An intelligent longitudinal-lateral predictive control system and method applied to an electric fork truck

Abstract

The application discloses an intelligent longitudinal and lateral predictive control system applied to an electric forklift, which comprises a longitudinal driving intention analysis and prediction subsystem, a lateral driving intention analysis and prediction subsystem, a traction motor speed predictive control system and an oil pump motor speed predictive control system. The longitudinal driving intention analysis and prediction subsystem is used for analyzing and predicting a longitudinal driving intention and obtaining an optimal longitudinal driving intention sequence. The lateral driving intention analysis and prediction subsystem is used for analyzing and predicting a lateral driving intention and obtaining an optimal lateral driving intention sequence. The traction motor speed predictive control system is used for controlling the speed of a traction motor. The oil pump motor speed predictive control system is used for controlling the speed of an oil pump motor. The application can predict and adjust parameters such as the speed of the traction motor and the speed of the oil pump motor in the manner of predicting the longitudinal and lateral driving intentions, so as to ensure the stability and smoothness of the vehicle in the longitudinal and lateral control, avoid longitudinal instability, jerk, poor micro-motion caused by the untimely response of the traction motor, and avoid steering instability, oil pump abnormal sound and other problems caused by unreasonable control of the lateral oil pump motor.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of motion control of electric forklifts, and in particular to an intelligent lateral and longitudinal predictive control system and method applied to electric forklifts. BACKGROUND

[0002] With the increasing global awareness of environmental protection and the requirement of reducing carbon emissions, electric forklifts have become an ideal alternative to traditional internal combustion forklifts due to their zero emissions and low noise characteristics, as well as lower maintenance costs and energy consumption. The application scenarios of electric forklifts have expanded from traditional warehouse handling to more industries, such as ports, automobiles, and steel industries. Technological advancements in batteries, motors, and electronic control systems have made electric forklifts not only more reliable and durable, but also more energy-efficient, enhancing the competitiveness of the products.

[0003] Although advanced technologies and control systems have been adopted in the design and manufacturing process of electric forklifts to optimize their operational performance, there may still be some challenges and issues in actual application. For example, in the longitudinal control and lateral control of electric forklifts, in the longitudinal control of electric forklifts, due to the inability of the vehicle's actions to match the driver's intentions in a timely manner, 1) acceleration and deceleration control is not smooth: if the accelerator pedal response is too fast or too slow, it may cause sudden acceleration changes when the forklift starts or stops, affecting driving comfort and possibly causing damage to the goods; 2) brake response lag: the brake system responds slowly or the braking force does not match the driver's expectations, which may result in excessive stopping distance in emergency situations, increasing the risk of accidents.

[0004] In the lateral control of electric forklifts, due to the inability of the vehicle's lateral actions to match the driver's intentions in a timely manner, 1) steering response inconsistency: when the driver operates the steering, the actual steering angle of the electric forklift does not match the driver's expectations in real time, which may cause the electric forklift to move laterally unevenly; 2) poor lateral stability: if the electric forklift carries high or unbalanced goods, lateral movement may increase the risk of rollover. If the design of the forklift does not fully consider lateral stability, the driver's operation intentions may not be safely realized; 3) unclear steering force feedback: on manually controlled electric forklifts, if the force feedback provided by the steering mechanism is not obvious or does not conform to the conventional driving habits, the driver may feel confused, which may affect the accuracy of the operation. Different drivers have different driving habits, and if the control system is too rigid and cannot adapt to individual needs, some drivers may feel uncomfortable, which may affect operational efficiency and safety. SUMMARY

[0005] In view of the problems in the prior art, the application provides a crusher lubricating oil replacement prompting alarm system and a working method thereof, so as to solve the problems of non-smooth acceleration, deceleration, starting and braking and non-smooth steering in longitudinal control and transverse control of an electric forklift.

[0006] In order to achieve the above object, the application adopts the technical scheme of an intelligent longitudinal and transverse predictive control system applied to an electric forklift, comprising:

[0007] A longitudinal driving intention analysis and prediction subsystem is used for analyzing and predicting a longitudinal driving intention and obtaining an optimal longitudinal driving intention sequence;

[0008] A transverse driving intention analysis and prediction subsystem is used for analyzing and predicting a transverse driving intention and obtaining an optimal transverse driving intention sequence;

[0009] A traction motor speed predictive control system is used for linear matching lookup of a traction motor and a longitudinal driving intention, and the traction motor is controlled in speed during the transition of each intention by smoothing the output speed of the traction motor; and an oil pump motor speed predictive control system is used for linear matching lookup of an oil pump motor and a transverse driving intention, and the oil pump motor is controlled in speed during the transition of each intention by smoothing the output speed of the oil pump motor.

[0010] Further, the longitudinal driving intention analysis and prediction subsystem comprises a longitudinal driving intention data processing, a longitudinal driver intention clustering analysis and a longitudinal motion parameter adaptive adjustment module.

[0011] Further, the transverse driving intention analysis and prediction subsystem comprises a transverse driving intention data processing, a hidden Markov model (HMM) building, an HMM model evaluation and a driver transverse driving intention prediction module.

[0012] Further, when the vehicle performs an action, the longitudinal driving intention analysis and prediction subsystem needs to process the detected longitudinal driving related data, and the processed data needs to be subjected to clustering analysis and longitudinal driving intention prediction; the traction motor speed predictive control system adjusts the acceleration gradient, the brake rate and the traction motor speed according to the monitored longitudinal driving intention data, so as to realize data predictive control processing of longitudinal driving; the transverse driving intention analysis and prediction subsystem processes the transverse action related data generated by the vehicle, and the processed data is subjected to HMM model building and optimal transverse driving intention sequence prediction of the next moment; and the oil pump motor speed predictive control system adjusts the speed of the oil pump motor in advance according to the predicted transverse driving intention sequence, so as to realize speed control of the oil pump motor.

[0013] An intelligent longitudinal and transverse predictive control method applied to an electric forklift, comprising a longitudinal predictive control method and a transverse predictive control method,

[0014] Longitudinal predictive control method:

[0015] When the driver controls the electric forklift movement, first, longitudinal driving intention analysis and prediction are performed: by collecting longitudinal driving intention data of the electric forklift in real time, characteristic data set is obtained after preliminary filtering processing, longitudinal driving intention data is processed by using wavelet function conversion and wavelet soft threshold denoising method, then K-means clustering analysis method is used to analyze and predict the longitudinal driving intention characteristic data, after clustering prediction, a first-order Markov model is built to predict the state, the longitudinal driving intention prediction data obtained by the first-order Markov model prediction analysis is input into the traction motor speed prediction control system, longitudinal motion parameter adaptive adjustment is performed, and the speed of the traction motor is predicted to ensure the stability and smoothness of the vehicle longitudinal driving;

[0016] Lateral predictive control method:

[0017] After collecting and processing the lateral driving intention data of the electric forklift, an HMM model is built, a parameter set of the lateral driving intention is constructed, input into the HMM for model solving, the lateral driving intention prediction of the next moment is realized, the driving intention prediction set is obtained, the Baum-Welch algorithm is called to obtain the optimal lateral driving intention prediction sequence; the lateral driving intention prediction sequence is output to the oil pump motor speed prediction control system, the oil pump motor speed prediction control system realizes the adaptive adjustment of the oil pump motor speed of the electric forklift in the lateral control according to the prediction sequence of the lateral driving intention, and the lateral control smoothness of the electric forklift is ensured.

[0018] Further, the longitudinal predictive control method is as follows:

[0019] First, collect longitudinal driving intention data, mainly including: accelerator pedal opening, vehicle speed, brake pedal opening data, after preliminary filtering processing of the data, obtain characteristic data set Data L1 ; use discrete wavelet function to process the data in Data L1 in time-frequency domain signal transformation, the discrete wavelet function is shown in formula 1 and formula 2:

[0020] φ(t)=∑ k g(k)φ(2t-k) (1)

[0021] ψ(t)=∑ k h(k)φ(2t-k) (2)

[0022] In the formula: g(k), g(k) are filter coefficients; ψ(t) is a wavelet function; φ(t) is a scale transformation function; k≥0 represents the translation of the function in the time domain;

[0023] The high-frequency component and the low-frequency component of each driving data are obtained by wavelet decomposition and reconstruction of the longitudinal driving intention data dataset DataL1 respectively, and the dataset Data L2 Since the low-frequency component data is close to the driving intention, and the high-frequency component data is mutation data, it is necessary to perform wavelet soft threshold denoising processing on the wavelet processed data Data L2 , that is, for each wavelet coefficient, if its absolute value exceeds a preset threshold, the coefficient is retained but its amplitude is reduced to the threshold value; if the absolute value is less than or equal to the threshold, the coefficient is set to zero, so as to remove the smaller coefficients caused by noise by the soft threshold processing method, as shown in formula 3:

[0024]

[0025] c' = sign(c).max(|c|-T,0) (4) In the formula, c is a wavelet coefficient; T is a threshold; sign(c) is a sign function of the wavelet coefficient c; max(|c|-T,0) is the larger one of |c|-T and 0; when |c| is greater than the threshold T, the output is |c|-T multiplied by its sign; c' is a new wavelet coefficient;

[0026] The threshold T can balance the relationship between the denoising effect and the signal feature retention. The threshold T is determined by using the empirical formula of Donoho and Johnstone, as shown in formula 5:

[0027]

[0028] In the formula, σ is the noise standard deviation; N is the signal length;

[0029] After processing by the wavelet soft threshold denoising method, Data L3 is obtained; the noise processed Data L3 dataset is processed by using the K-means clustering analysis method to cluster five types of driving intention feature data: speed keeping, vehicle acceleration, deceleration, and parking. First, five initial center points are set according to the five types of driving intention data: speed keeping, vehicle acceleration, deceleration, and parking, and each data point is assigned to the center point with the minimum Euclidean distance from the center point, as shown in formula 6:

[0030]

[0031] In the formula, d is the Euclidean distance; x i is the i-th data point; c i is the position of the i-th center point;

[0032] According to the average value of the coordinates of all data points in the cluster of the new center point, the position of the center point of each cluster is updated, as shown in formula (7):

[0033]

[0034] In the formula: C j is the set of all data points assigned to the jth center point; |C j | is the number of elements of the set |C j According to the optimization objective of the K-means algorithm, the sum of the squared errors within the cluster, the algorithm is updated iteratively according to the minimization of the sum of the squared errors within the cluster, as shown in formula (8):

[0035]

[0036] In the formula: J is the total intra-cluster error; k is the number of clusters; C j is the set of all data points assigned to the jth center point;

[0037] After real-time intention clustering analysis of longitudinal driving intention data, a first-order Markov model is built to predict the state of longitudinal driving intention, and the next time longitudinal driving intention feature data after state prediction by the first-order Markov model is transmitted to the traction motor speed prediction control system for subsequent traction motor speed prediction control.

[0038] Further, a first-order Markov model is built to predict the state of longitudinal driving intention: first, a Markov state set S is constructed, as shown in formula (9):

[0039] S={s2 s2...s k}(9)

[0040] In the formula: s1, s2, s k is the longitudinal driving intention data obtained by real-time clustering analysis,

[0041] Substitute the longitudinal driving intention data into the state transition matrix, as shown in formula (10):

[0042] A=[ξ ij ](10)

[0043] In the formula: A is an N×N state transition probability matrix, ξ ij is the probability of the longitudinal driving intention state s i to state s j ; according to the state prediction formula, the longitudinal driving intention state data is predicted, and the state prediction is as shown in formula (11):

[0044] p(t)=p(t-1)·A(11)

[0045] wherein p(t), p(t-1) are 1 x N dimensional row vectors representing the probability distribution at time t, t-1.

[0046] Further, the longitudinal motion parameter adaptive adjustment mainly includes: acceleration gradient adjustment, brake rate adjustment, to realize the smooth processing of the rotational speed of the traction motor involved in the longitudinal driving.

[0047] Further, the lateral prediction control method is specifically as follows:

[0048] First, collect the lateral driving intention data, mainly including: steering angle, vehicle speed and other data, and obtain the feature data set DataT1 after data filtering processing; construct the HMM model by inputting the HMM parameter set, the parameter set including: a limited set of driving intentions, a driving intention feature sequence, a driving intention transition probability, and a driving intention initial distribution probability, according to the set emergency steering and general steering, wherein the emergency steering is set as the steering wheel angle speed greater than β1° / s when the speed is α1km / h, and the general steering is set as the steering wheel angle speed greater than β2° / s when the speed is α2km / h; wherein the emergency steering and the general steering are set values; the steering behavior of the driver is divided into two actions, and the emergency steering action and the general steering action are taken as two hidden states of the lateral driving model, to construct the steering angle observation value set Cr={Cr1, Cr2,..., Crn}, and the optimal driving intention sequence is obtained by combining the Baum-Welch algorithm, and the corresponding parameter model of the HMM model is obtained by the vector n . As shown in formula (12):

[0049]

[0050] wherein Q=[Q1, Q2] is the driving intention set, Cr is the steering angle observation value set, α ij is the transition probability from driving intention i to driving intention j, wherein i, j ∈ [1, n], b j (k) is the probability of observing driving operation k at driving intention j, P is the distribution probability of the initial driving intention of the driver, and the driving intention transition probability is as shown in formula (13):

[0051]

[0052] wherein α t (i) is the probability of observing the operation before time t at driving intention i, β t (i) is the probability of observing the operation after time t at driving intention i; b j (Cr t+1 ) is the probability of observing the driving operation Cr t+1 at driving intention j

[0053] The estimated probability of the shift of driving intention can be obtained according to the Baum-Welch algorithm re-estimation formula, as shown in formulas (14), (15) and (16):

[0054]

[0055] P i = χ1(i) (16)

[0056] is the estimated probability of the shift of driving intention from i to j, is the estimated probability of observing driving operation k at time j when the driving intention is i, i is the estimated probability of the initial driving intention distribution, δ(o t , v k ) is the optimal driving intention operation sequence; χ t (i) is the probability of observing the complete driving operation sequence at time t when the driving intention is i;

[0057] The initial parameter vector of the HMM model is estimated by using the maximum likelihood estimation function, as shown in formula 17:

[0058]

[0059] In the formula, X is the observation sequence; L(θ; X) is the likelihood function; P(x i |θ) is the probability mass function; and θ is the target parameter;

[0060] The initial parameter vector of the HMM model is obtained After that, the initial parameter vector of the HMM model is obtained , and according to the observation value set Cr of the cornering experiment data, the parameter vector of the HMM model is estimated as , and the state transition probability is combined to realize the prediction of the driving intention at the next time, so as to obtain the driving intention prediction set Q f ; the lateral driving intention set Q f is predicted, and after the steering wheel turning angle value is predicted according to Q f , the turning angle prediction sequence Cr f ={Cr 1f , Cr 2f ,..., Cr nf} is obtained, the turning angle prediction sequence is transmitted to the oil pump motor speed prediction control system, the oil pump motor speed is predicted according to the turning angle prediction sequence, and thus the smooth steering control in the steering process is realized.

[0061] The beneficial effects of this invention are: by predicting and adjusting parameters such as the traction motor speed and oil pump motor speed by predicting the lateral and longitudinal driving intentions, the stability and smoothness of the vehicle in lateral and longitudinal control can be ensured, avoiding longitudinal instability, jerking, poor micro-motion caused by untimely response of the traction motor, and uneven steering and abnormal oil pump noise caused by unreasonable control of the lateral oil pump motor. Attached Figure Description

[0062] Figure 1 This is a control principle diagram of the present invention;

[0063] Figure 2 Here is the overall flowchart of the longitudinal driving intent analysis and prediction subsystem;

[0064] Figure 3 This is the overall flowchart of the lateral driving intention analysis and prediction subsystem. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the scope of the invention.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0067] like Figure 1 As shown, an intelligent longitudinal and lateral predictive control system for electric forklifts consists of four parts: a longitudinal driving intention analysis and prediction subsystem, a lateral driving intention analysis and prediction subsystem, a traction motor speed prediction control system, and an oil pump motor speed prediction control system.

[0068] When the vehicle is in motion, the longitudinal driving intention analysis and prediction subsystem processes the detected longitudinal driving-related data, performs cluster analysis on the processed data, and predicts longitudinal driving intentions. The traction motor speed prediction and control system adjusts the acceleration gradient, braking rate, and traction motor speed based on the monitored longitudinal driving intention data to achieve longitudinal driving data prediction and control processing. The lateral driving intention analysis and prediction subsystem processes the lateral motion-related data generated by the vehicle, builds an HMM model, and predicts the optimal lateral driving intention sequence for the next moment. The oil pump motor speed prediction and control system adjusts the oil pump motor speed in advance based on the predicted lateral driving intention sequence to achieve oil pump motor speed control.

[0069] The longitudinal driving intent analysis and prediction subsystem includes modules such as longitudinal driving intent data processing, longitudinal driver intent clustering analysis, and adaptive adjustment of longitudinal motion parameters.

[0070] like Figure 2 As shown, the longitudinal predictive control method is as follows:

[0071] First, longitudinal driving intention data is collected, mainly including accelerator pedal opening, vehicle speed, and brake pedal opening. After preliminary filtering, the feature dataset is obtained. L1 Using discrete wavelet functions to analyze data L1 The data in the table are subjected to signal transformation in the time-frequency domain using discrete wavelet functions, as shown in Equations 1 and 2:

[0072] φ(t)=∑ k g(k)φ(2t-k) (1)

[0073] ψ(t)=∑ k h(k)φ(2t-k) (2)

[0074] In the formula: g(k) and g(k) are filter coefficients; ψ(t) is the wavelet function; φ(t) is the scaling transformation function; k≥0 indicates the translation of the function in the time domain.

[0075] By performing wavelet decomposition and reconstruction on the longitudinal driving intention data dataset DataL1, the high-frequency and low-frequency components of each driving data were initially obtained, resulting in the dataset DataL1. L2 Since the low-frequency component data closely approximates driving intentions, while the high-frequency component data represents abrupt changes, wavelet processing of the data is necessary. L2 Wavelet soft thresholding denoising is performed: for each wavelet coefficient, if its absolute value exceeds a preset threshold, the coefficient is retained but its amplitude is reduced to the threshold value; if the absolute value is less than or equal to the threshold, the coefficient is set to zero, thus ensuring that smaller coefficients caused by noise are removed by soft thresholding, as shown in Formula 3.

[0076]

[0077] c' = sign(c).max(|c|-T, 0) (4)

[0078] In the formula: c is the wavelet coefficient; T is the threshold; sign(c) is the sign function of the wavelet coefficient c; max(|c|-T,0) is the larger of |c|-T and 0; when |c| is greater than the threshold T, the output is |c|-T multiplied by its sign; c' is the new wavelet coefficient.

[0079] The threshold T can balance the relationship between denoising effect and preservation of signal features. The selection of the threshold T is determined using the empirical formula of Donoho and Johnstone, as shown in Formula 5:

[0080]

[0081] In the formula: σ is the noise standard deviation; N is the signal length;

[0082] Data is obtained after processing with wavelet soft thresholding denoising method. L3 K-means clustering analysis was used to analyze the noise-processed data. L3 The dataset is processed and clustered into five categories of driving intention features: speed maintenance, vehicle acceleration, deceleration, and stopping. First, based on these five categories of driving intention data, five initial center points are set. The point with the smallest Euclidean distance from each data point to a center point is then assigned to that center point, as shown in Formula 6.

[0083]

[0084] In the formula: d is the Euclidean distance; x i It is the i-th data point; c i It is the position of the i-th center point.

[0085] Based on the average coordinates of all data points within the cluster, the center point position of each cluster is updated and calculated, as shown in formula (7):

[0086]

[0087] In the formula: C j It is the set of all data points assigned to the j-th center point; |C j | is a set| C j | The number of elements;

[0088] Based on the optimization objective of the K-means algorithm, which is the sum of squared errors within the cluster, the algorithm is updated iteratively by minimizing the sum of squared errors within the cluster, as shown in formula (8):

[0089]

[0090] In the formula: J is the total intra-cluster error; k is the number of clusters; C j It is the set of all data points assigned to the j-th center point;

[0091] After performing real-time intent clustering analysis on longitudinal driving intent data, a first-order Markov model is built to predict the state of longitudinal driving intent.

[0092] First, the Markov state set S is constructed as shown in equation (9):

[0093] S = {s2 s2... s k} (9)

[0094] In the formula: s1, s2, s k is the longitudinal driving intention data obtained by real-time clustering analysis,

[0095] Substitute the longitudinal driving intention data into the state transition matrix as shown in equation (10):

[0096] A = [ξ ij ] (10)

[0097] In the formula: A is an N*N state transition probability matrix, ξ ij is the probability of the longitudinal driving intention state s i to state s j ;

[0098] According to the state prediction formula, the longitudinal driving intention state data is predicted, and the state prediction is shown in equation (11):

[0099] p(t) = p(t-1) · A (11) In the formula: p(t), p(t-1) is a 1*N dimensional row vector, representing the probability distribution at time t and t-1;

[0100] The next time longitudinal driving intention feature data after state prediction by the first-order Markov model is transmitted to the traction motor speed prediction control system for subsequent traction motor speed prediction control. The traction motor prediction control system monitoring module monitors the longitudinal driving intention sequence and the first-order Markov state prediction sequence obtained by clustering for adaptive adjustment of longitudinal motion parameters, mainly including acceleration gradient adjustment and brake rate adjustment, to realize smooth processing of the speed of the traction motor involved in longitudinal driving.

[0101] The lateral driving intention analysis and prediction subsystem includes: lateral driving intention data processing, Hidden Markov Model (HMM) building, HMM model evaluation, and driver lateral driving intention prediction modules.

[0102] As shown in Figure 3 , the lateral prediction control method is as follows:

[0103] First, the lateral driving intention data is collected, mainly including: steering angle, vehicle speed and other data, and the characteristic data set DataT1 is obtained after data filtering processing; the HMM model is constructed by inputting the parameter set, including: the limited set of driving intention, the sequence of driving intention characteristics, the driving intention transition probability, and the driving intention initial distribution probability; according to the set emergency steering and general steering, wherein the emergency steering is set as the steering wheel angle speed greater than β1° / s when the speed is α1km / h, and the general steering is set as the steering wheel angle speed greater than β2° / s when the speed is α2km / h; wherein the emergency steering and the general steering are set values; the steering behavior of the driver is divided into two actions, and the emergency steering action and the general steering action are taken as two hidden states of the lateral driving model, and the steering angle observation value set Cr={Cr1, Cr2,..., Crn} is constructed, and the optimal driving intention sequence is obtained by combining the Baum-Welch algorithm, and the corresponding parameter model of the HMM model is obtained by the vector n} is constructed, and the optimal driving intention sequence is obtained by combining the Baum-Welch algorithm, and the corresponding parameter model of the HMM model is obtained by the vector As shown in formula (12):

[0104]

[0105] In the formula: Q=[Q1, Q2] is the driving intention set, Cr is the steering angle observation value set, α ij is the transition probability from driving intention i to driving intention j, wherein i, j ∈ [1, n], b j (k) is the probability of observing driving operation k at driving intention j, P is the distribution probability of the initial driving intention of the driver, and the driving intention transition probability is as shown in formula (13):

[0106]

[0107] In the formula: α t (i) is the probability of observing operation before time t at driving intention i, β t (i) is the probability of observing operation after time t at driving intention i; b j (Cr t+1 ) is the probability of observing driving operation Cr t+1 at driving intention j

[0108] According to the re-estimation formula of the Baum-Welch algorithm, the estimated probability of driving intention transition can be obtained, as shown in formula (14), (15) and (16):

[0109]

[0110]

[0111] P i =χ1(i) (16)

[0112] the estimated probability that the driving intention is transferred from i to j, the estimated probability that the driving operation k is observed at time j, P i the estimated probability of the distribution of initial driving intention, δ(o t ,v k is the optimal driving intention operation sequence; χ t (i) is the probability of observing the complete driving operation sequence at time t when the driving intention is i;

[0113] The initial parameter vector of the HMM model is estimated using the maximum likelihood estimation function, as shown in Equation 17:

[0114]

[0115] wherein: X is the observation sequence; L(θ; X) is the likelihood function; P(x i |θ) is the probability mass function; and θ is the target parameter;

[0116] The initial parameter vector of the HMM model is obtained After that, according to the initial parameter vector of the HMM model and the observation set Cr of the steering angle experiment data, the parameter vector of the HMM model is estimated as Combined with the state transition probability the driving intention prediction set Q f at the next time is obtained.

[0117] The lateral driving intention set Q f is predicted, and according to the steering wheel angle value predicted by Q f , the steering angle prediction sequence Cr f ={Cr 1f ,Cr 2f ,...,Cr nf} is obtained, which is transmitted to the oil pump motor speed prediction control system, and the oil pump motor speed is predicted according to the steering angle prediction sequence, so as to realize smooth steering control during steering.

[0118] The above only describes the preferred embodiments of the present application and should not be used to limit the present application, and any modification, equivalent replacement or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An intelligent longitudinal and lateral predictive control system applied to an electric forklift truck, comprising: a longitudinal driving intention analysis and prediction subsystem for analyzing and predicting longitudinal driving intentions, and predicting an optimal longitudinal driving intention sequence; a lateral driving intention analysis and prediction subsystem for analyzing and predicting lateral driving intentions, and predicting an optimal lateral driving intention sequence; a traction motor speed predictive control system for linear matching of the traction motor with the longitudinal driving intention, and for speed control of the traction motor by smoothing the output speed of the traction motor during the transition between each intention; an oil pump motor speed predictive control system for linear matching of the oil pump motor with the lateral driving intention, and for speed control of the oil pump motor by smoothing the output speed of the oil pump motor during the transition between each intention; the longitudinal driving intention analysis and prediction subsystem comprising longitudinal driving intention data processing, longitudinal driver intention clustering analysis, and longitudinal motion parameter adaptive adjustment modules; the lateral driving intention analysis and prediction subsystem comprising lateral driving intention data processing, hidden Markov model (HMM) building, HMM model evaluation, and driver lateral driving intention prediction modules; when the vehicle is in motion, the longitudinal driving intention analysis and prediction subsystem processes detected longitudinal driving-related data, and the processed data is subjected to clustering analysis and longitudinal driving intention prediction; the traction motor speed predictive control system adjusts the acceleration gradient, brake rate, and traction motor speed based on the monitored longitudinal driving intention data, to achieve data predictive control processing for longitudinal driving; the lateral driving intention analysis and prediction subsystem processes lateral motion-related data generated by the vehicle, and the processed data is subjected to HMM model building and prediction of the optimal lateral driving intention sequence for the next time; the oil pump motor speed predictive control system adjusts the speed of the oil pump motor in advance based on the predicted lateral driving intention sequence, to achieve speed control of the oil pump motor.

2. The method of claim 1, applied to an intelligent lateral-longitudinal predictive control system for an electric forklift truck, characterized in that: comprising: a longitudinal predictive control method and a lateral predictive control method, the longitudinal predictive control method: when the driver controls the motion of the electric forklift truck, first perform longitudinal driving intention analysis and prediction: by real-time acquisition of longitudinal driving intention data of the electric forklift truck, obtain a feature data set after preliminary filtering processing, use wavelet function conversion and wavelet soft threshold denoising method to perform signal transformation and data processing on the longitudinal driving intention data, then use K-means clustering analysis method to perform clustering analysis and prediction on the longitudinal driving intention feature data, after clustering prediction, build a first-order Markov model for state prediction, input the longitudinal driving intention prediction data obtained by first-order Markov model prediction analysis into the traction motor speed predictive control system, perform longitudinal motion parameter adaptive adjustment, and perform predictive control on the speed of the traction motor to ensure the stability and smoothness of the longitudinal driving of the vehicle; the lateral predictive control method: After collecting and processing the lateral driving intention data of the electric forklift, the HMM model is built, the parameter set of the lateral driving intention is constructed, and the model is solved by inputting it into the HMM to realize the prediction of the lateral driving intention at the next moment, obtain the driving intention prediction set, call the Baum-Welch algorithm to obtain the optimal lateral driving intention prediction sequence; the lateral driving intention prediction sequence is output to the oil pump motor speed prediction control system, and the oil pump motor speed prediction control system realizes the adaptive adjustment of the oil pump motor speed of the electric forklift during lateral control according to the prediction sequence of the lateral driving intention, and ensures the smoothness of the lateral control of the electric forklift.

3. The method of claim 2, wherein the intelligent lateral and longitudinal predictive control system applied to the electric forklift is characterized in that: The longitudinal predictive control method is as follows: First, the longitudinal driving intention data is collected, including accelerator pedal opening, vehicle speed, brake pedal opening data, and after preliminary filtering processing, the characteristic data set Data L1 is obtained L1 ; the data in Data L1 is respectively transformed in time-frequency domain by using discrete wavelet function, and the discrete wavelet function is shown in formula (1) and formula (2): (1) (2) wherein: , are filter coefficients; is a wavelet function; is a scaling function; k > 0 denotes a translation of the function in time domain; The high-frequency component and the low-frequency component of each driving data are obtained by wavelet decomposition and reconstruction of the longitudinal driving intention data dataset DataL1, respectively, to obtain the dataset Data L2 Since the low-frequency component data is close to the driving intention and the high-frequency component data is mutation data, it is necessary to perform wavelet soft threshold denoising processing on the wavelet processed data Data L2 If the absolute value of each wavelet coefficient exceeds a preset threshold, the coefficient is retained but its amplitude is reduced to the threshold value; if the absolute value is less than or equal to the threshold value, the coefficient is set to zero, thereby ensuring that the smaller coefficients caused by noise are removed by the soft threshold processing method, as shown in formulas (3) and (4): (3) (4) where: is a wavelet coefficient; is a threshold value; is a wavelet coefficient is a sign function; is and the larger of 0; when is greater than the threshold value the output is and multiplied by its sign; is a new wavelet coefficient; Threshold The relationship between the balance of denoising effect and the retention of signal characteristics, the selection of threshold The empirical formula of Donoho and Johnstone is used to determine the threshold, as shown in formula (5): (5) In the formulae: is the noise standard deviation; is the signal length; Data L3 ; K-means clustering analysis method for noise processing Data L3 data set are processed and clustered into five categories of driving intention feature data: speed maintenance, vehicle acceleration, deceleration, and parking. First, set five initial center points according to the five categories of driving intention data: speed maintenance, vehicle acceleration, deceleration, and parking. Assign each data point to the center point with the smallest Euclidean distance, as shown in equation (6): (6) wherein: is the Euclidean distance; is the i-th data point; is the j-th data point; is the i-th center point; is the j-th center point. According to the average value of the coordinates of all data points in the cluster of the new center point, the center point position of each cluster is updated and calculated, as shown in formula (7): (7) wherein: is the set of all data points assigned to the th center point; is the number of elements of the set . According to the optimization objective of the K-means algorithm, the sum of the squared errors in the cluster is updated and iterated according to the minimization of the sum of the squared errors in the cluster, as shown in formula (8): (8) wherein: is the total intra-cluster error; is the number of clusters; is the set of all data points assigned to the th center point. After real-time intention clustering analysis of the longitudinal driving intention data, a first-order Markov model is built to predict the state of the longitudinal driving intention, and the next moment longitudinal driving intention feature data after state prediction by the first-order Markov model is transmitted to the traction motor speed prediction control system for subsequent traction motor speed prediction control.

4. The method of claim 3, wherein the intelligent lateral and longitudinal predictive control system applied to the electric forklift is characterized in that: A first-order Markov model is built to predict the longitudinal driving intention state: First, the Markov state set is constructed As shown in equation (9): (9) In the formula: , , is longitudinal driving intention data obtained by real-time clustering analysis, The longitudinal driving intention data is substituted into the state transition matrix, as shown in formula (10): (10) In the formula: is a state transition probability matrix, is a longitudinal driving intention state to state the probability; According to the state prediction formula, the longitudinal driving intention state data is predicted, and the state prediction is as shown in formula (11): (11) where: , is a row vector of dimension V, representing , the probability distribution at time instant t.

5. The method of claim 2, applied to an intelligent lateral-longitudinal predictive control system for an electric forklift truck, characterized by: The longitudinal motion parameter adaptive adjustment includes acceleration gradient adjustment and brake rate adjustment to realize the speed smoothness processing of the traction motor involved in the longitudinal driving.

6. The method of claim 2, wherein the intelligent lateral and longitudinal predictive control system applied to the electric forklift is characterized in that: The lateral predictive control method is as follows: First, the lateral driving intention data includes: steering angle, vehicle speed data, and the characteristic data set DataT1 is obtained after data filtering processing; the HMM model is constructed by inputting the parameter set, the parameter set includes: the limited set of driving intention, the characteristic sequence of driving intention, the driving intention transition probability, and the driving intention initial distribution probability; according to the set emergency steering and general steering, wherein the emergency steering is set as the steering wheel angular velocity greater than β1° / s when the speed is α1 km / h, and the general steering is set as the steering wheel angular velocity greater than β2° / s when the speed is α2 km / h; wherein the emergency steering and the general steering are set values; the steering behavior of the driver is divided into two actions, and the emergency steering action and the general steering action are taken as two hidden states of the lateral driving model to construct the steering angle observation value set The optimal driving intention sequence is obtained by combining the Baum-Welch algorithm, and the corresponding parameter model of the HMM model is obtained by the vector As shown in formula (12): (12) wherein: is a set of driving intents, is a set of turn angle observations, is a driving intent to driving intent transition probabilities, wherein , is a probability of observed driving operation when driving intent is is a distribution probability of initial driving intent of the driver, and the driving intent transition probability is shown in equation (13): (13) wherein: is the probability of observing the time instant before the driving intention is the probability of observing the time instant after the driving intention is the probability of observing the time instant after the driving intention is the probability of observing the driving operation before the driving intention before the driving intention According to the Baum-Welch algorithm re-estimation formula, the estimated probability of driving intention transition is obtained, as shown in formulas (14), (15) and (16): (14) (15) (16) is the estimated probability of the driving intention being transferred to the driving intention is the estimated probability of the driving intention being at the time instant is the estimated probability of the driving operation at the time instant is the distribution estimate probability of the initial driving intention is the optimal driving intention operation sequence; is the probability of observing the complete driving operation sequence at the time instant when the driving intention is Using the maximum likelihood estimation function, the initial parameter vector of the HMM model is estimated, as shown in formula (17): (17) where: is the observation sequence; is the likelihood function; is the probability mass function; is the target parameter; Obtaining an initial parameter vector of an HMM model Then, according to the initial parameter vector of the HMM model , and according to the observation set of the cornering experiment data , the parameter vector of the HMM model can be estimated , combined with the state transition probability to realize the driving intention prediction of the next moment and obtain a driving intention prediction set ; A set of lateral driving intentions are predicted , according to After predicting the steering wheel angle value, a predicted sequence of angle values is obtained The predicted sequence of angles is passed to the oil pump motor speed prediction control system, and the oil pump motor speed is predicted according to the predicted sequence of angle values, thereby achieving smooth steering control during steering.

Images