Intelligent guide transport vehicle tire life cycle management system

Abstract

The invention discloses an intelligent guide transport vehicle tire life cycle management system. A tire selection module is used for selecting a tire; a time / mileage selection module is used for selecting the time / mileage needing to predict the pattern depth; the pattern depth prediction module calls a prediction model to predict the pattern depth of the selected tire in the selected time / mileage; the curve generation module generates a historical wear curve according to the historical pattern data of the selected tire, and also generates a predicted wear curve according to the pattern depth predicted by the pattern depth prediction module; and the curve display module displays the historical wear curve and the predicted wear curve generated by the curve generation module. The pattern depth of the selected tire in the selected time / mileage can be predicted, the historical wear curve and the predicted wear curve can be generated and displayed, manual direct measurement is not needed, the labor cost and the detection difficulty are reduced, the tire life cycle is convenient to manage, and the detection efficiency is improved. The technical problems of high cost and high difficulty of manual detection of the tire pattern depth in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of tire management analysis, and in particular, relates to a tire life cycle management system for an intelligent guided vehicle. BACKGROUND

[0002] Intelligent guided vehicles (IGV, Inteligent Guided Vehicle) are the main horizontal transportation equipment in an automated terminal and play an important role in connecting automated rail cranes and automated quays. For a fleet of dozens or even hundreds of intelligent guided vehicles in an automated terminal, managing a large number of tires is a very complex task. Scientific tire replacement cycle planning and the establishment of a database for different tire brands and processes are very important. A scientific tire life cycle analysis management system can save a large amount of labor costs and help to develop a more controllable vehicle scheduling plan, thereby improving the container turnover rate of the automated terminal, and the establishment of a tire database can also provide a reference for scientific tire procurement decisions.

[0003] However, scientific management system research in this field is still in a blank state, and the existing management method mainly relies on periodic inspection of vehicle tire pattern depth and proposes a tire replacement plan when the pattern depth is below a certain value. This method lacks planning and usually can only start pattern inspection at the beginning of the terminal operation window, and it is difficult to establish a tire database to scientifically develop the time for batch entry inspection.

[0004] The current intelligent guided vehicle tire life cycle management has not yet formed a more refined management scheme, and mainly faces the following technical problems:

[0005] (1) Window period inspection: vehicle tire pattern depth is inspected during the operation window period, and a tire replacement plan is proposed when the tire pattern depth is below a certain value. This method lacks planning and it is difficult to establish a tire database to scientifically develop the time for batch entry inspection.

[0006] (2) Difference in wear curve: due to the difference in wear curve between driven and driving axles, different brands of tires will also differ in wear curve. In the terminal, tire master-slave exchange may occur, and the existing simple tire curve recording method is difficult to cope with refined management.

[0007] (3) Limitation of data collection amount: in order to ensure that the terminal production does not stop and consider the labor cost, it is not possible to frequently measure the tire pattern depth data, and it is even difficult to ensure that all equipment is measured regularly every month. Therefore, it is not possible to use machine learning and other methods that rely on large amounts of data to train prediction models.

[0008] (4) Taylor series fitting problem: using Taylor series fitting method, too much rely on local approximation in low order, fitting effect is poor, high order is prone to overfitting, easy to be disturbed by noise. SUMMARY

[0009] The application provides a smart guided transport vehicle tire life cycle management system, which solves the technical problems of high cost and great difficulty in manual detection of tire pattern depth in the prior art.

[0010] To achieve the above technical purpose, the application adopts the following technical scheme:

[0011] The smart guided transport vehicle tire life cycle management system comprises:

[0012] A storage module for storing historical pattern depth of the tire;

[0013] A tire selection module for selecting a tire;

[0014] A time / mileage selection module for selecting a time / mileage for which the pattern depth needs to be predicted;

[0015] A pattern depth prediction module for calling a prediction model to predict the pattern depth of the tire selected by the tire selection module at the time / mileage selected by the time / mileage selection module;

[0016] A curve generation module for generating a historical wear curve according to the historical pattern data of the tire selected by the tire selection module, and for generating a predicted wear curve according to the pattern depth predicted by the pattern depth prediction module;

[0017] A curve display module for displaying the historical wear curve and the predicted wear curve generated by the curve generation module.

[0018] In some embodiments of the application, the tire selection module comprises:

[0019] A vehicle number selection box for selecting a vehicle number;

[0020] A plurality of tire selection buttons; each tire selection button corresponds to a tire at a different position of the vehicle selected by the vehicle number selection box, and each tire selection button displays the position of the corresponding tire in the selected vehicle and the current pattern depth;

[0021] The pattern depth prediction module, in response to a prediction instruction, calls a prediction model to predict the pattern depth of the tire selected by the tire selection button at the time / mileage selected by the time / mileage selection module, and sends it to the curve generation module;

[0022] The curve generation module generates a historical wear curve according to historical tread data of the tire corresponding to the tire selection button when receiving a click instruction of any tire selection button; and generates a predicted wear curve when receiving a predicted tread depth sent by the tread depth prediction module.

[0023] In some embodiments of the present application, the positional relationship of the plurality of tire selection buttons is consistent with the actual positional relationship of the corresponding tires in the vehicle.

[0024] In some embodiments of the present application, the color of the tire selection button is determined according to the current tread depth of the corresponding tire.

[0025] In some embodiments of the present application, the color of the tire selection button is determined according to the current tread depth of the corresponding tire, specifically comprising:

[0026] The color of the tire selection button is determined according to the following function:

[0027] color(d) = (R(d), G(d), B(d));

[0028] Wherein, d is the current tread depth of the tire, in millimeters;

[0029]

[0030] In some embodiments of the present application, the tire selection module further comprises a number selection box for selecting a tire number.

[0031] The tread depth prediction module, in response to a prediction instruction, calls a prediction model to predict the tread depth of the tire selected by the number selection box at the time / mileage selected by the time / mileage selection module, and sends it to the curve generation module.

[0032] The curve generation module generates a historical wear curve according to the historical tread data of the tire selected by the number selection box; and generates a predicted wear curve when receiving a predicted tread depth sent by the tread depth prediction module.

[0033] In some embodiments of the present application, the intelligent guided transport vehicle tire life cycle management system further comprises:

[0034] A single vehicle tire all selection box for selecting all tires of a single vehicle;

[0035] The curve generation module is further configured to generate an average wear curve of all tires of the vehicle according to the historical tread data of all tires of the selected single vehicle.

[0036] In some embodiments of the present application, the intelligent guided transport vehicle tire life cycle management system further comprises:

[0037] vehicle multi-select box, for selecting multiple vehicles;

[0038] tire selection box, for selecting positions of tires;

[0039] The curve generation module is further configured to generate an average wear curve of the tires at the selected positions of the selected multiple vehicles.

[0040] In some embodiments of the present application, the prediction model is established using a Taylor series expansion, and the least squares method is used for parameter updating.

[0041] In some embodiments of the present application, the prediction model is established using a Taylor series expansion, and the least squares method is used for parameter updating, specifically including:

[0042] (11) Establish a prediction model:

[0043]

[0044] wherein h(t n ) is the tire pattern depth at time point t n ; t0 is the expansion point of time; is a parameter;

[0045] Let h(t n ) = φ(t n ) T θ n ;

[0046] wherein the feature vector is a parameter vector

[0047] (12) Use the least squares method to update the parameters to minimize the prediction error:

[0048] i∈[1,n], i takes values from 1 to n, and the following calculations are performed respectively;

[0049] Calculate the gain matrix:

[0050] Calculate the prediction error: e i = h(t i ) - φ(t i ) T θ i-1 ;

[0051] Update the parameter vector: θ i = θ i-1 + K i e i ;

[0052] Update the covariance matrix:

[0053] where h(t1), h(t2), …, h(t n ) are known historical tread data;

[0054] λ∈(0,1] is a forgetting factor;

[0055] (13) Use the prediction model to predict the tire tread depth h(t n+k ) at time point t n+k ; where k≥1;

[0056] Calculate the gain matrix:

[0057] Calculate the prediction error: e n+k =e n+k-1 +e + ;

[0058] where e + is a random number in the range (-0.01, +0.01);

[0059] Update the parameter vector: θ n+k =θ n+k-1 +K n+k e n+k ;

[0060] Update the covariance matrix:

[0061] Calculate the tire tread depth: h(t n+k )=φ(t n+k ) T θ n+k ;

[0062] Or,

[0063] (21) Establish a prediction model:

[0064]

[0065] where h(s n ) is the tire tread depth at mileage point s n ; s0 is the expansion point of the mileage; is a parameter;

[0066] Let:

[0067] where the characteristic vector is the parameter vector

[0068] (22) Use the least squares method to update the parameters to minimize the prediction error:

[0069] i∈[1,n], where the value of i starts from 1 and goes up to n, and the following calculations are performed respectively;

[0070] Calculate the gain matrix:

[0071] Calculate the prediction error: e i =h(s i )-φ(s i ) T θ i-1 ;

[0072] Update parameter vector: θ i =θ i-1 +K i e i ;

[0073] Update the covariance matrix:

[0074] Among them, h(s1), h(s2), ..., h(s) n () represents known historical pattern data;

[0075] λ∈(0,1] is the forgetting factor;

[0076] (23) Using a prediction model, predict the mileage point s. n+k Tire tread depth h(s) n+k ); where k≥1;

[0077] Calculate the gain matrix:

[0078] Calculate the prediction error: e n+k =e n+k-1 +e + ;

[0079] Among them, e + A random number within the range (-0.01, +0.01);

[0080] Update parameter vector: θ n+k =θ n+k-1 +K n+k e n+k ;

[0081] Update the covariance matrix:

[0082] Calculate tire tread depth: h(s) n+k )=φ(s n+k ) Tθ n+k .

[0083] Compared with the prior art, the advantages and positive effects of the present application are that: the intelligent guided transport vehicle tire life cycle management system of the present application uses a tire selection module to select a tire that needs to predict the pattern depth or a tire that needs to generate a historical wear curve; uses a time / mileage selection module to select a time / mileage that needs to predict the pattern depth; a pattern depth prediction module calls a prediction model to predict the pattern depth of the selected tire at the selected time / mileage; a curve generation module generates a historical wear curve according to the historical pattern data of the selected tire, and also generates a predicted wear curve according to the pattern depth predicted by the pattern depth prediction module; and a curve display module displays the historical wear curve and the predicted wear curve generated by the curve generation module. Therefore, the intelligent guided transport vehicle tire life cycle management system of the present application can predict the pattern depth of the selected tire at the selected time / mileage, generate and display the historical wear curve and the predicted wear curve, without the need for manual direct measurement, thereby reducing the labor cost and detection difficulty, facilitating the management of the tire life cycle, and solving the technical problems of high cost and great difficulty in manual detection of the tire pattern depth in the prior art.

[0084] Other features and advantages of the present application will become more apparent after reading the detailed description of the embodiments of the present application in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0085] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0086] Figure 1 is a structural block diagram of an embodiment of the intelligent guided transport vehicle tire life cycle management system proposed by the present application; Figure 2 is a user interface diagram of an embodiment of the intelligent guided transport vehicle tire life cycle management system proposed by the present application; Figure 3 is a schematic diagram of a tire selection button; Figure 4 is a schematic diagram of a month search scroll bar; Figure 5 is a schematic diagram of a vehicle number selection box; Figure 6 is a schematic diagram of a historical wear curve and a predicted wear curve; Figure 7 is a schematic diagram of RLS updating a prediction model and an RLS parameter updating process; Figure 8 is a schematic diagram of a month and a pattern depth wear curve; Figure 9 is a schematic diagram of a mileage and a pattern depth wear curve. DETAILED DESCRIPTION

[0087] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort are within the scope of the present application.

[0088] In the description of the present application, it should be understood that the terms "center", "upper", "lower", "front", "back", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0089] In the description of the present application, it should be understood that the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0090] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connection" should be understood in a broad sense, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the communication inside two elements. For a person of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0091] The tire life cycle management system of the intelligent guided vehicle of the embodiment comprises a storage module, a tire selection module, a time / mileage selection module, a pattern depth prediction module, a curve generation module, a curve generation module, etc., as shown in Figure 1 .

[0092] The storage module is used for storing the historical pattern depth of the tire;

[0093] The tire selection module is used for selecting the tire;

[0094] The time / mileage selection module is used for selecting the time / mileage for which the pattern depth needs to be predicted;

[0095] a pattern depth prediction module configured to invoke a prediction model to predict the pattern depth of the tire selected by the tire selection module at the time / mileage selected by the time / mileage selection module;

[0096] a curve generation module configured to generate a historical wear curve according to the historical pattern data of the tire selected by the tire selection module, and to generate a predicted wear curve according to the pattern depth predicted by the pattern depth prediction module;

[0097] a curve display module configured to display the historical wear curve and the predicted wear curve generated by the curve generation module.

[0098] The intelligent guided transport vehicle tire life cycle management system of the embodiment can predict the pattern depth of the selected tire at the selected time / mileage, generate and display the historical wear curve and the predicted wear curve, without direct manual measurement, thereby reducing the labor cost and detection difficulty, facilitating the management of the tire life cycle, and solving the technical problems of high cost and great difficulty in manual detection of the tire pattern depth in the prior art.

[0099] The storage module is generally designed in the form of a database to facilitate the storage of various information of the vehicle and the tire.

[0100] The time / mileage selection module includes a time selection module or a mileage selection module and is configured to select the time or the mileage at which the pattern depth is to be predicted.

[0101] The tire selection module, the time / mileage selection module, and the curve display module are all designed on one user interface, as shown in FIG. 6, to facilitate the selection and observation of the user and to realize the tire position number screening and the wear curve visualization. Figure 2

[0102] The time selection module can be designed as a month scroll bar to facilitate the selection of the time by the user, as shown in FIG. 7. Figure 4

[0103] In some embodiments of the present application, to facilitate the selection of the tire by the user, the tire selection module includes a vehicle number selection box and a plurality of tire selection buttons.

[0104] ​​A vehicle number selection box is configured to select a vehicle number.

[0105] A plurality of tire selection buttons are configured to select different positions of tires of the selected vehicle. Each tire selection button displays the position of the corresponding tire in the selected vehicle and the current tread depth of the corresponding tire.

[0106] A tread depth prediction module is configured to, in response to a prediction instruction (e.g., a click instruction of a prediction button on the interface), invoke a prediction model to predict the tread depth of the tire selected by the tire selection button at the time / mileage selected by the time / mileage selection module, and send the predicted tread depth to the curve generation module.

[0107] The curve generation module is configured to, when receiving a click instruction of any tire selection button, generate a historical wear curve according to the historical tread data of the tire corresponding to the tire selection button; and when receiving the predicted tread depth sent by the tread depth prediction module, generate a predicted wear curve.

[0108] For example, an intelligent guided vehicle has eight tires, and therefore eight tire selection buttons are designed. Each tire selection button displays the position of the tire and the tread depth, so as to realize the visualization of the wear of all tires, as shown in FIG. 1. Figure 3

[0109] The vehicle number selection box can be a drop-down box, which is convenient for a user to select a vehicle number, as shown in FIG. 2. Figure 5

[0110] In some embodiments of the present application, in order to facilitate a user to accurately and quickly select a target tire, the positional relationship of the plurality of tire selection buttons is consistent with the actual positional relationship of the corresponding tires in the vehicle.

[0111] In some embodiments of the present application, the color of the tire selection button is determined according to the current tread depth of the corresponding tire. A user can conveniently and intuitively obtain the tread depth of the tire through the color of the tire selection button.

[0112] In some embodiments of the present application, the color of the tire selection button is determined according to the current tread depth of the corresponding tire, and specifically includes:

[0113] The color of the tire selection button is determined according to the following function:

[0114] color(d) = (R(d), G(d), B(d));

[0115] wherein d is the current tread depth of the tire, and the unit is millimeter.

[0116] ​​

[0117] When the pattern depth d of the tire is ≥ 20 mm, color(d) = (173, 216, 230);

[0118] When the pattern depth d of the tire is < 10 mm, R(d) = 255, G(d) < 255, B(d) = 0.

[0119] Therefore, the color change function color(d) of the tire selection button, which is expressed by a triple (R, G, B) with R being the red channel, G being the green channel, and B being the blue channel, can make the color of the tire selection button change smoothly from light blue to red as the pattern depth of the tire decreases, thereby warning the user through the color of the tire so that the user can take timely measures.

[0120] In some embodiments of the present application, in order to facilitate the user to select the tire, the tire selection module further comprises a number selection box for selecting the tire number.

[0121] The pattern depth prediction module, in response to a prediction instruction, calls the prediction model to predict the pattern depth of the tire selected by the number selection box at the time / mileage selected by the time / mileage selection module and sends it to the curve generation module;

[0122] The curve generation module generates a historical wear curve according to the historical pattern data of the tire selected by the number selection box and generates a predicted wear curve when receiving the predicted pattern depth sent by the pattern depth prediction module.

[0123] Since each tire has a unique number, the unique tire can be located through the tire number. When the historical wear curve and the predicted wear curve of the target tire are needed to be known, the target tire number is selected through the number selection box, the curve generation module generates the historical wear curve according to the historical pattern data of the tire selected by the number selection box and generates the predicted wear curve according to the predicted pattern depth of the tire selected by the number selection box.

[0124] Therefore, the curve generation module, when receiving the click instruction of any tire selection button, generates a historical wear curve according to the historical pattern data of the tire corresponding to the tire selection button; and can also generate a historical wear curve according to the historical pattern data of the tire selected by the number selection box.

[0125] The pattern depth prediction module, in response to a prediction instruction, calls the prediction model to predict the pattern depth of the tire selected by the number selection box at the time / mileage selected by the time / mileage selection module and sends it to the curve generation module;

[0126] The curve display module displays the historical wear curve and the predicted wear curve generated by the curve generation module, as shown in Figure 6

[0127] In some embodiments of the present application, in order to facilitate the user to select all the tires of a single vehicle, the intelligent guided transport vehicle tire life cycle management system further comprises a single vehicle tire full selection box.

[0128] The single vehicle tire full selection box is used to select all the tires of a single vehicle.

[0129] The curve generation module is further configured to generate an average wear curve of all the tires of the selected single vehicle according to historical tread data of all the tires of the selected single vehicle.

[0130] First, the vehicle number selection box is used to select the vehicle number to select the vehicle, and the single vehicle tire full selection box is checked to select all the tires of a single vehicle. The curve generation module generates an average wear curve of all the tires of the selected single vehicle according to historical tread data of all the tires of the selected single vehicle. Then, the curve display module displays the average wear curve generated by the curve generation module.

[0131] By designing the single vehicle tire full selection box, the user can easily select all the tires of a single vehicle, and then the curve generation module generates an average wear curve of all the tires of the vehicle, thereby enriching the functions of the intelligent guided transport vehicle tire life cycle management system.

[0132] In some embodiments of the present application, in order to facilitate the user to select the tires of a position of a plurality of vehicles, the intelligent guided transport vehicle tire life cycle management system further comprises a vehicle multi-selection box and a tire selection box.

[0133] The vehicle multi-selection box is used to select a plurality of vehicles.

[0134] The tire selection box is used to select the position of the tire. For example, left front 1, left front 2, left rear 1, left rear 2, right front 1, right front 2, right rear 1, and right rear 2.

[0135] The curve generation module is further configured to generate an average wear curve of the tires of the selected position of the selected plurality of vehicles according to historical tread data of the tires of the selected position of the selected plurality of vehicles.

[0136] First, the vehicle multi-selection box is used to select a plurality of vehicles, and then the tire selection box is used to select the position of the tire. The curve generation module generates an average wear curve of the tires of the selected position of the selected plurality of vehicles according to historical tread data of the tires of the selected position of the selected plurality of vehicles. Then, the curve display module displays the average wear curve generated by the curve generation module.

[0137] ​For example, if the left front tire of three vehicles is selected, the curve generation module generates the average wear curve of the left front tire of these three vehicles.

[0138] By designing vehicle multi-select boxes and tire selection boxes, users can easily select tires at specified locations on multiple vehicles. Then, the curve generation module generates the average wear curves of the selected tires at the selected locations on the multiple vehicles, enriching the functionality of the intelligent guided transport vehicle tire lifecycle management system.

[0139] In some embodiments of this application, the prediction model is established using Taylor series expansion and the parameters are updated using the least squares method to obtain an accurate prediction model, which is simple and convenient.

[0140] In some embodiments of this application, the prediction model is established using Taylor series expansion and the parameters are updated using the least squares method, specifically including:

[0141] (11) Establish a prediction model:

[0142]

[0143] Where h(t) n (t) represents time point t n The tire tread depth; t0 is the time expansion point; For parameters.

[0144] remember:

[0145] Among them, feature vector parameter vector

[0146] (12) Use the least squares method to update the parameters to minimize the prediction error:

[0147] i = 1, 2, 3, ..., n;

[0148] i∈[1,n], where the value of i starts from 1 and goes up to n, and the following calculations are performed respectively;

[0149] Calculate the gain matrix:

[0150] Calculate the prediction error: e i =h(t) i )-φ(t i ) T θ i-1 ;

[0151] Update parameter vector: θ i =θ i-1 +K i e i ;

[0152] Update the covariance matrix:

[0153] where h(t1), h(t2), …, h(tn) are known historical tire pattern data; n

[0154] λ∈(0,1] is the forgetting factor, and is the known data;

[0155]

[0156] The model parameters are corrected using the least squares method to minimize the prediction error.

[0157] When i = 1, the first iteration calculation:

[0158]

[0159] e1 = h(t1) - φ(t1) T θ0

[0160] θ1 = θ0 + K1e1;

[0161]

[0162] The predicted value of the tire pattern depth at time t1 is: h1pre = φ(t1) T θ1.

[0163] When i = 2, the second iteration calculation:

[0164]

[0165] e2 = h(t2) - φ(t2) T θ1

[0166] θ2 = θ1 + K2e2;

[0167]

[0168] The predicted value of the tire pattern depth at time t2 is: h2pre = φ(t2) T θ2.

[0169] When i = n, the nth iteration calculation:

[0170]

[0171] e n = h(t n ) - φ(t n ) T θ n-1 ​

[0172] θ n = θ n-1 + K n e n ;

[0173]

[0174] t n he n pre = φ(t n ) T θ n .

[0175] After the above iteration calculation, the prediction error is very small, and the prediction error has converged.

[0176] (13) using the prediction model, predicting the tire tread depth h(t n+k ) at time point t n+k ; wherein k≥1;

[0177] Calculate the gain matrix:

[0178] Calculate the prediction error: e n+k = e n+k-1 + e + ;

[0179] Wherein e + is a random number in the range of (-0.01, +0.01);

[0180] Update the parameter vector: θ n+k = θ n+k-1 + K n+k e n+k ;

[0181] Update the covariance matrix:

[0182] Calculate the tire tread depth: h(t n+k ) = φ(t n+k ) T θ n+k .

[0183] Specifically:

[0184] First calculation: predict the tire tread depth h(t n+1 ) at time point t n+1 , that is, perform the following calculation:

[0185] Calculate the gain matrix:

[0186] Compute prediction error: e n+1 = e n + e + ;

[0187] Update parameter vector: θ n+1 = θ n + K n+1 e n+1 ;

[0188] Update covariance matrix:

[0189] Compute tire tread depth at time t n+C : h(t n+1 ) = φ(t n+1 ) T θ n+1 .

[0190] 2nd computation: Compute tire tread depth h(t n+2 ) at prediction time point t n+2 , i.e., perform the following computation:

[0191] Compute gain matrix:

[0192] Compute prediction error: e n+2 = e n+1 + e + ;

[0193] Update parameter vector: θ n+2 = θ n+1 + K n+2 e n+2 ;

[0194] Update covariance matrix:

[0195] Compute tire tread depth at time t n+2 : h(t n+2 ) = φ(t n+2 ) T θ n+2 .

[0196] ...

[0197] kth computation: Compute tire tread depth h(t n+k ) at prediction time point t n+k , i.e., perform the following computation:

[0198] Compute gain matrix:

[0199] Compute prediction error: e n+k = e n+k-1 + e +;

[0200] Update parameter vector: θ n+k = θ n+k-1 + K n+k e n+k ;

[0201] Update covariance matrix:

[0202] Calculate tire pattern depth: h(t n+k ) = φ(t n+k ) T θ n+k .

[0203] Therefore, when predicting the tire pattern depth h(t n+k ) at time point t n+k , h(t n+1 ) at time point t n+1 , h(t n+2 ) at time point t n+2 , …, and h(t n+k ) at time point t n+k are sequentially predicted.

[0204] For example:

[0205] If k = 1, i.e., the tire pattern depth h(t n+1 ) at time point t n+1 is predicted by using the prediction model, one calculation is performed to calculate the gain matrix K n+1 , the prediction error e n+1 , the parameter vector θ n+1 , the covariance matrix P n+1 , and the tire pattern depth h(t n+1 ).

[0206] If k = 2, i.e., the tire pattern depth h(t n+2 ) at time point t n+2 is predicted by using the prediction model, first, the first calculation is performed to calculate the gain matrix K n+1 , the prediction error e n+1 , the parameter vector θ n+1 , the covariance matrix P n+1 , and the tire pattern depth h(t n+1 ), and then the second calculation is performed to calculate the gain matrix K n+2 , the prediction error e n+2 , the parameter vector θ n+2 , the covariance matrix P n+2 , and the tire pattern depth h(t n+2 ).

[0207] If k = 3, that is, using the prediction model to predict time point t n+3 Tire tread depth h(t) n+3 If so, the first calculation is performed to calculate the gain matrix K. n+1 Prediction error e n+1 , parameter vector θ n+1 Covariance matrix P n+1 Tire tread depth h(t) n+1 Then, a second calculation is performed to calculate the gain matrix K. n+2 Prediction error e n+2 , parameter vector θ n+2 Covariance matrix P n+2 Tire tread depth h(t) n+2 Then, a third calculation is performed to calculate the gain matrix K. n+3 Prediction error e n+3 , parameter vector θ n+3 Covariance matrix P n+3 Tire tread depth h(t) n+3 ).

[0208] Calculate the gain matrix:

[0209] Calculate the prediction error: e n+3 =e n+2 +e + ;

[0210] Update parameter vector: θ n+3 =θ n+2 +K n+3 e n+3 ;

[0211] Update the covariance matrix:

[0212] Calculate t n+3 Tire tread depth at time: h(t) n+3 )=φ(t n+3 ) T θ n+3 .

[0213] By establishing the above prediction model with time as the input variable and using the least squares method to update the parameters, accurate pattern depth prediction can be achieved.

[0214] The above prediction model takes a time point as input and outputs the predicted pattern depth. The parameters of the prediction model are updated using the least squares method. Here, n represents the number of historical pattern depth data points.

[0215] In some embodiments of the present application, the prediction model is established using a Taylor series expansion and the parameters are updated using a least square method, which specifically includes:

[0216] (21) Establishing a prediction model:

[0217]

[0218] wherein h(s n ) is the tire pattern depth of the mileage point s n ; s0 is the expansion point of the mileage; is a parameter;

[0219] Note: h(s n ) = φ(s n ) T θ n ;

[0220] wherein the feature vector is the parameter vector

[0221] (22) Update the parameters using a least square method to minimize the prediction error:

[0222] i∈[1,n], i takes values from 1 to n, and the following calculations are performed respectively;

[0223] Calculate the gain matrix:

[0224] Calculate the prediction error: e i = h(s i ) - φ(s i ) T θ i-1 ;

[0225] Update the parameter vector: θ i = θ i-1 + K i e i ;

[0226] Update the covariance matrix:

[0227] wherein h(s1), h(s2), …, h(s n ) are known historical pattern data;

[0228] λ∈(0,1] is a forgetting factor;

[0229] (23) Use the prediction model to predict the tire pattern depth h(s n+k ) of the mileage point s n+k ; wherein k≥1;

[0230] Compute gain matrix:

[0231] Compute prediction error: e n+k = e n+k-1 + e + ;

[0232] where e + is a random number in the range (-0.01, +0.01);

[0233] Update parameter vector: θ n+k = θ n+k-1 + K n+k e n+k ;

[0234] Update covariance matrix:

[0235] Compute tire tread depth: h(s n+k ) = φ(s n+k ) T θ n+k .

[0236] By establishing the above prediction model with mileage as the input variable, and using the least squares method to update the parameters, accurate tread depth prediction can be achieved. Wherein n is the number of historical tread depth data.

[0237] The input of the above prediction model is the mileage point, and the output is the predicted tread depth. The least squares method is used to update the parameters of the prediction model. The specific iterative calculation process of the parameter update and prediction of the prediction model with mileage as the input can refer to the description of the prediction model with time as the input, which will not be repeated here.

[0238] For example, the interval between two adjacent time points is 1 month, and the interval between two adjacent mileage points is 5000 meters.

[0239] Next, taking time as an input variable, the RLS adaptive Taylor series update prediction algorithm is described.

[0240] (1) Taylor series local approximation.

[0241] In the prediction model, the input variable can be time or mileage, etc. Here, time is taken as the input variable for expansion. The Taylor series expansion is:

[0242]

[0243] Where: h(t0) is the tire tread depth at the expansion point t0;

[0244] tn The time point is the expansion point t0, which is usually selected as a point close to the prediction point, such as predicting the wear curve in October, November and December, and the expansion point is August.

[0245] and are the first and second derivatives at the expansion point, respectively.

[0246] The above expansion is approximated as a second-order polynomial, i.e. the prediction model:

[0247]

[0248] Take the parameter vector: Corresponding to h(t0), the first derivative and half of the second derivative, respectively.

[0249] (2) Recursive least squares (RLS).

[0250] By continuously correcting the parameter vector θ, the prediction error is minimized:

[0251]

[0252] Where λ∈(0,1] is the forgetting factor, used to reduce the weight of old data and adapt to dynamic changes. When λ is equal to 1, old data is emphasized, and when λ is equal to 1, new data is dynamically updated. The model takes λ = 0.98, which is suitable for the tire tread depth prediction model with smooth transition of wear change. φ(t n ) = [1, (t n -t0), (t n -t0) 2 ] T is the feature vector.

[0253] Parameter update formula:

[0254] Each time a new data point (t, h(t)) is added, the parameter vector θ and the covariance matrix P are updated:

[0255] The initial covariance matrix P can be represented as: P0 = 100·I3;

[0256] Where I3 is a 3x3 identity matrix to correspond to the parameter vector.

[0257]

[0258] Gain matrix:

[0259] Control the weight proportion of new data.

[0260] Prediction error: e n = h(t n )-φ(tn ) T θ n-1 ;

[0261] Parameter vector update: θ n =θ n-1 +k n e n ;

[0262] Covariance matrix update: The iterative process of the RLS adaptive Taylor series update prediction model can be found in [link to relevant documentation]. Figure 7 As shown below, the prediction iteration process will be explained in detail below, using specific pattern depth data. Table 1 below shows the pattern depth for each month.

[0263] Table 1

[0264]

[0265]

[0266] I. Model Training Cycle:

[0267] Selection of the Taylor expansion point t0: Select the two dates before the start of the forecast data when data becomes available. For example, if the forecast data starts in September, then July is chosen as the Taylor expansion point t0.

[0268] 1. First iteration: Iterate to obtain θ1, P1 and K1.

[0269] Completed two simulations and predictions of tire tread depth for January:

[0270] The initial value of h1pre is 0; after θ1 correction, it becomes...

[0271] The specific calculation steps are as follows:

[0272] The eigenvector φ(t) is:

[0273]

[0274] When t = 1, the eigenvector φ(1) is:

[0275]

[0276] Covariance matrix P n Its initialization form P0 is:

[0277]

[0278] The gain matrix K1, prediction error e1, and parameter vector θ1 are respectively:

[0279]

[0280]

[0281] Tire pattern depth h1 prediction at t1 node under semi-supervised learning:

[0282]

[0283] Update covariance matrix P1:

[0284]

[0285] 2. Second round of iteration: iterate out K2, θ2 and P2.

[0286] Complete 2 months tire pattern depth simulation prediction:

[0287] The initial value of h2pre is 11.43; after θ2 correction, it is

[0288] The specific calculation steps are as follows:

[0289] When t = 2, the eigenvector φ(2) is:

[0290]

[0291] θ1, P1 and K1 are taken from the results of the first round of iteration, and the gain matrix K2, the prediction error e2 and the parameter vector θ2 are calculated as:

[0292]

[0293] Tire pattern depth h2 prediction at t2 node under semi-supervised learning:

[0294]

[0295] Update covariance matrix P2:

[0296]

[0297] 3. Third round of iteration: iterate out θ3 and P3.

[0298] e3 = 0.7534776636549498

[0299]

[0300] h3pre = 14.76

[0301]

[0302] 4. Fourth iteration: iteration out θ4 and P4.

[0303] e4 = 0.7534776636549498

[0304] h4pre = 13.802091259892602

[0305]

[0306]

[0307] e1 to e8 are: [16.37, 4.246873251794547, 0.7534776636549498, 1.2302321183369287, 1.2954241455576607, 0.7057324776222504,

[0308] 0.2974588681576531, 0.11780056039619602].

[0309] It can be seen that the error is getting smaller and smaller, and the predictability of the model is gradually increasing.

[0310] II. Model prediction period (from September):

[0311] Generally speaking, the closer to the Taylor expansion point, the better the convergence effect, but because θ9 is taken from θ8, so 8 months are still involved in the prediction, therefore, t0 is taken from the two nodes before the prediction month, so the effect will be better from July.

[0312] 1. About the prediction calculation of September:

[0313]

[0314] e9 = e8 + e + ;

[0315] θ9 = θ8 + K9e9;

[0316] The predicted September pattern depth is:

[0317] At this time, there is no step to compensate for the correction to the real data, and the e value is already small enough, and a random noise conforming to the normal distribution (Gaussian distribution) is added to e.

[0318]

[0319] e + ~ N(0, σ2), σ = 0.01

[0320] e9 = e8 + e+

[0321] 2. Forecasting calculation for October:

[0322]

[0323] e 10 = e9+ e + ;

[0324] Calculate θ 10 = θ9+ K 10 e 10 ;

[0325] The predicted October tread depth is:

[0326]

[0327] e + ~ N(0, σ2), σ = 0.01

[0328] The curve graph drawn according to the above data is shown in Figure 8 .

[0329] Since mileage is more strongly related to tire wear, the prediction model can also use mileage as an independent variable. The following Table 2 shows the tread depth corresponding to the mileage.

[0330] Table 2

[0331]

[0332] After replacing the month with the driving mileage, the prediction accuracy remains basically unchanged, and the predicted curve graph is shown in Figure 9 .

[0333] However, the calculation amount has reached an astronomical number level at this time. For example:

[0334]

[0335] The system architecture of the present application includes a data collection and management module, a user interface design, and an intelligent prediction module.

[0336] (1) Data collection and management module: a) Regularly record the tread depth data of each tire of each vehicle to form a basic data set. b) Design the database table structure, including the fields of tire number, vehicle number, tire position, and tread depth. c) Implement data addition, deletion, modification, and query operations, and support batch import and export functions.

[0337] (2) User interface design: a) Design the interface tire global map according to the IGV tire layout. b) Add database interface: vehicle number selection, tire position selection, number filter, month selection scroll bar, etc. c) Data visualization module: ① Draw the tire pattern depth curve according to the selected data; ② Add color change function, change the color of the tire display module from light blue to red according to the tire pattern depth.

[0338] (3) Intelligent prediction module: a) Based on historical data and RLS adaptive Taylor series update algorithm, provide tire wear prediction.

[0339] The intelligent guided transport vehicle tire life cycle management system based on least square method adaptive Taylor series update of the application relates to the technical field of intelligent guided transport vehicle management of automated wharf, in particular to an intelligent guided transport vehicle tire life cycle management system based on recursive least square method adaptive Taylor series update. The system realizes scientific and fine whole life cycle management of the tires of the intelligent guided transport vehicle of the automated wharf.

[0340] The application aims to solve the technical problems mentioned in the background art, and provides an intelligent guided transport vehicle tire life cycle management system based on recursive least square method (RLS) adaptive Taylor series update, which realizes scientific and fine tire management.

[0341] The system of the application realizes intelligent tire identification and distribution: (a) Each tire is assigned a unique number, usually using the tire's own number, while recording brand information, etc. (b) Each vehicle is regarded as an independent "shelf", and each tire position is regarded as an independent storage space, which is convenient for management and tracking.

[0342] The system of the application realizes user-friendly interface design: (a) Provides the function of viewing according to the order of storage position (tire position) after clicking the shelf (vehicle), and 8 tire frame top view option boxes, which facilitates users to intuitively understand the tire state. (b) Supports multiple search functions, including tire number search, vehicle number search, month search scroll bar, tire mileage search, tire position search, etc., which improves data query efficiency.

[0343] The system of the present application realizes advanced data analysis and visualization: (a) Realize tire position and wear visualization in the image user interface (GUI), refer to the IGV tire layout to display the tire display block (i.e. tire selection button), the tire display block (tire selection button) can simultaneously display the tire pattern depth, and color distinguish according to the depth of the pattern depth, when the pattern wears to a certain extent, the display block will become orange warning color, and will be changed to red if still continue to wear. (b) Provide the average wear curve of 8 tires of a single vehicle, the specific tire position wear curve of a single vehicle, support the check function, and flexibly select the wear curve of a specific position for comparison and analysis. (c) Have the ability to generate average wear curve of specified parts of all vehicles, help managers master the tire wear condition comprehensively.

[0344] The system of the present application realizes intelligent prediction and maintenance suggestion: based on historical data and RLS adaptive Taylor series update algorithm, provide tire wear prediction and tire replacement plan suggestion.

[0345] The system of the present application realizes the establishment of tire analysis database: the database design is reasonable, supports multi-dimensional retrieval and statistics, and improves work efficiency. Based on the collected and analyzed tire type, process, brand data, scientific data analysis is carried out, and data basis is established for tire procurement and selection.

[0346] The system of the present application has the following advantages:

[0347] (1) Efficient data management: a) Can quickly process and analyze large-scale data, ensure the accuracy and integrity of the data. b) The database design is reasonable, supports multi-dimensional retrieval and statistics, and improves work efficiency.

[0348] (2) Accurate wear monitoring: a) Regularly record the tire pattern depth, combine with data analysis, realize accurate monitoring of tire wear condition, and find potential problems in time. b) Through the visualization tool, intuitively display the tire wear trend, and facilitate the managers to make decisions.

[0349] (3) Flexible query and statistical function: a) Multiple search methods meet the query needs in different scenarios, improve user experience. b) Provide rich statistical functions, support single tire, single vehicle, all vehicles and other multi-dimensional data analysis, help managers understand the tire use condition comprehensively.

[0350] (4) Intelligent prediction and optimization suggestion: a) Combine historical data and RLS adaptive Taylor series update algorithm, provide tire wear prediction and replacement suggestion, scientifically plan downtime, and improve equipment utilization.

[0351] The application provides an innovative intelligent guided transport vehicle tire life cycle management system, which realizes efficient management and monitoring of the tire of the intelligent guided transport vehicle at the automatic wharf through intelligent tire identification and distribution, multi-dimensional data acquisition and storage, user-friendly interface design, advanced data analysis and visualization, and intelligent prediction. Compared with the traditional method, the application has higher data management efficiency, more accurate wear monitoring, and more flexible query and statistical functions, and provides strong technical support for modern industry and logistics industry.

[0352] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit it; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent guided transport vehicle tire lifecycle management system, characterized in that: include: A storage module used to store the historical tread depth of the tire; Tire selection module, used to select tires; The time / mileage selection module is used to select the time / mileage for which pattern depth needs to be predicted; The tread depth prediction module is used to call the prediction model to predict the tread depth of the tire selected by the tire selection module at the time / mileage selected by the time / mileage selection module. The curve generation module is used to generate a historical wear curve based on the historical tread data of the tire selected by the tire selection module, and also to generate a predicted wear curve based on the tread depth predicted by the tread depth prediction module. The curve display module is used to display the historical wear curves and predicted wear curves generated by the curve generation module.

2. The intelligent guided transport vehicle tire lifecycle management system according to claim 1, characterized in that: The tire selection module includes: The vehicle number selection box is used to select the vehicle number; Multiple tire selection buttons; each tire selection button corresponds to a different tire on the vehicle selected by the vehicle number selection box, and each tire selection button displays the position of the corresponding tire on the selected vehicle and the current tread depth; The tread depth prediction module, in response to the prediction command, calls the prediction model to predict the tread depth of the tire selected by the tire selection button at the time / mileage selected by the time / mileage selection module, and sends it to the curve generation module. When the curve generation module receives a click instruction for any tire selection button, it generates a historical wear curve based on the historical tread data of the tire corresponding to that tire selection button; when it receives a predicted tread depth sent by the tread depth prediction module, it generates a predicted wear curve.

3. The intelligent guided transport vehicle tire lifecycle management system according to claim 2, characterized in that: The positional relationship of the multiple tire selection buttons is consistent with the actual positional relationship of the corresponding tires on the vehicle.

4. The intelligent guided transport vehicle tire lifecycle management system according to claim 2, characterized in that: The color of the tire selection button is determined according to the current tread depth of the corresponding tire.

5. The intelligent guided transport vehicle tire lifecycle management system according to claim 4, characterized in that: The color of the tire selection button is determined according to the current tread depth of the corresponding tire, specifically including: The color of the tire selection button is determined according to the following function; color(d)=(R(d),G(d),B(d)); Where d is the current tread depth of the tire, in millimeters; 6. The intelligent guided transport vehicle tire lifecycle management system according to claim 1, characterized in that: The tire selection module also includes a number selection box for selecting tire numbers; The tread depth prediction module, in response to the prediction command, calls the prediction model to predict the tread depth of the tire selected by the number selection box at the time / mileage selected by the time / mileage selection module, and sends it to the curve generation module. The curve generation module generates a historical wear curve based on the historical tread data of the tire selected by the number selection box; and generates a predicted wear curve when it receives the predicted tread depth sent by the tread depth prediction module.

7. The intelligent guided transport vehicle tire lifecycle management system according to claim 1, characterized in that: The intelligent guided transport vehicle tire lifecycle management system also includes: The "Select All Tires for a Single Vehicle" box is used to select all tires of a single vehicle. The curve generation module is also used to generate the average wear curve of all tires of the selected single vehicle based on the historical tread pattern data of all tires of that vehicle.

8. The intelligent guided transport vehicle tire lifecycle management system according to claim 1, characterized in that: The intelligent guided transport vehicle tire lifecycle management system also includes: The vehicle selection box is used to select multiple vehicles. The tire selection box is used to select the location of the tire; The curve generation module is also used to generate the average wear curve of the tires at selected locations of the selected multiple vehicles.

9. The intelligent guided transport vehicle tire lifecycle management system according to any one of claims 1 to 8, characterized in that: The prediction model is established using Taylor series expansion and the parameters are updated using the least squares method.

10. The intelligent guided transport vehicle tire lifecycle management system according to claim 9, characterized in that: The prediction model is established using Taylor series expansion and updated using the least squares method, specifically including: (11) Establish a prediction model: Where h(t) n (t) represents time point t n The tire tread depth; t0 is the time expansion point; For parameters; Note: h(t n ) = φ(t n ) T θ n ; Among them, feature vector parameter vector (12) Use the least squares method to update the parameters to minimize the prediction error: i∈[1,n], where the value of i starts from 1 and goes up to n, and the following calculations are performed respectively; Calculate the gain matrix: Calculate the prediction error: e i =h(t) i )-φ(t i ) T θ i-1 ; Update parameter vector: θ i =θ i-1 +K i e i ; Update the covariance matrix: Where h(t1), h(t2), ..., h(t) n () represents known historical pattern data; λ∈(0,1] is the forgetting factor; (13) Using the prediction model, predict time point t n+k Tire tread depth h(t) n+k ); where k≥1; Calculate the gain matrix: Calculate the prediction error: e n+k =e n+k-1 +e + ; Among them, e + A random number within the range (-0.01, +0.01); Update parameter vector: θ n+k =θ n+k-1 +K n+k e n+k ; Update the covariance matrix: Calculate tire tread depth: h(t) n+k )=φ(t n+k ) T θ n+k ; or, (21) Establish a prediction model: Where h(s) n ) represents the mileage point s n The tire tread depth; s0 is the mileage expansion point; For parameters; Note: h(s n ) = φ(s n ) T θ n ; Among them, feature vector parameter vector (22) Use the least squares method to update the parameters to minimize the prediction error: i∈[1,n], where the value of i starts from 1 and goes up to n, and the following calculations are performed respectively; Calculate the gain matrix: Calculate the prediction error: e i =h(s i )-φ(s i ) T θ i-1 ; Update parameter vector: θ i =θ i-1 +K i e i ; Update the covariance matrix: Among them, h(s1), h(s2), ..., h(s) n () represents known historical pattern data; λ∈(0,1] is the forgetting factor; (23) Using a prediction model, predict the mileage point s. n+k Tire tread depth h(s) n+k ); where k≥1; Calculate the gain matrix: Calculate the prediction error: e n+k =e n+k-1 +e + ; Among them, e + A random number within the range (-0.01, +0.01); Update parameter vector: θ n+k =θ n+k-1 +K n+k e n+k ; Update the covariance matrix: Calculate tire tread depth: h(s) n+k )=φ(s n+k ) T θ n+k .

Images