nonlinear quality estimation.

Notably, are: compared with the prior art, the method is superior to the traditional vehicle quality estimation method, the method can estimate the real-time quality of the vehicle by combining a physical kinematic model without combining additional sensors such as a digital elevation map or a gyroscope, the model deployment is simple, the method is economical and practical, the preliminary preparation work is less, the nonlinear digital dictionary method is simple to use, the solving speed is high, the accuracy is reliable, and the method is a new algorithm worthy of popularization and application.

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

Claims (7)

1. The nonlinear data dictionary quality estimation method of the automatic driving vehicle is characterized by comprising the following steps of:

s1: friction coefficient calculation, namely, under the influence of weather and geographical conditions, the friction factor is a dynamic change value, and the measurement is carried out by establishing a dynamic friction calculation model;

s2: wind resistance estimation;

s3: establishing a ramp kinematics equation, and converting the original problem into a nonlinear problem;

s4: solving a kinematic equation on the ramp using nonlinear Newton iterations;

s5: constructing a high-dimension digital dictionary, and preparing for quality estimation query of a nonlinear dictionary;

s6: the method for searching the adjacent points determines the actual motion condition of the vehicle at a certain moment according to the input variable and searches out the theoretical mass value closest to the actual motion condition;

s7: and outputting a quality estimation result and calculating an error range of the quality estimation.

2. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 1, wherein: the friction coefficient calculation in the S1 adopts two strategies to calculate different road sections respectively, and the specific steps comprise:

s1a: according to the friction coefficient estimation strategy I, approximate equivalent application of friction coefficients of different road segments is realized by adopting a default value and a manual input method;

s1b: adopting a deep learning automatic identification method to identify a corresponding label according to the roughness of a video pavement by adopting a level road friction coefficient estimation strategy II;

s1c: the road surface type label of the deep learning test sample corresponds to the theoretical friction coefficient, namely the road surface type label can be obtained, namely the corresponding friction coefficient is obtained.

3. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 1, wherein: s2 Stroke resistance F _air The evaluation steps of the values are as follows:

s2a: calculating the wind shielding area of the front section of the commercial vehicle, and obtaining the current running speed of the vehicle;

s2b: giving a piecewise function of wind resistance coefficient calculation, and finally outputting wind resistance F _air Is used for the evaluation of the (c).

4. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 1, wherein: the specific steps in the step S4 include:

s4a: establishing a kinematic equation of a slope, and converting the kinematic equation into a nonlinear problem containing unknown parameters m and theta;

s4b: solving a nonlinear motion equation by using a Newton nonlinear iteration method, and setting the iteration times k and an error threshold epsilon;

s4c: outputting an approximate solution through multiple iterationsAnd +.>

5. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 1, wherein: the specific steps in the step S5 include:

s5a: performing spatial dispersion on important variables in the equation according to a known range, and establishing a spatial mapping relation between the multidimensional variable and the traction mass ratio R, namely constructing a high-dimensional nonlinear digital dictionary;

s5b: the mass of the commercial vehicle at this moment in theory can be estimated using the ratio of the current actual tractive effort F to the tractive effort mass ratio R.

6. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 1, wherein: the space of the adjacent point search in the S6 is based on the space in the high-dimensional digital dictionary constructed in the S4, and the specific steps comprise:

s6a: searching a theoretical quality reference value closest to the actual vehicle running state in a high-dimensional digital dictionary by using a neighbor point searching method, and preparing for quality estimation output;

s6b: the search is calculated using a neighbor point search method, including a minimum distance error search method and a Ball Tree search algorithm.

7. The method for non-linear data dictionary quality estimation of an autonomous vehicle of claim 5, wherein: the specific steps in the step S5a include:

respectively carrying out equidistant discrete on each variable of (F, theta, v, a) to obtain high-dimensional discrete coordinates, and preparing for subsequent dictionary inquiry;

the vehicle weight estimate is queried using a non-linear dictionary and the result is output, depending on the input parameters (F, θ, v, a), where θ may be iteratively obtained by Newton. And (F, v, a) may be acquired by the vehicle sensor.