A global system status online determination method based on chassis modularization

Abstract

The present invention discloses an online determination method for the global system state based on chassis modularization. Based on the dynamic optimization of the target tolerance value and the time step, a dynamic constraint function is established by considering the original response data set at the current time T, the time step variable and the change rate of the original response data set, and the maximum effective step length that both meets the redundant error constraint and maximizes the fusion speed is found to realize the data fusion of multiple module sensors. The statistical characteristics of the fused information data set are extracted by calculating the standard deviation thereof, and valid and invalid data are divided and inversion calculations are performed to obtain valid original response data and invalid original response data. On this basis, the information inefficiency of the response data of each module can be obtained, thereby completing the dynamic determination and online decision-making of the chassis sub-module state and the global performance of the system.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data fusion and vehicle chassis control, and specifically relates to an online determination method for global system status based on chassis modularization. Background Art

[0002] With the deepening trend toward chassis control-by-wire, modularization, and intelligence, online stability assessment, dynamic control, and multi-module coordination of vehicle chassis systems in volatile unmanned environments and complex road conditions have become hot research topics. In real-world environments, vehicle system states dynamically change in response to external factors. In manned vehicles, drivers make timely decisions and adjust their driving behavior based on environmental changes. Under these circumstances, the vehicle chassis system's dynamics are relatively limited, merely adapting to maintain vehicle stability and meeting practical operational requirements. In contrast, in the research and development of advanced driver assistance systems and unmanned vehicles, the role of humans and their involvement in the system are further reduced, with drivers increasingly removed from the decision-making process, with the vehicle chassis's own control system taking primary responsibility. The vehicle chassis is a multi-module system, with information from each module intertwined and complex. Furthermore, as the system evolves over time, the degree of information fusion between modules becomes increasingly complex due to time and environmental factors, further increasing the difficulty of analysis. Therefore, within this vast information architecture, extracting valid information from this voluminous multi-dimensional information while eliminating invalid information to achieve dynamic assessment and online decision-making is a pressing technical challenge. Summary of the Invention

[0003] In view of this, the present invention proposes a global system status online determination method based on chassis modularization, which can dynamically determine the vehicle driving status and make online decisions to assist vehicle control.

[0004] The technical solution for achieving the purpose of the present invention is:

[0005] A method for online determination of global system status based on chassis modularization, comprising the following steps:

[0006] Step 1: Based on the response data sets of the sensors of each sub-module of the chassis at the current time T and the historical time T-ΔT, calculate the tolerance value of the response data set at the current time T, and compare it with the preset target tolerance value. When the tolerance value does not exceed the target tolerance value, calculate the fusion data set based on the response data set at the current time, and enter step 2; when the tolerance value exceeds the target tolerance value, dynamically optimize the time step ΔT to obtain the maximum effective step ΔT max , update the historical time to T-ΔT max , return to the beginning of step 1 and continue.

[0007] Step 2: Calculate the standard deviation of the fused dataset and its subsets, and construct a standard deviation matrix; calculate the probability of each standard deviation in the standard deviation matrix, construct a probability matrix, and obtain the maximum standard deviation in the standard deviation matrix and its corresponding subset in the fused dataset. Will Perform inversion operations as features of the fused dataset to obtain invalid inversion response datasets And calculate the information inefficiency η of the fusion data set; analyze The number of submodules corresponding to the response data contained in is combined with the information inefficiency η to determine the chassis status, and a decision is made based on the chassis status to complete the online determination of the global system status.

[0008] Furthermore, based on the response data sets of the sensors of each submodule of the chassis at the current time T and the historical time T-ΔT, the specific method of calculating the tolerance value of the response data set at the current time T is:

[0009] The response data sets of each sub-module sensor of the chassis at the current time T and the historical time T-ΔT are dimensionally unified to obtain the response data matrix at the current time T And the response data change rate matrix at historical time T-ΔT

[0010] according to and Calculate the sensor tolerance value ω, the expression is:

[0011]

[0012] Where n=1, 2, ..., N is the sequence number of the submodule, and N is the total number of submodules. is the response data matrix of the nth submodule at the current time T, is the response data change rate matrix of the nth submodule at the historical moment T-ΔT.

[0013] Furthermore, the specific method of calculating the fused dataset based on the response dataset at the current time T is:

[0014] According to the target tolerance value γ, tolerance value ω, and the response data matrix at the current time T And the response data change rate matrix at historical time T-ΔT Calculate the fused dataset Its expression is:

[0015]

[0016] in, is the subset corresponding to the nth submodule in the fusion dataset.

[0017] Furthermore, the specific method of dynamically optimizing the time step ΔT is as follows:

[0018] Change the time step ΔT = ΔT j , ΔT j is the jth time step variable; according to the response data set at the current time T and the time T-ΔT j The response data set establishes its dynamic constraint conditions J j , whose expression is:

[0019]

[0020] Among them, E rj is the dynamic constraint function corresponding to the j-th time step, is the subset corresponding to the i-th submodule in the original response dataset at the current time T, is time T-ΔT j The subset corresponding to the i-th submodule in the original response dataset, for j = 1, 2, 3, ..., L is the number of the time step, L is the total number of time steps; i = 1, 2, ..., N is the sequence number of the submodule, δ is the redundant error.

[0021] According to the time T-ΔT j The dynamic constraint function value is calculated based on the response data set. Any time step ΔT1 where the dynamic constraint function value exceeds the redundant error δ is selected as the lower limit, and any time step where the dynamic constraint function value does not exceed the redundant error δ is selected as the upper limit ΔT2. The optimization interval [ΔT1, ΔT2] is established. The average value ΔT3 of the upper and lower limits is taken and its constraint function value E is calculated. r3 , when E r3 >δ, the update optimization interval is [ΔT3, ΔT2]. r3 When ≤δ, the optimization interval is updated to [ΔT1, ΔT3]; the gradient function of the time step is established, and the maximum value is solved in the updated optimization interval as the maximum effective step length ΔT max , whose expression is:

[0022]

[0023] in, A preset positive number.

[0024] Furthermore, the specific method of step 2 is:

[0025] Fusion dataset After k-times data partitioning, k×k subsets are obtained, and the standard deviation of each subset of each data partitioning is calculated; k is selected according to the dimension of the fused dataset, and its expression is:

[0026]

[0027] Construct a standard deviation matrix based on the standard deviation of each subset of each data division Calculate the standard deviation matrix The probability of each standard deviation in , construct the probability matrix The standard deviation matrix With the probability matrix Dot product to get the standard deviation matrix The largest element S in max ; According to the largest element S max Find the corresponding subset before data partitioning right and Perform inversion operations separately to obtain the inversion results of the response information set at the current time T and invalid subsets Subtract the two to get the valid subset

[0028] Calculate the information inefficiency η, its expression is:

[0029]

[0030] Among them, Num I Invalid subset The total number of response data included, Num V Valid subset The total number of response data included.

[0031] Further analysis The number of submodules corresponding to the response data contained in is combined with the information inefficiency η to determine the chassis status, and a decision is made based on the chassis status. The specific method for completing the online determination of the global system status is as follows:

[0032] when When only one submodule’s response data is included, the information inefficiency η is divided into three cases: when the information inefficiency η∈(68%,100%], the status of the submodule corresponding to the invalid subset is analyzed online dynamically; when the information inefficiency η∈(45%,68%] and the duration exceeds 10 control response cycle lengths T ... c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 40 km / h. When the information efficiency η∈[0%,45%], the vehicle's global system is troubleshooted and the vehicle is parked.

[0033] when When only the response data of two submodules are included, the value range of information inefficiency η is divided into three cases: when information inefficiency η∈(64%,100%], The status of the corresponding submodule is analyzed online dynamically; when the information inefficiency η∈(40%,64%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 35 km / h. When the information inefficiency η∈[0%,40%], the vehicle's global system is troubleshooted and the vehicle is stopped.

[0034] when When the response data is contained in more than 2 submodules, the value range of information inefficiency η is divided into three cases: when the information inefficiency η∈(60%,100%], the sampling rate of the response data is increased to 5 times the sampling rate in the previous time period, and the spike signal of the response data is checked. The status of the corresponding submodule is analyzed online dynamically; when the information inefficiency η∈(40%, 60%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 30 km / h within 15 seconds. When the information efficiency η∈[0%,40%], the global system of the vehicle is checked for faults and the vehicle is stopped.

[0035] Beneficial effects:

[0036] 1. The present invention proposes an online determination method for the global system status based on chassis modularization. The method realizes data fusion of multi-module sensors based on dynamic optimization of target tolerance value and time step. The statistical characteristics of the fused information data set are extracted by calculating the standard deviation, valid and invalid data are divided, and inversion calculation is performed to obtain valid original response data and invalid original response data. On this basis, the information inefficiency of the response data of each module can be obtained, and the dynamic determination and online decision-making of the chassis sub-module status and the global performance of the system can be completed.

[0037] 2. Before calculating the tolerance value, considering the problem of inconsistent dimensions of the response data of each submodule at the same time, the present invention first unifies the dimensions of all response data sets to reduce the amount of calculation and improve the calculation speed.

[0038] 3. The present invention comprehensively considers the original response data set at the current time T, the time step variable and the change rate of the original response data set to establish a dynamic constraint function, and finds the maximum effective step size that meets the redundant error constraint and maximizes the fusion speed.

[0039] 4. The present invention adopts the form of random and multiple grouping to construct the standard deviation matrix to better extract the statistical characteristics of the fused data set, pave the way for distinguishing valid data from invalid data, and calculate the information inefficiency based on the distinction results.

[0040] 5. In order to ensure the reliability of the information pool, the present invention needs to extract as much chassis response data as possible in the case of multiple data, multiple modules and mutual coupling of each module, so as to analyze the chassis status more comprehensively. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0042] like Figure 1 As shown, the present invention proposes a global system status online determination method based on chassis modularization, and the specific steps include:

[0043] (1) Create a fusion dataset based on chassis data:

[0044] Step 1.1. Collect the response data of each submodule of the vehicle's chassis system: For the target vehicle chassis modular system, assume that it is composed of N submodules. During the interaction process, information is transmitted to each other through the submodules, and the specific module corresponding to the information is known. At a certain historical moment T-ΔT before the current moment, the chassis controller applies an excitation signal to the cascade submodule of the chassis system, and the sensors of each submodule output response data and transmit it to the cascade controller module. The cascade controller module receives all the response data and divides the corresponding response data into subset 1, subset 2, ..., subset n, ..., subset N according to the submodule number. All subsets constitute the original response data set of the chassis system at the current moment T. in is the original response data of the sensor of the nth submodule, n=1,2,...,N. Figure 1 As shown in the figure, due to the differences in sensor accuracy under different working conditions, in order to minimize the influence of external disturbances, multi-dimensional information fusion calculations should be performed in the cascade controller module for the response information sets of different sub-modules.

[0045] Step 1.2: To facilitate the integration of response data of different dimensions, unify the dimensions of the response data of all submodules: solve the transition matrix of each subset in the original response data set to obtain a response data matrix with unified dimensions. Its expression is:

[0046]

[0047] in, is the subset corresponding to the nth submodule in the response data set at the current time T, M N ×M N In response to the dimension of the dataset, M n1 ×M n2 is the dimension of the original response dataset, They are The pre-transition transfer matrix and post-transition transfer matrix are determined according to the system measurement accuracy under the corresponding time step ΔT.

[0048] Step 1.3: Based on the original response data set at the historical moment T-ΔT, unify its dimensions and calculate the response data change rate matrix Its expression is:

[0049]

[0050] in, is the subset corresponding to the nth submodule in the original response dataset at historical time T-ΔT, for is the first-order differential of , and ΔT is the time step.

[0051] Step 1.4, based on matrix and Calculate the tolerance value of the chassis system at the current moment, and determine whether the tolerance value exceeds the target tolerance value γ. If it exceeds, directly calculate the fusion data set Execute step 2.1, otherwise execute step 1.5. The expression of tolerance value ω is:

[0052]

[0053] in, express The 2-norm of .

[0054] Fusion dataset The expression is:

[0055]

[0056] Step 1.5: In order to find a time step that satisfies both the redundant error δ and the maximum computational speed, and to make the fused data set effective, a dynamic constraint J is established, the time step ΔT is changed, and dynamic optimization is performed to find the maximum effective step ΔT. max , update the historical time to T-ΔT max Return to step 1.3 to continue. The expression of dynamic constraint J is:

[0057]

[0058] Among them, Er is the dynamic constraint function, is the subset corresponding to the i-th submodule in the response data set at the current time T, is the subset corresponding to the i-th submodule in the response data set at time T-ΔT, for The first-order differential of ; i = 1, 2, ..., N is the serial number of the submodule, used to calculate the integral.

[0059] Specifically, a method for dynamically optimizing the time step ΔT is:

[0060] Change the time step ΔT to ΔT j , solve the dynamic constraint function value E at each time step rj , to determine whether it meets the dynamic constraints at the corresponding moment, the expression is:

[0061]

[0062] Among them, J j is the time step ΔT j Dynamic constraints, E rj is the time step ΔT j The corresponding dynamic constraint function, is time T-ΔT j The subset corresponding to the i-th submodule in the original response dataset, for The first-order differential of ; j = 1, 2, 3..., L is the number of time steps, and L is the total number of time steps.

[0063] According to the time T-ΔT j The dynamic constraint function value is calculated based on the response data set. Any time step ΔT1 where the dynamic constraint function value exceeds the redundant error δ is selected as the lower limit, and any time step where the dynamic constraint function value does not exceed the redundant error δ is selected as the upper limit ΔT2. The optimization interval [ΔT1, ΔT2] is established. The average value ΔT3 of the upper and lower limits is taken and its constraint function value E is calculated. r3 , when E r3 >δ, the update optimization interval is [ΔT3, ΔT2]. r3 When ≤δ, the optimization interval is updated to [ΔT1, ΔT3]; the gradient function of the time step is established, and the maximum value is solved in the updated optimization interval as the maximum effective step length ΔT max , whose expression is:

[0064]

[0065] in, A preset positive number.

[0066] (2) If Figure 1 As shown, after completing the fusion calculation of multi-dimensional information, the cascade controller responds to the fusion data set Perform feature extraction:

[0067] Step 2.1: Divide the subsets of the fused dataset into k groups of subsets each time. Calculate the standard deviation of each subset and the standard deviation S of the fused dataset:

[0068] The fused dataset is divided into two categories according to the dimension M. N ×M N Perform the first data division to obtain k subsets, and the dimension of each subset is: The dimension is M N ×M N11 , the second subset The dimension is M N ×M N12 , the kth subset The dimension is M N ×M N1k , where M N =M N11 +M N12 +…+M N1k The τth subset obtained by the first data partition For example, The standard deviation S 1τ The expression is:

[0069]

[0070] Among them, τ is the sequence number of the subset, n 1τ is the number of response data in the τth subset of the first data partitioning process, x 1ξ for The ξth response data in is x 1ξ The mean of .

[0071] Calculate the standard deviation S of all subsets obtained by the first data division 11 ,S 12 ,…,S 1k , perform the same operation as the first data division on all subsets obtained from each data division, obtain the standard deviation of all subsets obtained from each data division, and form the standard deviation matrix of the fused data set To highlight the characteristics of the fused dataset, its expression is:

[0072]

[0073] Among them, S kkis the standard deviation of the kth subset of the kth data partition. k is selected according to the dimension of the fused dataset, and its expression is:

[0074]

[0075] If the calculated value of k is not an integer, then rounding it up can obtain the corresponding specific k value.

[0076] Fusion dataset The expression of the standard deviation S is:

[0077]

[0078] in, To fusion dataset The dataset at row p and column q in To fusion dataset The mean of , p is the number of rows, and q is the number of columns.

[0079] Step 2.2: Based on the standard deviation matrix of the fused data set Calculate the probability matrix Its expression is:

[0080]

[0081] Among them, S Pkk To select S in S kk The probability P(S kk |S).

[0082] Step 2.3: Standard deviation matrix and the probability matrix Perform dot multiplication to find the maximum standard deviation S in the standard deviation matrix max , whose expression is:

[0083]

[0084] Looking for S max The corresponding d-th subset of the c-th data partition Perform inversion operation on this subset to obtain invalid inversion response data set Perform inversion operation on the fused data set to obtain the inversion response data set The effective inversion response data set That is exist The complement of . The expression of the inversion response data set is:

[0085]

[0086] in, is a matrix The pseudo-inverse matrix of is a matrix The pseudo-inverse matrix of .

[0087] Step 2.4: Calculate information inefficiency η, which is expressed as:

[0088]

[0089] Among them, Num I Invalid inversion response dataset The total number of response data included, Num V Valid inversion response dataset The total number of response data included.

[0090] (3) Determine the global performance of the vehicle system and make online decisions: The invalid inversion response dataset at the current time T Put the valid inversion response data set at the current time T into the invalid information pool in chronological order Put them into the valid information pool in chronological order. Figure 1 As shown in Figure 2, the invalid information pool also contains invalid inversion response data sets before the current moment, and the valid information pool also contains valid inversion response data sets before the current moment, which are divided into perception information sets and their corresponding state information sets. Global information network module analyzes invalid subset X TI The number of submodules corresponding to the data is combined with the information inefficiency η to determine the chassis status, and the decision is made based on the chassis status, specifically:

[0091] when When only one submodule’s response data is included, the information inefficiency η is divided into three cases: when the information inefficiency η∈(68%,100%], The status of the corresponding submodule is dynamically analyzed online to determine whether the chassis can achieve the expected target performance indicators 1 to M; when the information inefficiency η∈(45%,68%] and the duration exceeds 10 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 40 km / h. When the information inefficiency η∈[0%,45%], the vehicle's global system is troubleshooted and the vehicle is parked.

[0092] when When only the response data of two submodules are included, the value range of information inefficiency η is divided into three cases: when information inefficiency η∈(64%,100%], The status of the corresponding submodule is dynamically analyzed online to determine whether the chassis can achieve the expected target performance indicators 1 to M; when the information inefficiency η∈(40%, 64%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 35 km / h. When the information inefficiency η∈[0%,40%], the vehicle's global system is troubleshooted and the vehicle is stopped.

[0093] when When the response data is contained in more than 2 submodules, the value range of information inefficiency η is divided into three cases: when the information inefficiency η∈(60%,100%], the sampling rate of the response data is increased to 5 times the sampling rate in the previous time period, and the spike signal of the response data is checked. The status of the corresponding submodule is dynamically analyzed online to determine whether the chassis can achieve the expected target performance indicators 1 to M; when the information inefficiency η∈(40%, 60%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 30 km / h within 15 seconds. When the information efficiency η∈[0%,40%], the global system of the vehicle is checked for faults and the vehicle is stopped.

[0094] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A global system status online determination method based on chassis modularization, characterized in that the steps include: Step 1: Based on the response data sets of the chassis submodule sensors at the current time T and the historical time T-ΔT, calculate the tolerance value of the response data set at the current time T and compare it with the preset target tolerance value. If the tolerance value does not exceed the target tolerance value, calculate the fused data set based on the response data set at the current time and proceed to step 2. When the tolerance value exceeds the target tolerance value, the time step ΔT is dynamically optimized to obtain the maximum effective step ΔT max , update the historical time to T-ΔT max , return to the beginning of step 1 and continue; Step 2: Calculate the standard deviation of the fused dataset and its subsets, and construct a standard deviation matrix; Calculate the probability of each standard deviation in the standard deviation matrix, construct a probability matrix, obtain the maximum standard deviation in the standard deviation matrix and its corresponding subset in the fused data set Will Perform inversion operations as features of the fused dataset to obtain invalid inversion response datasets And calculate the information inefficiency η of the fusion data set; analyze The number of submodules corresponding to the response data contained in is combined with the information inefficiency η to determine the chassis status, and make decisions based on the chassis status to complete the online determination of the global system status; The specific method of calculating the tolerance value of the response data set at the current time T based on the response data sets of the sensors of each submodule of the chassis at the current time T and the historical time T-ΔT is: The response data sets of each sub-module sensor of the chassis at the current time T and the historical time T-ΔT are dimensionally unified to obtain the response data matrix at the current time T And the response data change rate matrix at historical time T-ΔT according to and Calculate the sensor tolerance value ω, the expression is: Where n=1, 2, ..., N is the sequence number of the submodule, and N is the total number of submodules. is the response data matrix of the nth submodule at the current time T, is the response data change rate matrix of the nth submodule at the historical moment T-ΔT; The specific method of calculating the fused dataset based on the response dataset at the current time T is: According to the target tolerance value γ, tolerance value ω, and the response data matrix at the current time T And the response data change rate matrix at historical time T-ΔT Calculate the fused dataset Its expression is: in, is the subset corresponding to the nth submodule in the fusion dataset; The specific method of step 2 is: Fusion dataset After k-times data partitioning, k×k subsets are obtained, and the standard deviation of each subset of each data partitioning is calculated; k is selected according to the dimension of the fused dataset, and its expression is: Construct a standard deviation matrix based on the standard deviation of each subset of each data division Calculate the standard deviation matrix The probability of each standard deviation in , construct the probability matrix The standard deviation matrix With the probability matrix Dot product to get the standard deviation matrix The largest element S in max ; According to the largest element S max Find the corresponding subset before data partitioning right and Perform inversion operations separately to obtain the inversion results of the response information set at the current time T and invalid subsets Subtract the two to get the valid subset Calculate the information inefficiency η, its expression is: Among them, Num I Invalid subset The total number of response data included, Num V Valid subset The total number of response data included.

2. The method according to claim 1, wherein The specific method of dynamically optimizing the time step ΔT is: Change the time step ΔT = ΔT j , ΔT j is the jth time step variable; according to the response data set at the current time T and the time T-ΔT j The response data set establishes its dynamic constraint conditions J j , whose expression is: Among them, E rj is the dynamic constraint function corresponding to the j-th time step, is the subset corresponding to the i-th submodule in the original response dataset at the current time T, is time T-ΔT j The subset corresponding to the i-th submodule in the original response dataset, for The first-order differential of ; j = 1, 2, 3..., L is the number of the time step, L is the total number of time steps; i = 1, 2,..., N is the sequence number of the submodule, δ is the redundant error; According to the time T-ΔT j The dynamic constraint function value is calculated based on the response data set. Any time step ΔT1 where the dynamic constraint function value exceeds the redundant error δ is selected as the lower limit, and any time step where the dynamic constraint function value does not exceed the redundant error δ is selected as the upper limit ΔT2. The optimization interval [ΔT1, ΔT2] is established. The average value ΔT3 of the upper and lower limits is taken and its constraint function value E is calculated. r3 , when E r3 >δ, the update optimization interval is [ΔT3, ΔT2]. r3 When ≤δ, the optimization interval is updated to [ΔT1, ΔT3]; the gradient function of the time step is established, and the maximum value is solved in the updated optimization interval as the maximum effective step length ΔT max , whose expression is: in, A preset positive number.

3. The method according to claim 1, wherein The analysis The number of submodules corresponding to the response data contained in is combined with the information inefficiency η to determine the chassis status, and a decision is made based on the chassis status. The specific method for completing the online determination of the global system status is as follows: when When only one submodule’s response data is included, the information inefficiency η is divided into three cases: when the information inefficiency η∈(68%,100%], the status of the submodule corresponding to the invalid subset is analyzed online dynamically; when the information inefficiency η∈(45%,68%] and the duration exceeds 10 control response cycle lengths T ... c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 40 km / h. When the information inefficiency η∈[0%,45%], the vehicle's global system is troubleshooted and the vehicle is parked. when When only the response data of two submodules are included, the value range of information inefficiency η is divided into three cases: when information inefficiency η∈(64%,100%], The status of the corresponding submodule is analyzed online dynamically; when the information inefficiency η∈(40%,64%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 35 km / h. When the information inefficiency η∈[0%,40%], the vehicle's global system is troubleshooted and the vehicle is stopped. when When the response data is contained in more than 2 submodules, the value range of information inefficiency η is divided into three cases: when the information inefficiency η∈(60%,100%], the sampling rate of the response data is increased to 5 times the sampling rate in the previous time period, and the spike signal of the response data is checked. The status of the corresponding submodule is analyzed online dynamically; when the information inefficiency η∈(40%, 60%] and the duration exceeds 5 control response cycle lengths T c ,right The corresponding submodule starts fault diagnosis and reduces the vehicle speed to below 30 km / h within 15 seconds. When the information efficiency η∈[0%,40%], the global system of the vehicle is checked for faults and the vehicle is stopped.

Images