Brake energy recovery method for electric drive commercial vehicle considering braking intention

Abstract

The application discloses a braking energy recovery method for electrically-driven commercial vehicles considering braking intention, wherein, based on a driver braking intention recognition process of a hidden Markov-interval two-type fuzzy neural network, uncertain information in the braking process of the driver and the braking system can be effectively processed, and the recognition accuracy of the driver braking intention is significantly improved; based on the braking energy recovery process of the driver braking intention, braking torque distribution is performed by fully considering the driver braking intention, and the slow response problem of the friction braking process of the commercial vehicle is made up by utilizing the quick response advantage of the regenerative braking system, so that the braking safety and the braking comfort are fully ensured.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of electric drive commercial vehicles, and particularly relates to an electric drive commercial vehicle braking energy recovery method considering braking intention. BACKGROUND

[0002] As a core component of new energy systems, the electric drive axle of a commercial vehicle is a key technology for promoting the technical upgrading and market development of new energy commercial vehicles. As an effective means to improve the driving range of electric vehicles, the research on the composite braking system of an electric commercial vehicle is of great significance to the improvement of the electric drive axle technology of a commercial vehicle.

[0003] The existing research on the braking energy recovery technology of a commercial vehicle focuses on the coordinated distribution of braking torque. In Chinese Patent Application No. CN201811570181.6, a new energy commercial vehicle braking energy recovery system and control method are proposed, which enables the energy recovery system to be compatible with the ABS system without changing the braking system through a novel structural scheme. In Chinese Patent Application No. CN201910765142.X, an electric commercial vehicle braking energy recovery system and method are proposed, which formulates a braking strategy according to the vehicle state and brake pedal information. However, the two electric commercial vehicle braking energy recovery methods of the above patent applications do not consider the braking intention of the driver and the response difference between the actuators. The vehicle braking process is a dynamic time-varying nonlinear process, and the existing braking intention recognition methods, such as Markov chain and fuzzy theory, are based on the assumptions of fixed state change probability, correct rules, and independent distribution. The above methods have deficiencies in recognition accuracy and real-time performance, and the ability to handle uncertain information in the process of recognizing the braking intention of the driver needs to be improved.

[0004] The commercial vehicle drive-by-wire chassis technology has certain advantages in vehicle handling performance and safety, and the use of regenerative braking systems can effectively improve the endurance of commercial vehicles. However, the braking intention of the driver has a great influence on the braking torque distribution, and how to accurately identify the braking intention of the driver while considering the braking intention of the driver and the working characteristics of the actuators for braking torque distribution has a great improvement space for the braking energy recovery and braking comfort of commercial vehicles. SUMMARY

[0005] To solve the above problems, the main purpose of the present application is to design an electric drive commercial vehicle braking energy recovery method considering braking intention, which solves the problem that the existing electric commercial vehicle braking energy recovery method does not fully consider the difference between the braking intention of the driver and the response of the actuators, resulting in low braking energy recovery efficiency and reduced braking comfort and safety performance.

[0006] To achieve the above purpose, the present application adopts the following technical scheme:

[0007] The application discloses a braking energy recovery method for an electric drive commercial vehicle considering braking intention, which is an energy recovery method for driver braking intention based on a hidden Markov-interval type-2 fuzzy neural network, comprising a braking intention recognition process of the hidden Markov-interval type-2 fuzzy neural network and a braking energy recovery process based on the driver braking intention.

[0008] The method comprises the following steps:

[0009] Step 1: a hidden Markov model of driver braking intention is established, and after optimization of parameters of the hidden Markov model, a maximum likelihood estimation array is obtained based on collected brake pedal data;

[0010] Step 2: an interval type-2 fuzzy neural braking intention recognition network is established based on processing of the maximum likelihood estimation array as input, the interval type-2 fuzzy neural braking intention recognition network outputs braking intention, and input of the interval type-2 fuzzy neural braking intention recognition network is fuzzified;

[0011] Step 3: fuzzy reasoning of the interval type-2 fuzzy neural braking intention recognition network is performed based on fuzzification of the input variable of step 2;

[0012] Step 4: the interval type-2 fuzzy neural braking intention recognition network is defuzzified;

[0013] Step 5: the interval type-2 fuzzy neural braking intention recognition network outputs;

[0014] Step 6: the driver braking intention is recognized based on the hidden Markov-interval type-2 fuzzy neural network, and the driver braking intention is judged, the braking intention comprising slight braking, moderate braking and emergency braking;

[0015] It is judged whether the driver braking intention is slight braking, if yes, step 7 is entered, if not, it is further judged whether the driver braking intention is moderate braking, if yes, step 8 is entered, if not, it is emergency braking, and step 9 is entered;

[0016] Step 7: when the driver braking intention is slight braking, the maximum regenerative braking torque of an electric machine carried by an electric drive axle of the commercial vehicle is calculated, and after the regenerative braking torque and the friction braking torque are obtained, step 10 is entered;

[0017] Step 8: when the driver braking intention is moderate braking, the maximum regenerative braking torque of the electric machine carried by the electric drive axle of the commercial vehicle and the rear wheel braking torque are calculated, and after the regenerative braking torque and the friction braking torque are obtained, step 10 is entered;

[0018] Step 9: When the driver braking intention is emergency braking, the maximum regenerative braking torque and the anti-lock braking torque of the motor carried by the electric drive axle of the commercial vehicle are calculated, and then the regenerative braking torque and the friction braking torque are obtained to enter step 10;

[0019] Step 10: The obtained friction braking torque is compensated by the regenerative braking system;

[0020] Step 11: The vehicle controller controls the regenerative braking system and the friction braking system to implement corresponding braking torques according to the size of the regenerative braking torque and the friction braking torque until the vehicle speed is reduced to zero.

[0021] As a further description of the application, in step 1, establishing the driver braking intention hidden Markov model includes the following steps:

[0022] S11: According to the driver braking operation and the vehicle system state, the input of the driver braking intention hidden Markov model is determined, and the observation sequence vector is:

[0023]

[0024] Wherein, O t is the observation sequence vector of the hidden Markov model, d(t) represents the displacement of the brake pedal, represents the brake pedal speed, and t is the time;

[0025] S12: Based on the pedal displacement d(t) and the brake pedal speed signal Baum-Welch algorithm is used to identify the driver braking intention hidden Markov model parameters, and the driver braking intention hidden Markov model parameters are:

[0026] λ=(π,A,B);

[0027] Wherein, λ is the driver braking intention hidden Markov model, π is the initial probability vector, A is the transition probability matrix, and B is the observation probability density function;

[0028] The expression of the observation probability density function is:

[0029]

[0030] Wherein, N(O,μ ij ,σ ij ) represents the q-dimensional Gaussian density function of state p, ω pq represents the qth mixed function weight coefficient under state p, μ pq represents the vector mean of the qth mixed Gaussian function under state p, and σ mixed function represents the covariance matrix of the mixed function;

[0031] S13: If ε t (p,q) is the probability of the state p observation sequence at t time the qth mixed Gaussian function, the Markov chain probability of state p at t time and state q at t+1 time is:

[0032]

[0033] Wherein, P[O, μ pq ,σ pq ] is the qth dimension Gaussian density function in state p, alpha t (p) is the forward probability of state p, beta t (p) is the backward probability of state p, sigma pq is the covariance matrix of the qth mixed function in state p;

[0034] After optimizing the parameters of the hidden Markov model, based on the collected brake pedal data, the matching degree with the multi-dimensional hidden Markov model is calculated based on the forward-backward algorithm, and the maximum likelihood estimation array is obtained.

[0035] As a further description of the application, in step 2, the process of establishing the interval type-2 fuzzy neural brake intention recognition network includes the following steps:

[0036] S21: Determine the input and output of the interval type-2 fuzzy neural brake intention recognition network;

[0037] The maximum likelihood estimation array is normalized to obtain the Markov model state cumulative probability vector under three brake intentions, and it is taken as the input x i ;

[0038]

[0039] Wherein, i=1, 2, 3; 1 represents slight braking, 2 represents moderate braking, and 3 represents emergency braking; ω0 is the initial weight, w k , sigma k respectively the radial basis function center and width of the kth neuron;

[0040] S22: Fuzzification of interval type-2 fuzzy neural brake intention recognition network input;

[0041] The interval type-2 fuzzy set of input layer neuron x1 is The type-2 fuzzy set of input layer neuron x2 is The type-2 fuzzy set of input layer neuron x3 is

[0042] Wherein, NB represents negative big, NS represents negative small, ZE represents zero, PS represents positive small, and PB represents positive big.

[0043] Therefore, the number of hidden layer neurons is 15, and the activation function of each neuron, i.e. the membership function, is:

[0044]

[0045] wherein, is the upper bound of the output value of the jth membership function layer neuron, u ij is the lower bound of the output value of the jth membership function layer neuron, c i is the lower bound of the center value of the membership function of the jth membership function layer neuron, ij is the upper bound of the center value of the membership function of the jth membership function layer neuron, σ ij is the width of the membership function of the jth membership function layer neuron, j = 1, 2, …, 15.

[0046] As a further description of the present application, in step 3, the fuzzy reasoning of the interval two-type fuzzy neural brake intention recognition network includes the following:

[0047] After the input variables of the interval two-type fuzzy neural brake intention recognition network are fuzzified by the membership functions, the membership interval of the variables is obtained Each node in the excitation layer represents a fuzzy rule, and the activation interval F j (x1, x2, x3) of each rule is expressed as:

[0048]

[0049] wherein, f j is the lower bound of the activation of the jth rule, is the upper bound of the activation of the jth rule.

[0050] As a further description of the present application, in step 4, the fuzzy deflation of the interval two-type fuzzy neural brake intention recognition network includes the following:

[0051] The deflation method adopts the Enhanced Karnic-Mendel (EKM) algorithm, and the output expression of this layer is:

[0052]

[0053] wherein, J is the number of fuzzy rules, y iL is the left end point of the network deflation output, y iR is the right end point of the network deflation output, L is the left turning point of the EKM algorithm, and R is the right turning point of the EKM algorithm.

[0054] ​As a further description of the present invention, in step 5, the output of the interval type II fuzzy neural braking intention recognition network includes the following:

[0055] The fifth layer of the interval type II fuzzy neural braking intent recognition network is the output layer, and its nodes represent the output of the entire network. The network's output expression is:

[0056]

[0057] The neural network is trained using the backpropagation algorithm and gradient descent. The performance metric is defined as follows:

[0058]

[0059] The parameters that need to be learned in the above network include the uncertainty center c of the type II Gaussian membership function. j and Rule consequent parameters w i and w i The formula for parameter correction is:

[0060]

[0061]

[0062] Where, τ w For rule consequent parameters w j and The learning rate, τ c For the center of uncertainty c j and The learning rate.

[0063] As a further description of the present invention, in step 7, the maximum regenerative braking torque T of the motor mounted on the electric drive axle of the commercial vehicle is calculated. r_max It includes the following steps:

[0064] S71: Determine the maximum regenerative braking torque T of the motor r_max Is it greater than the required braking torque T? b_i ;

[0065] If so, then the vehicle braking torque T b for:

[0066] T b =T r ;

[0067] If not, then the vehicle braking torque T b for:

[0068]

[0069] Tf is the front axle braking torque, T b_f Tf is the front axle braking torque, T b_r Tf is the front axle braking torque, T r Tf is the front axle braking torque, T m Tf is the front axle braking torque, T

[0070] S72: get the regenerative braking torque T r Tf is the front axle braking torque, T m go to step 10.

[0071] As a further description of the application, in step 8, the maximum regenerative braking torque T r_max and the rear wheel braking force T b_r of the motor carried by the commercial vehicle electric drive axle, comprising the following steps:

[0072] S81: determine whether the maximum regenerative braking torque T r_max of the motor is greater than the required braking torque T b_r ;

[0073] If yes, the vehicle braking torque T b is:

[0074]

[0075] If no, the vehicle braking torque T b is:

[0076]

[0077] Tf is the front axle braking torque, T b_f Tf is the front axle braking torque, T b_r Tf is the front axle braking torque, T r Tf is the front axle braking torque, T m Tf is the front axle braking torque, T

[0078] S82: get the regenerative braking torque T r Tf is the front axle braking torque, T m go to step 10.

[0079] As a further description of the application, in step 9, the maximum regenerative braking torque T r_max and the anti-lock required braking torque T b_λ of the motor carried by the commercial vehicle electric drive axle, comprising the following steps:

[0080] S91: determine whether the maximum regenerative braking torque T r_max of the motor is greater than the anti-lock required braking torque T b_λ ;

[0081] If yes, the vehicle braking torque Tb is:

[0082] T b = T m + T r ;

[0083] If not, the vehicle braking torque T b is:

[0084] T b = T r ;

[0085] wherein T r is the regenerative braking torque, T m is the friction braking torque;

[0086] S92: obtaining the regenerative braking torque T r and the friction braking torque T m , and then entering step 10.

[0087] As a further description of the present application, in step 10, the regenerative braking compensation torque is expressed as:

[0088]

[0089] wherein T r is the regenerative braking torque, T m is the friction braking torque, is the actual friction braking braking torque.

[0090] Compared with the prior art, the technical effects of the present application are:

[0091] The present application provides a braking energy recovery method for electric drive commercial vehicles considering braking intention, wherein the driver braking intention recognition process based on hidden Markov-interval type-2 fuzzy neural network can effectively process the uncertain information in the braking process of the driver and the braking system, significantly improving the driver braking intention recognition accuracy; the braking energy recovery process based on the driver braking intention fully considers the driver braking intention for braking torque distribution, and simultaneously utilizes the rapid response advantage of the regenerative braking system to make up for the slow response problem of the commercial vehicle friction braking process, fully guaranteeing the braking safety and improving the braking comfort. BRIEF DESCRIPTION OF DRAWINGS

[0092] Fig. 1 is a schematic diagram of the driver braking intention recognition process of the present application based on hidden Markov-interval type-2 fuzzy neural network;

[0093] Fig. 2 is a schematic diagram of the braking energy recovery process flow based on the driver braking intention of the present application. DETAILED DESCRIPTION

[0094] The application will be described in detail below with reference to the accompanying drawings:

[0095] In an embodiment of the application, a brake energy recovery method for electric drive commercial vehicles considering brake intention is disclosed, as shown in the figure, the method is an energy recovery method for driver brake intention based on hidden Markov-interval type-2 fuzzy neural network, including: a brake intention recognition process of hidden Markov-interval type-2 fuzzy neural network, and a brake energy recovery process based on driver brake intention. Figs. 1-2

[0096] The method includes the following steps:

[0097] Step 1: establish a hidden Markov model of driver brake intention, and after optimizing the parameters of the hidden Markov model, obtain a maximum likelihood estimation array based on the collected brake pedal data;

[0098] Step 2: after processing based on the maximum likelihood estimation array as input, establish an interval type-2 fuzzy neural brake intention recognition network, the output of the interval type-2 fuzzy neural brake intention recognition network is brake intention, and the input of the interval type-2 fuzzy neural brake intention recognition network is fuzzified;

[0099] Step 3: based on the fuzzification of the input variable of step 2, perform fuzzy reasoning of the interval type-2 fuzzy neural brake intention recognition network;

[0100] Step 4: defuzzification of the interval type-2 fuzzy neural brake intention recognition network;

[0101] Step 5: output of the interval type-2 fuzzy neural brake intention recognition network;

[0102] Step 6: identify the driver brake intention based on the hidden Markov-interval type-2 fuzzy neural network, and judge the driver brake intention, the brake intention including slight braking, moderate braking, and emergency braking;

[0103] Judge whether the driver brake intention is slight braking, if yes, go to step 7, if not, further judge whether the driver brake intention is moderate braking, if yes, go to step 8, if not, it is emergency braking, go to step 9;

[0104] Step 7: when the driver brake intention is slight braking, calculate the maximum regenerative braking torque of the electric motor carried by the electric drive axle of the commercial vehicle, and go to step 10 after obtaining the regenerative braking torque and the friction braking torque;

[0105] ​Step 8: When the driver's braking intention is moderate braking, the maximum regenerative braking torque of the motor carried by the electric drive axle of the commercial vehicle and the rear wheel braking torque are calculated, and then the regenerative braking torque and the friction braking torque are obtained to enter step 10;

[0106] Step 9: When the driver's braking intention is emergency braking, the maximum regenerative braking torque of the motor carried by the electric drive axle of the commercial vehicle and the anti-lock braking demand torque are calculated, and then the regenerative braking torque and the friction braking torque are obtained to enter step 10;

[0107] Step 10: The obtained friction braking torque is compensated by the regenerative braking system;

[0108] Step 11: The vehicle controller controls the regenerative braking system and the friction braking system to implement corresponding braking torque according to the size of the regenerative braking torque and the friction braking torque until the vehicle speed is reduced to zero.

[0109] More specifically, the present embodiment details the above steps as follows:

[0110] In the above step 1, the driver's braking intention hidden Markov model is established, including the following steps:

[0111] S11: According to the driver's braking operation and the vehicle system state, the input of the driver's braking intention hidden Markov model is determined, and the observation sequence vector is:

[0112]

[0113] Wherein, O t is the observation sequence vector of the hidden Markov model, d(t) represents the displacement of the brake pedal, represents the brake pedal speed, and t is the time;

[0114] S12: Based on the pedal displacement d(t) and the brake pedal speed signal Baum-Welch algorithm is used to identify the driver's braking intention hidden Markov model parameters, and the driver's braking intention hidden Markov model parameters are:

[0115] λ=(π,A,B);

[0116] Wherein, λ is the driver's braking intention hidden Markov model, π is the initial probability vector, A is the transition probability matrix, and B is the observation probability density function;

[0117] The expression of the observation probability density function is:

[0118]

[0119] Wherein, N(O,μ) ij ,σ ij Let ω denote the q-dimensional Gaussian density function of state p. pq μ represents the weight coefficient of the q-th mixture function in state p. pq Let σ represent the vector mean of the q-th Gaussian mixture function under state p, and let σ be the covariance matrix of the mixture function.

[0120] S13: If ε t Let (p,q) be the probability of the q-th mixture Gaussian function of the observation sequence at time t for state p. Then the Markov chain probabilities of state p at time t and state q at time t+1 are:

[0121]

[0122] Wherein, P[O,μ pq ,σ pq Let α be the Gaussian density function of dimension q in state p. t (p) represents the forward probability of state p, β t (p) represents the backward probability of state p, σ pq Let be the covariance matrix of the q-th mixture function in state p;

[0123] After optimizing the parameters of the hidden Markov model, the maximum likelihood estimation array is obtained by calculating the matching degree with the multidimensional hidden Markov model based on the collected brake pedal data using a forward-backward algorithm.

[0124] In this embodiment, step 2 above, the process of establishing the interval type II fuzzy neural braking intention recognition network, includes the following steps:

[0125] S21: Determine the input and output of the interval type II fuzzy neural braking intention recognition network;

[0126] The maximum likelihood estimation array is normalized to obtain the cumulative probability vector of the Markov model state under three braking intentions, and this vector is used as the input x. i ;

[0127]

[0128] Where i = 1, 2, 3; 1 represents light braking, 2 represents moderate braking, and 3 represents emergency braking; ω0 is the initial weight, w k , σ k These are the center and width of the radial basis function of the k-th neuron, respectively;

[0129] S22: Input fuzzification for the interval type II fuzzy neural braking intention recognition network;

[0130] The interval type-2 fuzzy set of the input layer neuron x1 is: Type II fuzzy set of input layer neurons x2 Type II fuzzy set of input layer neurons x3

[0131] Where NB represents negative large, NS represents negative small, ZE represents zero, PS represents positive small, and PB represents positive large; therefore, the number of neurons in the hidden layer is 15, and the activation function, i.e., the membership function, of each neuron is:

[0132]

[0133] in, u is the upper bound of the output value of the neuron in the j-th membership function layer. ij (x i ) is the lower bound of the output value of the neuron in the j-th membership function layer, c ij This is the lower bound of the center value of the membership function of the neuron in the j-th membership function layer. σ is the upper bound of the center value of the membership function of the neuron in the j-th membership function layer. ij Let be the width of the membership function of the neuron in the j-th membership function layer, where j = 1, 2, ..., 15.

[0134] In this embodiment, step 3 above includes the following fuzzy inference of the interval type II fuzzy neural braking intention recognition network:

[0135] The input variables of the interval-based type II fuzzy neural braking intent recognition network are fuzzified using a membership function to obtain the membership intervals of the variables. Each node in the activation layer represents a fuzzy rule, and the activation range F of each rule is... j (x1,x2,x3), the expression is:

[0136]

[0137] Among them, f j Let j be the lower bound of the activation degree of the j-th rule. Let be the upper bound of the activation degree of the j-th rule.

[0138] In this embodiment, step 4 above, the fuzzy downscaling of the interval type-2 fuzzy neural braking intent recognition network includes the following:

[0139] The reduction method employs the Enhanced Karnic-Mendel (EKM) algorithm, and the output expression of this layer is:

[0140]

[0141] Where J is the number of fuzzy rules, y iL The left endpoint of the network's downscaling output is y.iR L is the left turning point of the EKM algorithm, and R is the right turning point of the EKM algorithm.

[0142] In the step 5, the interval type-2 fuzzy neural braking intention recognition network outputs include the following:

[0143] The fifth layer of the interval type-2 fuzzy neural braking intention recognition network is an output layer, and the nodes thereof represent the output of the entire network. The output expression of the network is as follows:

[0144]

[0145] The error back propagation algorithm is used to train and learn the neural network by using the gradient descent method. The performance index is defined as follows:

[0146]

[0147] The parameters to be learned in the network include the uncertain center c of the type-2 Gaussian membership function j and the rule consequent parameter w i and The parameter correction formula is expressed as follows:

[0148]

[0149]

[0150] wherein τ w is the learning rate of the rule consequent parameter w j and τ c is the learning rate of the uncertain center c j and

[0151] Since there are left and right turning points in the EKM reduction algorithm, the parameters should be adjusted according to the turning points in the process of function derivation. The adjustment formula (wherein the uncertain center c of the type-2 Gaussian membership function j and The parameter update formula is only listed for the first neuron, and the parameter update formula for the remaining neurons is similar) is as follows:

[0152]

[0153] ​Traditional one-type fuzzy logic has poor ability to resist external interference. In the process of driver braking intention recognition, interval two-type fuzzy neural network is adopted to play the ability of two-type fuzzy system to process uncertain information and the anti-interference ability. By enhancing the description of inter-individual and intra-individual uncertainty, the anti-interference ability of fuzzy logic algorithm is enhanced, and the recognition accuracy of driver braking intention is improved.

[0154] In the above step 6, the driver braking intention is different, and the different steps are entered.

[0155] Specifically, in the above step 7, when the braking intention is slight braking, the maximum regenerative braking torque T r_max of the motor carried by the electric drive axle of the commercial vehicle is calculated.

[0156] S71: Determine whether the maximum regenerative braking torque T r_max of the motor is greater than the required braking torque T b_i .

[0157] If yes, the vehicle braking torque T b is:

[0158] T b =T r .

[0159] If no, the vehicle braking torque T b is:

[0160]

[0161] Wherein, T b_f is the front axle braking torque, T b_r is the rear axle braking torque, T r is the regenerative braking torque, and T m is the friction braking torque.

[0162] S72: After obtaining the regenerative braking torque T r and the friction braking torque T m , enter step 10.

[0163] In the above step 8, when the braking intention is moderate braking, the maximum regenerative braking torque T r_max of the motor carried by the electric drive axle of the commercial vehicle and the rear wheel braking force T b_r are calculated, including the following steps:

[0164] S81: Determine whether the maximum regenerative braking torque T r_max of the motor is greater than the required braking torque T b_r .

[0165] If yes, the vehicle braking torque T b is:

[0166]

[0167] If not, the vehicle braking torque T b is:

[0168]

[0169] wherein T b_f is the front axle braking torque, T b_r is the rear axle braking torque, T r is the regenerative braking torque, and T m is the friction braking torque;

[0170] S82: obtaining the regenerative braking torque T r and the friction braking torque T m and then entering step 10.

[0171] In this embodiment, in step 9, when the braking intention is emergency braking, the maximum regenerative braking torque T r_max of the motor of the electric drive axle of the commercial vehicle is obtained, and the anti-lock demand braking torque T b_λ is obtained, including the following steps:

[0172] S91: judging whether the maximum regenerative braking torque T r_max of the motor is greater than the anti-lock demand braking torque T b_λ ;

[0173] If yes, the vehicle braking torque T b is:

[0174] T b = T m + T r ;

[0175] If not, the vehicle braking torque T b is:

[0176] T b = T r ;

[0177] wherein T r is the regenerative braking torque, and T m is the friction braking torque;

[0178] S92: obtaining the regenerative braking torque T r and the friction braking torque T m and then entering step 10.

[0179] In this embodiment, in step 10, the obtained friction braking torque is compensated by using the rapid response characteristic of the regenerative braking system, and the regenerative braking compensation torque is expressed as:

[0180]

[0181] wherein, T r is the regenerative braking torque, T m is the friction braking torque, is the actual friction braking braking torque.

[0182] Through the above embodiments, the technical solutions of the present application are disclosed, and the present application has the following advantages over the prior art:

[0183] 1. The driver braking intention recognition method based on the hidden Markov-interval second-type fuzzy neural network of the present application can effectively process uncertain information in the braking process of the driver and the braking system, and significantly improve the recognition accuracy of the driver braking intention.

[0184] 2. The electric drive commercial vehicle braking energy recovery method based on the driver braking intention of the present application fully considers the driver braking intention for braking torque distribution, and simultaneously utilizes the rapid response advantage of the regenerative braking system to make up for the slow response problem of the friction braking process of the commercial vehicle, thereby fully guaranteeing the braking safety and improving the braking comfort.

[0185] The above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application, and other modifications or equivalent replacements to the technical solutions of the present application made by those skilled in the art should be covered in the scope of the claims of the present application as long as they do not deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for brake energy recovery of an electrically driven commercial vehicle considering braking intention, characterized in that: The method is an energy recovery method for driver braking intention based on a hidden Markov-interval type-2 fuzzy neural network, comprising: a braking intention recognition process of the hidden Markov-interval type-2 fuzzy neural network, and a braking energy recovery process based on the driver braking intention; The method comprises the following steps: Step 1: a hidden Markov model of driver braking intention is established, and after the parameters of the hidden Markov model are optimized, a maximum likelihood estimation array is obtained based on collected brake pedal data; Step 2: an interval type-2 fuzzy neural braking intention recognition network is established based on the maximum likelihood estimation array after processing as input, the interval type-2 fuzzy neural braking intention recognition network outputs a braking intention, and the interval type-2 fuzzy neural braking intention recognition network input is fuzzified; Step 3: based on the fuzzification of the input variable of step 2, the interval type-2 fuzzy neural braking intention recognition network is subjected to fuzzy reasoning; Step 4: the interval type-2 fuzzy neural braking intention recognition network is subjected to fuzzy defuzzification; Step 5: the interval type-2 fuzzy neural braking intention recognition network is outputted; Step 6: the driver braking intention is recognized based on the hidden Markov-interval type-2 fuzzy neural network, and the driver braking intention is judged, the braking intention including slight braking, moderate braking and emergency braking; whether the driver braking intention is slight braking is judged, if yes, step 7 is entered, if not, whether the driver braking intention is moderate braking is further judged, if yes, step 8 is entered, if not, it is emergency braking, and step 9 is entered; Step 7: when the driver braking intention is slight braking, the maximum regenerative braking torque of the motor carried by the electric drive axle of the commercial vehicle is calculated, and after the regenerative braking torque and the friction braking torque are obtained, step 10 is entered; Step 8: when the driver braking intention is moderate braking, the maximum regenerative braking torque of the motor carried by the electric drive axle of the commercial vehicle and the rear wheel braking torque are calculated, and after the regenerative braking torque and the friction braking torque are obtained, step 10 is entered; Step 9: when the driver braking intention is emergency braking, the maximum regenerative braking torque of the motor carried by the electric drive axle of the commercial vehicle and the anti-lock braking torque are calculated, and after the regenerative braking torque and the friction braking torque are obtained, step 10 is entered; Step 10: the obtained friction braking torque is compensated by the regenerative braking system; Step 11: the vehicle controller controls the regenerative braking system and the friction braking system to implement corresponding braking torque according to the size of the regenerative braking torque and the friction braking torque until the vehicle speed is reduced to zero.

2. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 1, the establishment of the hidden Markov model of the driver braking intention comprises the following steps: S11: according to the driver braking operation and the vehicle system state, the input of the hidden Markov model of the driver braking intention is determined, and the observation sequence vector is: wherein O t is a hidden Markov model observation sequence vector, d(t) represents brake pedal displacement, represents brake pedal velocity, and t is time; S12: based on the pedal displacement d(t) and the brake pedal speed signal v(t) collected by the brake pedal sensor The Baum-Welch algorithm is used to identify the parameters of the driver braking intention hidden Markov model, and the parameters of the driver braking intention hidden Markov model are: λ=(π,A,B); Wherein, λ is the hidden Markov model of the driver braking intention, π is the initial probability vector, A is the transition probability matrix, and B is the observation probability density function; The expression of the observation probability density function is: Wherein, N(O,μ) ij ,σ ij Let ω denote the q-dimensional Gaussian density function of state p. pq μ represents the weight coefficient of the q-th mixture function in state p. pq Let σ represent the vector mean of the q-th Gaussian mixture function under state p, and let σ be the covariance matrix of the mixture function. S13: If ε t (p, q) is the probability of the state p observation sequence at the t time the qth mixed Gaussian function, and the Markov chain probability of state p at t time and state q at t+1 time is: where P[O, μ pq ,σ pq ] is the qth dimensional Gaussian density function at state p, α t (p) is the forward probability of state p, β t (p) is the backward probability of state p, and σ pq is the covariance matrix of the qth mixture function at state p. After the optimization of the parameters of the hidden Markov model, based on the collected brake pedal data, the matching degree of the multi-dimensional hidden Markov model is calculated based on the forward-backward algorithm to obtain a maximum likelihood estimation array.

3. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 2, the process of establishing the interval type-2 fuzzy neural braking intention recognition network includes the following steps: S21: Determine the input and output of the interval type-2 fuzzy neural braking intention recognition network; The maximum likelihood estimation array is normalized to obtain Markov model state cumulative probability vectors under three braking intentions, and the Markov model state cumulative probability vectors are taken as inputs x i ; where i = 1, 2, 3; 1 represents slight braking, 2 represents moderate braking, and 3 represents emergency braking; ω0is an initial weight, w k , σ k are a radial basis function center and width of the kth neuron, respectively. S22: Fuzzification of the input of the interval type-2 fuzzy neural braking intention recognition network; The interval-valued intuitionistic fuzzy set of input layer neuron x1 is The interval-valued intuitionistic fuzzy set of input layer neuron x2 is The interval-valued intuitionistic fuzzy set of input layer neuron x3 is NB represents negative big, NS represents negative small, ZE represents zero, PS represents positive small, and PB represents positive big; Therefore, the number of hidden layer neurons is 15, and the activation function of each neuron, i.e., the membership function, is: wherein, is an upper bound of the output value of the jth membership function layer neuron, u ij (x i ) is a lower bound of the output value of the jth membership function layer neuron, c ij is a lower bound of the center value of the membership function of the jth membership function layer neuron, is an upper bound of the center value of the membership function of the jth membership function layer neuron, σ ij is a width of the membership function of the jth membership function layer neuron, j = 1, 2, …, 15.

4. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 3, characterized in that: In step 3, the fuzzy reasoning of the interval type-2 fuzzy neural braking intention recognition network includes the following: The input variable of the interval type fuzzy neural braking intention recognition network is subjected to fuzzy processing by a membership function to obtain a membership interval of the variable Each node in the excitation layer represents a fuzzy rule, and each rule has an activation interval F j (x1,x2,x3), and the expression is: wherein f j is the lower bound of the activation level of the jth rule, is the upper bound of the activation level of the jth rule.

5. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 4, the defuzzification of the interval type-2 fuzzy neural braking intention recognition network includes the following: The defuzzification method adopts the Enhanced Karnic-Mendel (EKM) algorithm, and the output expression of this layer is: where J is the number of fuzzy rules, y iL is the left end point of the network down-type output, y iR is the right end point of the network down-type output, L is the left turning point of the EKM algorithm, and R is the right turning point of the EKM algorithm.

6. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 5, characterized in that: In step 5, the output of the interval type-2 fuzzy neural braking intention recognition network includes the following: The fifth layer of the interval type-2 fuzzy neural braking intention recognition network is the output layer, and its nodes represent the output of the entire network. The output expression of the network is: The error backpropagation algorithm is used to train and learn the neural network using the gradient descent method, and the performance index is defined as: The parameters to be learned in the above network include the uncertain center c of the two-type Gaussian membership function j and The rule consequent parameter w i and The parameter correction formula expression is: where τ w is a rule consequent parameter w j and is a learning rate for the uncertain center c c and c j and c j .

7. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 7, the maximum regenerative braking torque T of the motor mounted on the electric drive axle of the commercial vehicle is calculated r_max comprising the following steps: S71: Determine the maximum regenerative braking torque T of the motor r_max whether it is greater than the demand braking torque T b_i ; If yes, the vehicle braking torque T b is: T b = T r ; If not, the vehicle braking torque T b is: where T b_f is the front axle braking torque, T b_r is the rear axle braking torque, T r is the regenerative braking torque, T m is the friction braking torque; S72: Obtain regenerative braking torque T r with friction braking torque T m The step 10 is entered.

8. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 8, the maximum regenerative braking torque T of the motor mounted on the electric drive axle of the commercial vehicle is calculated r_max and the rear wheel braking force T b_r , comprising the following steps: S81: Determine the maximum regenerative braking torque T of the motor r_max whether it is greater than the demand braking torque T b_r ; If yes, the vehicle braking torque T b is: If not, the vehicle braking torque T b is: wherein T b_f is the front axle braking torque, T b_r is the rear axle braking torque, T r is the regenerative braking torque, T m is the friction braking torque; S82: Obtain regenerative braking torque T r with friction braking torque T m The step 10 is entered.

9. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 9, the maximum regenerative braking torque T of the motor mounted on the electric drive axle of the commercial vehicle r_max and the anti-lock demand braking torque T b_λ , comprising the following steps: S91 : determine the maximum regenerative braking torque T of the electric motor r_max whether greater than the anti-lock braking torque demand T b_λ ; If yes, the vehicle braking torque T b is: T b = T m + T r ; If not, the vehicle braking torque T b is: T b = T r ; where T r is the regenerative braking torque, T m is the friction braking torque; S92: Obtain regenerative braking torque T r with friction braking torque T m The post enters step 10.

10. The electric drive commercial vehicle brake energy recovery method considering brake intention according to claim 1, characterized in that: In step 10, the regenerative braking compensation torque is expressed as: where T r is the regenerative braking torque, T m is the friction braking torque, is the actual friction braking braking torque.

Images