Vehicle control system online feedback correction state prediction method based on multi-order information

Abstract

The invention belongs to the technical field of vehicle dynamics control, and particularly relates to a vehicle control system on-line feedback correction state prediction method based on multi-order information, which comprises the following steps of: firstly, realizing on-line prediction calculation on the multi-order information according to a current system state value, a state change rate, a change acceleration and a permissible tolerance range; according to the information calculation result and the actual deviation degree of the actual state value, differentiated judgment and correction are carried out on the state change rate and the prediction time step length, and a data identification process under limited information is obtained; and reverse iterative optimization is carried out according to different adopted correction modes, so that the optimal prediction and correction calculation process of the system state under different constraint conditions is realized. According to the method, the finite information is analyzed and deeply mined, so that effective analysis of data is realized on the premise of finite data capacity, information prediction in a confidence time step length and interval is realized on the basis of not increasing extra cost, and effective prediction state information is acquired in a time domain.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle dynamics control technology, and specifically relates to a vehicle control system online feedback correction state prediction method based on multi-order information. The state prediction method mainly performs multi-order expansion and online prediction calculation of information based on the currently available vehicle control system state information and its limited tolerance range. In the prediction process of the vehicle control system state information, the information state quantity, information change rate, information change acceleration, and information sampling step length at the historical and current moments should be fully utilized, and combined with external random disturbances, constraints and the upper and lower limits of the tolerance range, the system state judgment, prediction, selection, correction, iterative optimization and correction under multi-order and multi-dimensional information should be performed. In the actual calculation process, it is necessary to simultaneously consider the variation range of the direct sampling information and its multi-order information characteristics, and predict the state value and correct different forms of information under the premise of integrating multiple features, so as to complete the online effective prediction of the state under a limited data set, and provide an effective control criterion for the vehicle control system. Background Art

[0002] With the advent of the information age and the intelligent age, vehicle systems are gradually developing and transforming from mechanical structural components to intelligence, modularization, and integration. Passive mechanical and control systems can no longer meet actual performance and usage requirements. Human-machine collaborative control, intelligent decision-making assistance, highly integrated systems, fully intelligent judgment, and the comprehensive integration of man-machine-environment have provided guidance for the reform and in-depth development of the vehicle industry.

[0003] However, due to the limited installation space of the vehicle itself, the lightweight design requirement also makes the vehicle more sensitive to weight, but at the same time, drivers and users also have higher requirements for the vehicle's stability, safety, driving experience, intelligence and sustainable driving mileage. There is a contradiction between the two, which is manifested in the actual application process that the vehicle can only be equipped with a limited number of sensors, while the control system modules are also increasing and complex. How to obtain more system information based on a single state, how to use existing sensors and their data for feature extraction, and how to conduct deep information mining have become difficulties.

[0004] The actual system is generally divided into three types: white system, black system and gray system according to the degree of information available. Figure 1 As shown. The white system has completely clear information, the black system has completely unclear information, and the gray system has partially clear information. Multi-order feature analysis based on existing data and state prediction based on this are closer to effective data correction and calculation under the gray system.

[0005] Under known quantitative constraints, whether it is possible to extract features based on existing data under the premise of limited scale data through data calculation and deep mining, and complete data prediction and calculation under limited time steps through real-time calculation and feedback correction of errors, this has also become the direction and goal of research. In addition, under existing computing conditions, being able to manifest the intrinsic characteristics of data based on the differentiated errors of data characteristics has become another direction of practical research and exploration. Summary of the invention

[0006] 1. Technical issues to be resolved

[0007] The technical problem to be solved by the present invention is: how to provide a state prediction method based on online feedback correction of multi-order information, so as to realize the confidence state prediction and online correction under the premise of limited data set. The method not only considers the multi-order system state information, but also considers the influence of error range constraints, effective time step and real-time error feedback, and makes a reasonable prediction of the system state on the basis of comprehensive consideration of information fusion and cross-talk, and finally obtains the state prediction quantity that meets the actual needs.

[0008] (II) Technical solution

[0009] In order to solve the above technical problems, the present invention provides a method for predicting the online feedback correction state of a vehicle control system based on multi-order information, the method comprising:

[0010] Step 1: Multi-dimensional state calculation;

[0011] Step 2: Correction of state values ​​within the low tolerance range;

[0012] Step 3: Correction of state values ​​within the high tolerance range;

[0013] Wherein, the step 1 comprises:

[0014] Step 101: multi-order state calculation;

[0015] Step 102: Correction of system state change rate;

[0016] Step 103: Optimization calculation of correction coefficient;

[0017] The step 2 comprises:

[0018] Step 201: time step optimization calculation;

[0019] Step 202: state correction based on the optimized time step;

[0020] The step 3 comprises:

[0021] Step 301: time step optimization calculation;

[0022] Step 302: state correction based on optimized time step.

[0023] The step 101 is as follows:

[0024] The state of the vehicle control system at the current time T and its change rate and change acceleration are x T , and The permissible tolerance range of the vehicle control system state is [x l ,x u ], thus obtaining the vehicle control system state x after the initial time step ΔT, that is, at time (T+1) T+1 and its rate of change The predicted amount is:

[0025]

[0026] At this time, according to the upper limit of the vehicle control system state tolerance range x u , calculate its state change rate and its change acceleration at time (T+1) and They are:

[0027]

[0028] According to the lower limit of the vehicle control system state tolerance range x l , calculate its state change rate and its change acceleration at time (T+1) and They are:

[0029]

[0030] On this basis, the system predicts the state x p|T+1 Make a judgment, if x p|T+1 ∈[x l ,x u ], then further according to the rate of change of the state Make a judgment; if At this time, the state prediction is accurate and no feedback correction is needed; if or Then we need to predict the state change rate at time (T+1) Make corrections.

[0031] The step 102 is as follows:

[0032] According to the determination result in step 101, if x p|T+1 ∈[x l ,xu ]and At this time, set the ideal predicted state change rate for:

[0033]

[0034] Where α is the dynamic time step, and ε is the upward correction parameter, which aims to ensure The specific calculation equation is:

[0035]

[0036] If x p|T+1 ∈[x l ,x u ]and At this time, set the ideal predicted state change rate for:

[0037]

[0038] In the formula, λ is the downlink correction parameter, which aims to ensure The specific calculation equation is as follows:

[0039]

[0040] The step 103 is as follows:

[0041] The dynamic time step coefficient α of step 102 is calculated, and the equation is as follows:

[0042]

[0043] In the formula, we have:

[0044]

[0045] In the formula, θ l and θ u are the error following coefficients respectively.

[0046] The error following coefficient θ in step 103 is l and θ u The specific calculation method is as follows:

[0047]

[0048] Based on the corrected Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0049]

[0050] The step 201 is as follows:

[0051] If the x calculated in step 1 p|T+1 <x l , then further calculate the state change rate calculated in step 101 Make a judgment; if At this time, ΔT needs to be optimized and corrected. The time step at this time needs to be based on the state prediction value x p|T+1 With the lower threshold value x l Correct the error between them;

[0052] The predicted state error value e at this time l As shown below:

[0053] e l =x l -x pT+1

[0054] According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R l for:

[0055]

[0056] At this time, there are:

[0057]

[0058] Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined l As shown below:

[0059]

[0060] At this point, the solution to the above equation can be calculated as:

[0061]

[0062] In the formula, k pl , k il and k dl For the error e l The proportional, integral and differential coefficients are calculated according to the error e l Set the limit value of

[0063] If ΔT 1 >0 and ΔT 2 >0, then ΔT=min(ΔT1 ,ΔT 2 ); if ΔT 1 >0 and ΔT 2 <0, then ΔT * =ΔT 1 ; ΔT 1 <0 and ΔT 2 >0, then ΔT * =ΔT 2 ; If ΔT 1 <0 and ΔT 2 <0, then

[0064] The step 202 is as follows:

[0065] According to the optimization step length ΔT calculated in step 201 * , for the predicted state quantity x p|T+1 If the correction is made, the corresponding state change rate correction value is:

[0066]

[0067] Based on the calculation and determination results in claim 6, if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0068]

[0069] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0070]

[0071] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0072]

[0073] Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0074]

[0075] The step 301 is as follows:

[0076] If the x calculated in step 101 p|T+1>x u , then further calculate the state change rate calculated in step 101 Make a judgment; if At this time, ΔT needs to be optimized and corrected. The time step at this time needs to be based on the state prediction value x p|T+1 and the upper threshold value x u Correct the error between them;

[0077] The predicted state error value e at this time u As shown below:

[0078] e u =x u -x pT+1

[0079] According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R u for:

[0080]

[0081] Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined u As shown below:

[0082]

[0083] At this point, the solution to the above equation can be calculated as:

[0084]

[0085] In the formula, k pu , k iu and k du For the error e u The proportional, integral and differential coefficients are calculated according to the error e u Set the limit value of

[0086] If ΔT 3 >0 and ΔT 4 >0, then ΔT=min(ΔT 3 ,ΔT 4 ); if ΔT 3 >0 and ΔT 4 <0, then ΔT * =ΔT 3 ; ΔT 3 <0 and ΔT 4 >0, then ΔT * =ΔT4 ; If ΔT 3 <0 and ΔT 4 <0, then

[0087] The step 302 is as follows:

[0088] According to the optimization step length ΔT calculated in step 301 * ,, for the predicted state quantity x p|T+1 If the correction is made, the corresponding state change rate correction value is:

[0089]

[0090] Based on the calculation and determination results in step 301, if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0091]

[0092] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0093]

[0094] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0095]

[0096] Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0097]

[0098] (III) Beneficial effects

[0099] Compared with the prior art, the present invention proposes an online feedback correction state prediction method for a vehicle control system based on multi-order information based on limited information input. The method performs multi-dimensional and multi-scale comprehensive judgment according to the currently acquired state value and its change rate, change acceleration, and the allowable tolerance range of the state, and then performs online correction according to the offset of the specific state value and its offset range, thereby realizing a differentiated correction process of the state value prediction value, thereby achieving accurate prediction of the state under a limited data set. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 Schematic diagram of classification of different systems.

[0101] Figure 2 It is a schematic diagram of the principle architecture of the state prediction and correction method of the present invention. DETAILED DESCRIPTION

[0102] In order to make the purpose, content, and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings and examples.

[0103] In order to solve the above technical problems, the present invention provides an online feedback correction state prediction method based on multi-order information. The specific process and architecture of the method are as follows: Figure 2 shown.

[0104] In the implementation process of this method, firstly, according to the system state and its rate of change and acceleration, they correspond to x T , and And the permissible tolerance range [x l ,x u ], we can calculate that the system state and its rate of change at time (T+1) correspond to x T+1 and According to the upper limit value x of the system state tolerance range u and the lower limit x l , we can calculate that its state change rate and its change acceleration at time (T+1) are and as well as and Therefore, we can use the predicted state x within the tolerance range to p|T+1 , for the predicted state change rate Make corrections and then realize the predicted state quantity x p|T+1 Online correction of .

[0105] In addition, according to the upper deviation (exceeding the upper limit value of the tolerance range x u ) or under deviation (below the lower limit of the tolerance range x l ) Predicted state x p|T+1 The actual deviation degree of the initial prediction time step ΔT is optimized, and the optimal step value ΔT can be obtained by calculation and judgment. * In view of this, we further consider the predicted state change rate The degree of deviation, thus according to the optimal step value ΔT after iteration * Correct it, and the corresponding optimization value can be obtained Finally, the deviation degree of the variables and the influence of the optimized variable values ​​are comprehensively considered to achieve the prediction of the state quantity x p|T+1 Online correction of .

[0106] Example 1

[0107] The online feedback correction state prediction method based on multi-order information of this embodiment is further described.

[0108] Step 1: Correction of status value within tolerance range

[0109] 1.1 Multi-order state calculation

[0110] At the current time T, the corresponding state of the system and its rate of change and acceleration are x T , and The permissible tolerance range of the system status is [x l ,x u ], from which we can get the system state x after the initial time step ΔT, that is, at time (T+1) T+1 and its rate of change The predicted quantities are:

[0111]

[0112] At this time, according to the upper limit of the system state tolerance range x u , calculate its state change rate and its change acceleration at time (T+1) respectively and Specifically:

[0113]

[0114] According to the system status tolerance range lower limit x l , calculate its state change rate and its change acceleration at time (T+1) respectively and Specifically:

[0115]

[0116] On this basis, the system predicts the state x p|T+1 Make a judgment, if x p|T+1 ∈[x l ,x u ], then further according to the rate of change of the state Make a judgment. If At this time, the state prediction is accurate and no feedback correction is needed; if or Then we need to predict the state change rate at time (T+1) Make corrections.

[0117] 1.2 System state change rate correction

[0118] According to the judgment result in step 1.1, if At this time, set the ideal predicted state change rate for:

[0119]

[0120] Where α is the dynamic time step, and ε is the upward correction parameter, which aims to ensure The specific calculation equation is:

[0121]

[0122] if At this time, set the ideal predicted state change rate for:

[0123]

[0124] In the formula, λ is the downlink correction parameter, which aims to ensure The specific calculation equation is as follows:

[0125]

[0126] 1.3 Correction coefficient optimization calculation

[0127] Further calculation of the dynamic time step coefficient α in step 1.2 gives the following equation:

[0128]

[0129] In the formula, we have:

[0130]

[0131] In the formula, θ l and θ u They are the error following coefficients, which are dynamically adjusted according to the change of the error. The specific calculation method is as follows:

[0132]

[0133] Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0134]

[0135] Step 2: Correction of status values ​​within the low tolerance range

[0136] 2.1 Time step optimization calculation

[0137] If x in step 1.1 p|T+1 <x l , then further the rate of change of state Make a judgment. If At this time, ΔT needs to be corrected, and the time step at this time needs to be based on the state prediction value x p|T+1 With the lower threshold value x l Correct the error between them.

[0138] The predicted state error value e at this time l As shown below:

[0139] e l =x l -x pT+1

[0140] According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R l for:

[0141]

[0142] At this time, there are:

[0143]

[0144] Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined l As shown below:

[0145]

[0146] At this point, the solution to the above equation can be calculated as:

[0147]

[0148] In the formula, k pl , k il and k dl For the error e l The proportional, integral and differential coefficients are calculated according to the error e l The limit value is set.

[0149] If ΔT 1 >0 and ΔT 2>0, then ΔT=min(ΔT 1 ,ΔT 2 ); if ΔT 1 >0 and ΔT 2 <0, then ΔT * =ΔT 1 ; ΔT 1 <0 and ΔT 2 >0, then ΔT * =ΔT 2 ; If ΔT 1 <0 and ΔT 2 <0, then

[0150] 2.2 State correction based on optimized time step

[0151] According to the optimized step length ΔT calculated in step 2.1 * , for the predicted state quantity x p|T+1 The corresponding correction value is:

[0152]

[0153] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0154]

[0155] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0156]

[0157] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0158]

[0159] Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0160]

[0161] Step 3: Correction of status values ​​within high tolerance range

[0162] 3.1 Time step optimization calculation

[0163] If x in step 1.1 p|T+1 >x u , then further the rate of change of state Make a judgment if At this time, ΔT needs to be corrected, and the time step at this time needs to be based on the state prediction value x p|T+1 and the upper threshold value x u Correct the error between them.

[0164] The predicted state error value e at this time u As shown below:

[0165]

[0166] According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R u for:

[0167]

[0168] Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined u As shown below:

[0169]

[0170] At this point, the solution to the above equation can be calculated as:

[0171]

[0172] In the formula, k pu , k iu and k du For the error e u The proportional, integral and differential coefficients are calculated according to the error e u The limit value is set.

[0173] If ΔT 3 >0 and ΔT 4 >0, then ΔT=min(ΔT 3 ,ΔT 4 ); if ΔT 3 >0 and ΔT 4 <0, then ΔT * =ΔT 3 ; ΔT 3 <0 and ΔT 4 >0, then ΔT * =ΔT 4 ; If ΔT3 <0 and ΔT 4 <0, then

[0174] 3.2 State correction based on optimized time step

[0175] According to the optimized step length ΔT calculated in step 3.1 * ,, for the predicted state quantity x p|T+1 The corresponding correction value is:

[0176]

[0177] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0178]

[0179] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0180]

[0181] if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is:

[0182]

[0183] Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

[0184]

[0185] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting the state of online feedback correction of a vehicle control system based on multi-order information, characterized in that: The method comprises: Step 1: Multi-dimensional state calculation; Step 2: Correction of state values ​​within the low tolerance range; Step 3: Correction of state values ​​within the high tolerance range; Wherein, the step 1 comprises: Step 101: multi-order state calculation; Step 102: Correction of system state change rate; Step 103: Optimization calculation of correction coefficient; The step 2 comprises: Step 201: time step optimization calculation; Step 202: state correction based on the optimized time step; The step 3 comprises: Step 301: time step optimization calculation; Step 302: state correction based on optimized time step.

2. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 1, characterized in that: The step 101 is as follows: The state of the vehicle control system at the current time T and its change rate and change acceleration are x T , and The permissible tolerance range of the vehicle control system state is [x l ,x u ], thus obtaining the vehicle control system state x after the initial time step ΔT, that is, at time (T+1) T+1 and its rate of change The predicted amount is: At this time, according to the upper limit of the vehicle control system state tolerance range x u , calculate its state change rate and its change acceleration at time (T+1) and They are: According to the lower limit of the vehicle control system state tolerance range x l , calculate its state change rate and its change acceleration at time (T+1) and They are: On this basis, the system predicts the state x p|T+1 Make a judgment, if x p|T+1 ∈[x l ,x u ], then further according to the rate of change of the state Make a judgment; if At this time, the state prediction is accurate and no feedback correction is needed; if or Then we need to predict the state change rate at time (T+1) Make corrections.

3. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 2, characterized in that: The step 102 is as follows: According to the determination result in step 101, if x p|T+1 ∈[x l ,x u ]and At this time, set the ideal predicted state change rate for: Where α is the dynamic time step, and ε is the upward correction parameter, which aims to ensure The specific calculation equation is: If x p|T+1 ∈[x l ,x u ]and At this time, set the ideal predicted state change rate for: In the formula, λ is the downlink correction parameter, which aims to ensure The specific calculation equation is as follows:

4. The method for predicting the state of vehicle control system online feedback correction based on multi-order information as claimed in claim 3, characterized in that: The step 103 is as follows: The dynamic time step coefficient α of step 102 is calculated, and the equation is as follows: In the formula, we have: In the formula, θ l and θ u are the error following coefficients respectively.

5. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 4, characterized in that: The error following coefficient θ in step 103 is l and θ u The specific calculation method is as follows: Based on the corrected Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

6. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 5, characterized in that: The step 201 is as follows: If the x calculated in step 1 p|T+1 <x l , then further calculate the state change rate calculated in step 101 Make a judgment; if At this time, ΔT needs to be optimized and corrected. The time step at this time needs to be based on the state prediction value x p|T+1 With the lower threshold value x l Correct the error between them; The predicted state error value e at this time l As shown below: e l =x l -x pT+1 According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R l for: At this time, there are: Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined l As shown below: At this point, the solution to the above equation can be calculated as: In the formula, k pl , k il and k dl For the error e l The proportional, integral and differential coefficients are calculated based on the error e l Set the limit value of If ΔT1>0 and ΔT2>0, then ΔT=min(ΔT1,ΔT2); if ΔT1>0 and ΔT2<0, then ΔT * =ΔT1; ΔT1<0 and ΔT2>0, then ΔT * =ΔT2; if ΔT1<0 and ΔT2<0, then 7. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 6, characterized in that: The step 202 is as follows: According to the optimization step length ΔT calculated in step 201 * , for the predicted state quantity x p|T+1 If the correction is made, the corresponding state change rate correction value is: Based on the calculation and determination results in claim 6, if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

8. The method for predicting the online feedback correction state of a vehicle control system based on multi-order information as claimed in claim 7, characterized in that: The step 301 is as follows: If the x calculated in step 101 p|T+1 >x u , then further calculate the state change rate calculated in step 101 Make a judgment; if At this time, ΔT needs to be optimized and corrected. The time step at this time needs to be based on the state prediction value x p|T+1 and the upper threshold value x u Correct the error between them; The predicted state error value e at this time u As shown below: e u =x u -x pT+1 According to the rate of change of state The limit variation range of x p|T+1 The controllable range of variation R u for: Considering the influence of external disturbances and system uncertainty, the controllable variation range R is defined u As shown below: At this point, the solution to the above equation can be calculated as: In the formula, k pu , k iu and k du For the error e u The proportional, integral and differential coefficients are calculated based on the error e u Set the limit value of If ΔT3>0 and ΔT4>0, then ΔT=min(ΔT3,ΔT4); if ΔT3>0 and ΔT4<0, then ΔT * =ΔT3; ΔT3<0 and ΔT4>0, then ΔT * =ΔT4; if ΔT3<0 and ΔT4<0, then 9. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 8, characterized in that: The step 302 is as follows: According to the optimization step length ΔT calculated in step 301 * ,, for the predicted state quantity x p|T+1 If the correction is made, the corresponding state change rate correction value is: Based on the calculation and determination results in step 301, if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: if At this time, based on the initial time step ΔT, After the correction calculation is performed, the correction value is: Based on the revised Further predict the state x p|T+1 If the correction is made, the corresponding correction value is:

10. The vehicle control system online feedback correction state prediction method based on multi-order information as claimed in claim 9, characterized in that: The method performs a multi-dimensional and multi-scale comprehensive judgment based on the currently acquired state value and its change rate, change acceleration, and the permissible tolerance range of the state, and then performs online correction according to the offset of the specific state value and its offset range, thereby realizing a differentiated correction process of the state value prediction value, thereby achieving accurate prediction of the state under a limited data set.

Images