Research on Multi-Mode Variable Parameter Intelligent Shift Control Method of Loader Based on RBF Network

: The loader is one of the most widely used pieces of engineering machinery in the world for soil transportation, loading and unloading materials, and low-intensity shovel digging operations in harsh and complex operating conditions; it requires very frequent shifting and has other challenging characteristics. In order to realize automatic frequent shifting, we need to better design the shifting rules in the shifting process, improve the shifting quality and working efficiency, and solve the key engineering problems of energy saving and high efficiency in the shifting process of loaders. In this paper, a 7-ton wheel loader is taken as the research object, the loader shoveling process of the four operating modes is analyzed, and a multi-mode variable parameter shift law is designed. Aiming at the complicated and nonlinear characteristics of the power transmission system of the loader, an intelligent shift control method based on an RBF neural network is proposed. Finally, the simulation test and the clutch shift oil pressure test are carried out. From the test results, the clutch test oil pressure curve obviously shows a four-stage upward trend during shifting, and the buffering effect is obvious. The designed multi-mode variable-parameter intelligent shift law of the loader is reasonable and feasible, and the shift recognition rate reaches 97.92%, which provides theoretical support for the realization of intelligent automatic speed change control of the loader, and it certainly has engineering value.


Introduction
The loader is one of the most widely used pieces of engineering machinery in the world, mainly used for shoveling transportation, loading and unloading materials, lowintensity shovel-digging operations, round-trip shoveling, and loading cycle operations [1].However, the loader operating environment is harsh, the operating conditions are complex, and the driver needs to complete the shoveling operation through frequent gear shift operation, mainly to complete the two conditions of the six stages of operational tasks, namely, walking transportation conditions and loading and unloading conditions [2].Under different working conditions, the output characteristics and traction characteristics of the loader vary greatly, and the work pump can reach 40% to 60% of the engine output power when fully loaded, while the power consumed when unloaded is less than 10% [3].The loader, as an operating cycle of engineering vehicles, requires very frequent operation of the gear shift.According to statistics, in 1 h, the gear shift is operated on average up to more than 1000 times, that is, less than 4 s to shift once, so that the operator in control of the operating device has to shift frequently, greatly increasing the labor intensity of the operator, distracting the attention of the operator, and delaying the operating cycle of the loader.This greatly increases the operator's labor intensity, distracts his attention, delays the loader's cycle operation time, affects the efficiency of the operation, and increases the possibility of accidents [4]; on the other hand, the loader's work performance and work status are completely dependent on the operator's experience and intention to control, and once the operation is not proper, the loader's power performance cannot be guaranteed, and the torque converter is often in the inefficient zone, which leads to a decrease in the loader's economy, and the quality of the shifting gears will also be affected [5][6][7].
In order to solve the above problems, loader researchers widely draw lessons from the theory of automobile automatic transmission design, and since loader cycle operating conditions are complex and changeable, the working environment is harsh, the operating process requires frequent shifting, and due to other characteristics, the use of automatic shift technology is used to achieve a reduction in the operator's labor intensity, improve the dynamics of the loader and economy, improve the quality of shifting and operational efficiency, and for other purposes.At present, in the design of the loader shift law, there is mainly the two-parameter shift law (throttle opening and speed), three-parameter shift law (throttle opening, speed, and acceleration) [8-10], and four-parameter shift law (throttle opening, torque converter pump wheel speed, turbo speed, and oil pump pressure) [11,12], but most of these shift laws are directly quoted from the automobile's shifting strategy, which basically does not take into account the consumption of the oil pump power changes with working conditions, which makes it difficult to obtain a good automatic gearshift effect [13].Although the four-parameter shift law takes into account the power consumed by the working pump and the power flow fluctuation caused by the work of the oil pump, the power consumed by the oil pump is equated to the reduction of the engine throttle opening, which is not in line with the actual working condition of the loader [14].At the same time, because the engine and torque converter in the hydraulic transmission system of the loader is a highly nonlinear transmission system, it is difficult to solve such complex problems in the traditional linear control system, although some scholars have done some research in the intelligent control of the loader [15][16][17], but most of them use neural networks for shift control, and there is the problem of a low shift recognition rate [18,19].Therefore, this paper takes a 7-ton wheel loader of a certain type as the research object, analyzes the operation mode of the loader, studies the multi-mode variable parameter shift law in depth, puts forward the RBF neural network intelligent shift control method for the application of the loader, and carries out the simulation and shift oil pressure test.

Loader Operation Process Analysis
Wheel loaders, as a kind of widely used engineering machinery, are mainly used for short-distance shoveling and loading cycle operation.According to the different loading and transportation routes, they are generally divided into four operation modes: I-type, L-type, T-type, and V-type, as shown in Figure 1.Among them, I-type belongs to shuttle operation, L-type belongs to rotary operation, T-type belongs to right-angle rotary operation, and V-type belongs to semi-rotary operation.The above four operation modes are mainly based on saving operation time and reducing the number of turns, combined with the distribution of materials and the size of the operating site, in accordance with the principle of the most economical and efficient selection.Therefore, due to the high operating efficiency and short cycle time of the V-type operation mode, it is more representative to analyze the V-type operation mode of the loader [20,21].This paper takes the typical "V" cycle operation mode of wheel loaders as the research object, as shown in Figure 2. The shoveling operation process, as shown in the schematic diagram of the wheel loader V-type shoveling material work cycle, mainly includes six kinds of work phases: V1 unloaded to the stockpile, V2 shoveling material, V3 fully loaded backward to the operating origin, V4 fully loaded to the mining truck, V5 unloading, and V6 unloaded back to the starting point of the operation [22].This paper takes the typical "V" cycle operation mode of wheel loaders as the r search object, as shown in Figure 2. The shoveling operation process, as shown in th schematic diagram of the wheel loader V-type shoveling material work cycle, mainly in cludes six kinds of work phases: V1 unloaded to the stockpile, V2 shoveling material, V fully loaded backward to the operating origin, V4 fully loaded to the mining truck, V unloading, and V6 unloaded back to the starting point of the operation [22].V1 unloaded to the stockpile: the loader starts to move forward towards the materi and slows down when it is approaching the material.
V2 shoveling material: the loader inserts into the material at a certain speed, and a the resistance increases, the loader speed decreases.Adopt the shoveling method of in serting the material while lifting the working device until the bucket is full of material.
V3 Full load backward to the original point of operation: after the bucket is full material, the loader backs up to the original position with low-speed reverse gear.
V4 Driving to the mining truck with full load: the bucket starts from the original p sition and advances with a full load to approach the transportation loader and raises th  V1 unloaded to the stockpile: the loader starts to move forward towards the material and slows down when it is approaching the material.
V2 shoveling material: the loader inserts into the material at a certain speed, and as the resistance increases, the loader speed decreases.Adopt the shoveling method of inserting the material while lifting the working device until the bucket is full of material.
V3 Full load backward to the original point of operation: after the bucket is full of material, the loader backs up to the original position with low-speed reverse gear.
V4 Driving to the mining truck with full load: the bucket starts from the original position and advances with a full load to approach the transportation loader and raises the  V1 unloaded to the stockpile: the loader starts to move forward towards the material and slows down when it is approaching the material.
V2 shoveling material: the loader inserts into the material at a certain speed, and as the resistance increases, the loader speed decreases.Adopt the shoveling method of inserting the material while lifting the working device until the bucket is full of material.
V3 Full load backward to the original point of operation: after the bucket is full of material, the loader backs up to the original position with low-speed reverse gear.
V4 Driving to the mining truck with full load: the bucket starts from the original position and advances with a full load to approach the transportation loader and raises the moving arm while approaching the transportation truck.When the loader approaches the unloading position, the driver applies the brake, and the bucket is at the height of the unloading position.
V5 Unloading: Manipulate the bucket cylinder, and the bucket finishes unloading.V6 Retreating to the starting point of operation with no load: after unloading is completed, the loader retires to the original place with no load.

Loader Vehicle Parameters
This study focuses on a 7-ton wheeled loader of a certain type, with the vehicle parameters as shown in Table 1.Table 2 presents the transmission ratio and transmission efficiency values of the gearbox, while Table 3 shows the total transmission ratio and transmission efficiency values of the transmission system [23,24].

Loader Multi-Mode Variable Parameter Shift Law Design
According to the above loader operating conditions, the loader shift modes are mainly divided into three types: power mode, standard mode, and economic mode.The power mode is primarily used for shoveling and digging operations, the standard mode is mainly suited for light load and transportation conditions, and the economic mode is mainly applied for unladen driving [25].Therefore, in different modes, the main working condition objectives of the loader are different, which, in turn, affects the selection of shift parameters.Since the power mode and standard mode are mainly based on operating conditions and transportation conditions, in order to reflect the loader's workload more comprehensively, the shifting parameters are selected to be throttle, speed, and workload coefficient to realize three-parameter dynamic shifting; economic mode is not involved in the operating conditions, and neither are the pursuit of driving speed and fuel economy, so two parameters, throttle and speed, are selected to shift gears.Next, the three-mode variable parameter shift law proposed in this paper is designed as follows.

Power Mode Shift Law Design
The power mode of the loader is mainly based on the shoveling operation conditions, necessitating the assurance of optimal power output to fully utilize the loader's traction characteristics and efficiently complete shoveling tasks; that is, the shifting strategy prioritizes the dynamics of the loader.Therefore, based on the main task of power mode, this paper adopts the shifting strategy of optimal traction [20].The traction characteristic refers to the relationship between traction force, power, and speed of the loader in different gears.These can be analyzed and calculated according to the results of the common output characteristics of the diesel engine and torque converter, as well as the performance parameters of the transmission system, resulting in the expression of the speed calculation as in Equation ( 1).The formula for traction calculation is shown in Equation (2).
v-vehicle speed; n T -turbine speed; r k -tire radius; i z -total driveline ratio; i z -total transmission system ratio F K -loader driving force; T e -diesel engine torque; n p -output speed of working oil pump; n e -diesel engine speed; T w -torque consumed by working oil pump; K-torque coefficient of torque converter; i n -transmission n ratio at transmission gear; i 0 -drive axle ratio; η z -total transmission efficiency, 0.828.
From Equation ( 2), it can be seen that the traction force of the loader is the engine output torque minus the torque consumed by the oil pump set and then the force output to the driving wheels through the power transmission system.
In the power mode with different throttle openings and different workload coefficients, the loader's engine and torque converter have different common output characteristics and different traction characteristics.In this paper, the conditions of full throttle and full load are taken as examples to calculate the traction characteristics.The calculation results of the diesel engine and torque converter's combined output characteristics are presented in Table 4.The results of the loader's traction characteristics calculations are shown in Table 5.By following the aforementioned calculation method, changing the throttle opening and workload coefficient, and carrying out a traversal calculation, you can determine the correspondence between the shift parameters and the gear position.By fitting the calculated traction force and vehicle speed to generate a polynomial curve, the traction characteristic curve can be obtained.The full throttle/full load traction characteristic curve of the loader in power mode is shown in Figure 3.As can be seen from the figure, the optimal dynamic shift point between two adjacent gears is the intersection of the traction characteristic curve of each gear.The traction force is the same before and after the shift point of the two gears, which has less impact on the process of shifting gears, and it also ensures the demand for traction force before and after the shift.teristic curve can be obtained.The full throttle/full load traction characteristic curve of the loader in power mode is shown in Figure 3.As can be seen from the figure, the optimal dynamic shift point between two adjacent gears is the intersection of the traction characteristic curve of each gear.The traction force is the same before and after the shift point of the two gears, which has less impact on the process of shifting gears, and it also ensures the demand for traction force before and after the shift.Under the condition of full load in power mode, according to the above analytical calculations, the corresponding shift points can be obtained by traversing the throttle openings, as shown in Table 6.Similarly, by traversing the operating load factors under power mode conditions again, all the shift points corresponding to each throttle opening at different load factors can be identified.At this point, by plotting the three shift parameters of vehicle speed, throttle opening, and load factor in the same coordinate system, the three-parameter shift Under the condition of full load in power mode, according to the above analytical calculations, the corresponding shift points can be obtained by traversing the throttle openings, as shown in Table 6.Similarly, by traversing the operating load factors under power mode conditions again, all the shift points corresponding to each throttle opening at different load factors can be identified.At this point, by plotting the three shift parameters of vehicle speed, throttle opening, and load factor in the same coordinate system, the three-parameter shift law in the power mode of the loader can be designed, which consists of three sets of spatial surfaces, as shown in Figure 4.
From Figure 4, it can be observed that the three sets of shift surfaces correspond to downshifting from second gear to first gear and upshifting from first gear to second gear in power mode; downshifting from third gear to second gear and upshifting from second gear to third gear in power mode; and downshifting from fourth gear to third gear and upshifting from third gear to fourth gear in power mode.Furthermore, it is evident from the figure that the shifting strategy employs delayed downshifting, with the delay calculated as 10% of the full throttle shift point for the current gear.This approach effectively mitigates frequent fluctuations in gear engagement, reduces shifting shocks, and extends the lifespan of relevant shifting components.law in the power mode of the loader can be designed, which consists of three sets of spatial surfaces, as shown in Figure 4. From Figure 4, it can be observed that the three sets of shift surfaces correspond to downshifting from second gear to first gear and upshifting from first gear to second gear in power mode; downshifting from third gear to second gear and upshifting from second gear to third gear in power mode; and downshifting from fourth gear to third gear and upshifting from third gear to fourth gear in power mode.Furthermore, it is evident from the figure that the shifting strategy employs delayed downshifting, with the delay calculated as 10% of the full throttle shift point for the current gear.This approach effectively mitigates frequent fluctuations in gear engagement, reduces shifting shocks, and extends the lifespan of relevant shifting components.

Standard Mode Shift Law Design
As the standard mode mainly focuses on light load and transportation working conditions, this mode still focuses on the pursuit of dynamics in the process of loader working conditions, so it also takes the optimal traction as the shifting strategy and still adopts the three parameters of the throttle, vehicle speed, and workload coefficient as the shifting parameters to realize the design of the standard mode shifting law.With reference to the power mode shift law design method, using the matching standard mode output characteristic data, traversing the throttle opening, using the workload coefficient, and analyzing and calculating the standard mode traction characteristics of the loader and the corresponding shift point of all the shifts, you can design a three-parameter shift law in the standard mode and its shift law surface, as shown in Figure 5.

Standard Mode Shift Law Design
As the standard mode mainly focuses on light load and transportation working conditions, this mode still focuses on the pursuit of dynamics in the process of loader working conditions, so it also takes the optimal traction as the shifting strategy and still adopts the three parameters of the throttle, vehicle speed, and workload coefficient as the shifting parameters to realize the design of the standard mode shifting law.With reference to the power mode shift law design method, using the matching standard mode output characteristic data, traversing the throttle opening, using the workload coefficient, and analyzing and calculating the standard mode traction characteristics of the loader and the corresponding shift point of all the shifts, you can design a three-parameter shift law in the standard mode and its shift law surface, as shown in Figure 5.

Economic Mode Shift Law Design
Through the working condition analysis and shift law design of the power mode and the standard mode of the loader, it can be seen that these two modes are mainly used for shoveling and loading operations and transporting conditions with loads, and the traction characteristics of the loader are more demanding, so the traction performance index of the loader is mainly considered, followed by economic performance.However, the economic

Economic Mode Shift Law Design
Through the working condition analysis and shift law design of the power mode and the standard mode of the loader, it can be seen that these two modes are mainly used for shoveling and loading operations and transporting conditions with loads, and the traction characteristics of the loader are more demanding, so the traction performance index of the loader is mainly considered, followed by economic performance.However, the economic mode is mainly based on the unloaded driving conditions, which are primarily concerned with the driving speed and fuel consumption of the loader.Therefore, the shift law is mainly based on the economic indicators of the loader, that is, the shift law design with the best economic performance.The law aims to reduce the fuel consumption of the loader, which improves the transmission efficiency and reduces the operating cost, and aims to optimize the fuel economy of the loader by selecting the optimal shift point, that is, maximizing the output power at the unit fuel consumption.However, for the hydromechanical transmission, which is affected by the efficiency of the torque converter, only using the engine fuel consumption rate as the basis for automatic gear shifting does not achieve good results, and it is necessary to take the diesel engine and the torque converter as a transmission whole and to comprehensively consider the fuel consumption rate of the diesel engine and the efficiency of the transmission system.Therefore, when designing the optimal economic shift schedule, it is necessary to ensure that the hydraulic torque converter is always maintained at a high working efficiency as a prerequisite; for this reason, this paper will set the torque converter efficiency up to 70% as the high-efficiency zone, and the maximum efficiency of the torque converter used can be up to 88%.Thus, the higher the efficiency of the torque converter, the better it is when designing the shift law for the economic mode.Under this shift law, the loader does not carry out shoveling and transportation operations, the power requirements are relatively weak, and the fuel consumption is low, so for the selection of shift parameters, the design of the shift law is directly based on the two parameters of throttle opening and vehicle speed.
Using the common output characteristic data of the diesel engine matched with the torque converter in the economy mode and combined with Equation (1), this paper analyzes and calculates the speed of each gear with the full throttle opening as an example and plots the corresponding relationship between the vehicle speed and the efficiency of the torque converter by fitting the generated curve, as shown in Figure 6.From the figure, it can be observed that the intersection point of two adjacent torque converter efficiencies is higher than 70% or more, which satisfies the requirement that the torque converter is working in the high-efficiency zone, and this intersection point is the optimal shifting point of the economy shifting law when the throttle is fully in the economic mode.It can also be seen from the figure that the torque converter efficiency at the shift points of the second and third gears is low, which can be improved later through the perspective of torque converter structure design.Referring to the above design method and traversing the throttle openings, the two-parameter optimal economic shift law in the economic mode can be designed, and its shift-point data for each throttle opening is From the figure, it can be observed that the intersection point of two adjacent torque converter efficiencies is higher than 70% or more, which satisfies the requirement that the torque converter is working in the high-efficiency zone, and this intersection point is the optimal shifting point of the economy shifting law when the throttle is fully in the economic mode.It can also be seen from the figure that the torque converter efficiency at 2024, 13, 234 10 of 18 the shift points of the second and third gears is low, which can be improved later through the perspective of torque converter structure design.Referring to the above design method and traversing the throttle openings, the two-parameter optimal economic shift law in the economic mode can be designed, and its shift-point data for each throttle opening is shown in Table 7.Similarly, the equal delay two-parameter shift rule is adopted, and the delay is calculated according to 10% of the current shift point of full throttle.In its economic mode, the optimal economy shift law is shown in Figure 7.

Intelligent Gear Shift Control Method for Loader Based on RBF Networks
The key problem of automatic loader shift control is how to find the best gear position suitable for the current working conditions according to a certain mapping relationship.With this complex, nonlinear mapping relationship, it is difficult to use a mathematical expression, and the radial basis function (referred to as the RBF) neural network is precisely the solution to this type of complex, nonlinear problem of intelligent algorithms [26,27].In this paper, the intelligent gear shift control method of the loader with an RBF neural network is used to realize the automatic speed shift control strategy of the loader.

RBF Network Learning Algorithm
The selection of the RBF neural network basis function center is very important to network performance, and improper selection will directly affect network performance.The most commonly used is the clustering method; the algorithm has a hidden layer unit with a number of nodes that does not need to be determined in advance, the completion of the clustering of the RBF neural network is optimal, and the algorithm can be dynamically learned online, among other characteristics.This paper adopts the k-mean learning algorithm.
The clustering criterion function for the k-means method is:

Intelligent Gear Shift Control Method for Loader Based on RBF Networks
The key problem of automatic loader shift control is how to find the best gear position suitable for the current working conditions according to a certain mapping relationship.With this complex, nonlinear mapping relationship, it is difficult to use a mathematical expression, and the radial basis function (referred to as the RBF) neural network is precisely the solution to this type of complex, nonlinear problem of intelligent algorithms [26,27].In this paper, the intelligent gear shift control method of the loader with an RBF neural network is used to realize the automatic speed shift control strategy of the loader.

RBF Network Learning Algorithm
The selection of the RBF neural network basis function center is very important to network performance, and improper selection will directly affect network performance.The most commonly used is the clustering method; the algorithm has a hidden layer unit with a number of nodes that does not need to be determined in advance, the comple-tion of the clustering of the RBF neural network is optimal, and the algorithm can be dynamically learned online, among other characteristics.This paper adopts the k-mean learning algorithm.
The clustering criterion function for the k-means method is: The implication is that the sum of the squares of the errors between each class of samples u in the k subsets and the mean c i of the samples to which it belongs is then summed over all k classes of samples.Since the entire sample set has different categories, the algorithm produces different subsets of samples Γ i and their means c i , which results in different values of J e .Among these different values of J e , the best clustering is obtained by the k-means method when J e is the minimum value.
The steps of the k-means method are as follows: (1) Initialization: given the initial center c i (0) of each node.
Similar matching: calculate the distance and find the node with the minimum distance: Adjustment Center: Among these, learning efficiency β, 0 < β < 1.
(1) Add 1 to the value of t and perform the second step until rc the variable has a very small value end.
The RBF network learning process consists of the following two steps: 1 ⃝ Determination of center c i .The k-means cluster analysis technique was used to determine c i ; 2 ⃝ Determination of the radius σ j .The size of the radius σ j determines the range of the RBF unit's response to the input vector and affects the final classification accuracy of the network; 3 ⃝ Regulating weight w.Here, the regulating weights w are the connection weights between the output layer and the implicit layer of the network, which can be realized by the following two methods to regulate the weights of the network, respectively; 4 ⃝ Linear least squares method.The network output of this method is: The iterative formula is as follows:

Creation of RBF Neural Network Function Newrb
The instruction for the function newrb is [net,tr] = newrb(P,T,goal,spread), where P is the input vector; T is the target vector; goal is the network mean square error target; and spread is the smoothing factor.As a result, it is possible to obtain RBF networks of smaller sizes than newrbe.

RBF Neural Network Loader Intelligent Shift Control Simulation Experiments
A 7-ton wheel loader was used as the research object for the real vehicle test.Due to the limitations of the test process for collecting data, this paper only carried out the automatic shift test on the real vehicle in economic mode using the two parameters of engine speed and vehicle speed and conducted three upshifts.The engine speed and vehicle speed data were collected by sensors, and two sets of data were selected: the first group of speed and 2024, 13,234 engine speed data curves is shown in Figure 8, and the second group of speed and engine speed data curves is shown in Figure 9.

RBF Neural Network Loader Intelligent Shift Control Simulation Experiments
A 7-ton wheel loader was used as the research object for the real vehicle test.Due to the limitations of the test process for collecting data, this paper only carried out the automatic shift test on the real vehicle in economic mode using the two parameters of engine speed and vehicle speed and conducted three upshifts.The engine speed and vehicle speed data were collected by sensors, and two sets of data were selected: the first group of speed and engine speed data curves is shown in Figure 8, and the second group of speed and engine speed data curves is shown in Figure 9.According to the loader multi-mode variable parameter shifting law designed in the previous paper, a 7-ton wheel loader of a certain type is taken as the research object, and based on the RBF neural network model, the RBF neural network is established as the recognition model of the automatic shifting control network, and the RBF recognition network is established by using the RBF neural network of net = newrb (P,T,goal,bestX), in which P, T is the input/output of the network and goal is the training target.The recognition efficiency of the RBF neural network is highest when the expansion coefficient of the RBF neural network is 19.91.This extended coefficient is used to build the RBF neural network recognition model, and the recognition results are shown in Figure 10, from which it can be seen that the RBF neural network recognition model correctly recognizes up to 97.92% of the test samples.

RBF Network-Based Loader Multi-Mode Variable Parameter Intelligent Shift Law Test
In order to verify the correctness of the multi-mode variable-parameter intelligent According to the loader multi-mode variable parameter shifting law designed in the previous paper, a 7-ton wheel loader of a certain type is taken as the research object, and based on the RBF neural network model, the RBF neural network is established as the recognition model of the automatic shifting control network, and the RBF recognition network is established by using the RBF neural network of net = newrb (P,T,goal,bestX), in which P, T is the input/output of the network and goal is the training target.The recognition efficiency of the RBF neural network is highest when the expansion coefficient of the RBF neural network is 19.91.This extended coefficient is used to build the RBF neural network recognition model, and the recognition results are shown in Figure 10, from which it can be seen that the RBF neural network recognition model correctly recognizes up to 97.92% of the test samples.According to the loader multi-mode variable parameter shifting law designed in the previous paper, a 7-ton wheel loader of a certain type is taken as the research object, and based on the RBF neural network model, the RBF neural network is established as the recognition model of the automatic shifting control network, and the RBF recognition network is established by using the RBF neural network of net = newrb (P,T,goal,bestX), in which P, T is the input/output of the network and goal is the training target.The recognition efficiency of the RBF neural network is highest when the expansion coefficient of the RBF neural network is 19.91.This extended coefficient is used to build the RBF neural network recognition model, and the recognition results are shown in Figure 10, from which it can be seen that the RBF neural network recognition model correctly recognizes up to 97.92% of the test samples.

RBF Network-Based Loader Multi-Mode Variable Parameter Intelligent Shift Law Test
In order to verify the correctness of the multi-mode variable-parameter intelligent shifting law of the loader based on the RBF network, shifting oil pressure and time test

RBF Network-Based Loader Multi-Mode Variable Parameter Intelligent Shift Law Test
In order to verify the correctness of the multi-mode variable-parameter intelligent shifting law of the loader based on the RBF network, shifting oil pressure and time test experiments were used.The pressure sensor was installed to the hydraulic transmission clutch oil circuit, the shift oil pressure curve of each clutch was collected on the hydraulic transmission test bench, and the shift time was analyzed [28,29].The test and simulation results were compared to check the correctness and reliability of the simulation model.

Shift Law Test System
The test system is mainly composed of the test bench transmission device and acquisition and control parts.The hydraulic transmission test bench transmission device is mainly composed of a power device, hydraulic transmission, coupling, speed booster box, loading device, and other components.The hydraulic transmission test bench is shown in Figure 11.

Shift Law Test System
The test system is mainly composed of the test bench transmission device and acquisition and control parts.The hydraulic transmission test bench transmission device is mainly composed of a power device, hydraulic transmission, coupling, speed booster box, loading device, and other components.The hydraulic transmission test bench is shown in Figure 11.The test bench uses a three-phase asynchronous motor as a power device instead of the engine, through coupling with the hydraulic transmission before and after the installation of speed and torque sensors and an eddy current dynamometer as a loading device.Because the dynamometer has a limitation on the minimum output of the hydraulic transmission, in the test of the transmission, a dynamometer is installed beside a speed booster box, and the test uses a fixed transmission ratio.The control and acquisition part of the test system is mainly composed of a frequency converter, PLC, operation table, acquisition card, and control program.The main technical parameters of each component of the hydraulic transmission test bench are shown in Table 8.The hydrodynamic shift test bench adopts LabVIEW2018 software for control system program development.Through this, the control program can realize the function of shifting (manual or automatic), and it can also carry out various shifting tests, such as automatic shifting, shifting quality test, etc. [30].The control and acquisition part, as shown in Figures 12 and 13, represents the two-parameter automatic shifting program based on the vehicle speed and the degree of throttle opening.The test bench uses a three-phase asynchronous motor as a power device instead of the engine, through coupling with the hydraulic transmission before and after the installation of speed and torque sensors and an eddy current dynamometer as a loading device.Because the dynamometer has a limitation on the minimum output of the hydraulic transmission, in the test of the transmission, a dynamometer is installed beside a speed booster box, and the test uses a fixed transmission ratio.The control and acquisition part of the test system is mainly composed of a frequency converter, PLC, operation table, acquisition card, and control program.The main technical parameters of each component of the hydraulic transmission test bench are shown in Table 8.The hydrodynamic shift test bench adopts LabVIEW2018 software for control system program development.Through this, the control program can realize the function of shifting (manual or automatic), and it can also carry out various shifting tests, such as automatic shifting, shifting quality test, etc. [30].The control and acquisition part, as shown in Figures 12 and 13, represents the two-parameter automatic shifting program based on the vehicle speed and the degree of throttle opening.

Intelligent Shift Law Test Experiment
The test was conducted in manual shift mode, with forward gears shifting in the order from 0-1-2-3-4 and reverse gears shifting in the order from 0-1-2-3.The oil pressure curves of related clutches in each gear were collected in each gear and compared with the simulation results.
Figures 14-17 show the comparison between the test and simulation results of clutch shift oil pressure in each gear of the forward gears.It can be seen that the test oil pressure curves (solid line) of clutches KV and K4 in the process of shifting gears obviously show a four-segment upward trend, and the buffering effect is obvious.The oil pressure test curves of clutches K1, K2, and K3 show a three-segment upward trend, and the buffering effect is not as pronounced.Figures 18 and 19 show the comparison between the shift oil pressure test and simulation results for each gear of the reverse gears.It can be seen that the test oil pressure curve (realized) of the clutch KR during the shift process obviously shows a four-segment change, and the buffering effect is obvious.Clutch K1, K2, and K3 oil pressure rise faster, and the cushioning effect is not obvious.

Intelligent Shift Law Test Experiment
The test was conducted in manual shift mode, with forward gears shifting in the order from 0-1-2-3-4 and reverse gears shifting in the order from 0-1-2-3.The oil pressure curves of related clutches in each gear were collected in each gear and compared with the simulation results.
Figures 14-17 show the comparison between the test and simulation results of clutch shift oil pressure in each gear of the forward gears.It can be seen that the test oil pressure curves (solid line) of clutches KV and K4 in the process of shifting gears obviously show a four-segment upward trend, and the buffering effect is obvious.The oil pressure test curves of clutches K1, K2, and K3 show a three-segment upward trend, and the buffering effect is not as pronounced.Figures 18 and 19 show the comparison between the shift oil pressure test and simulation results for each gear of the reverse gears.It can be seen that the test oil pressure curve (realized) of the clutch KR during the shift process obviously shows a four-segment change, and the buffering effect is obvious.Clutch K1, K2, and K3 oil pressure rise faster, and the cushioning effect is not obvious.

Intelligent Shift Law Test Experiment
The test was conducted in manual shift mode, with forward gears shifting in the order from 0-1-2-3-4 and reverse gears shifting in the order from 0-1-2-3.The oil pressure curves of related clutches in each gear were collected in each gear and compared with the simulation results.
Figures 14-17 show the comparison between the test and simulation results of clutch shift oil pressure in each gear of the forward gears.It can be seen that the test oil pressure curves (solid line) of clutches KV and K4 in the process of shifting gears obviously show a four-segment upward trend, and the buffering effect is obvious.The oil pressure test curves of clutches K1, K2, and K3 show a three-segment upward trend, and the buffering effect is not as pronounced.Figures 18 and 19 show the comparison between the shift oil pressure test and simulation results for each gear of the reverse gears.It can be seen that the test oil pressure curve (realized) of the clutch KR during the shift process obviously shows a four-segment change, and the buffering effect is obvious.Clutch K1, K2, and K3 oil pressure rise faster, and the cushioning effect is not obvious.By comparing with the simulation results, it is found that the test and simulation curves of clutch shift oil pressure are very close, and the oil pressure in each stage of the clutch is basically the same.Because some factors, such as oil viscosity change and flow and other parameter influences, are ignored in the modeling process, the simulated oil pressure has a certain error in the shift time compared with the test data, but the error is not large.In summary, the simulation results are basically consistent with the test data, indicating the correctness of the simulation and providing a test basis for further improving the shift quality.

Conclusions
(1) Taking a 7-ton wheel loader as the research object, four modes of operation in the shoveling process of the loader are analyzed, and according to the operating conditions, three shift modes of the loader are proposed, namely, the power mode, the standard mode, and the economic mode.
(2) According to the characteristics of loader operation tasks and aiming at three shift modes, it is proposed that the power mode and standard mode are mainly based on operation conditions and transportation conditions, and the shift parameters are selected as throttle, speed, and workload coefficient so as to realize the design of the three-parameter dynamic shift law; the economic mode does not involve operation conditions, but only mainly focuses on transition driving, and the shift parameters are selected as throttle and speed so as to realize the design of optimal power and optimal economic two-parameter shift law design and give the shifting law surface.
(3) Aiming at the high complexity and nonlinear characteristics of the power transmission system of the loader, an intelligent shift control method based on an RBF neural network is proposed, and simulation tests and clutch shift oil pressure test tests are carried out.The test results show that the designed multi-mode and variable-parameter intelligent shift law of the loader is reasonable, and the shift recognition rate reaches 97.92%, which can solve the nonlinear and real-time variable automatic shift control problem.

Figure 1 .
Figure 1.Four common operating modes of loaders.

Figure 2 .
Figure 2. A schematic diagram of the V-type cycle operation of the loader.

Figure 1 .
Figure 1.Four common operating modes of loaders.
Wheel loader "V"model o

Figure 2 .
Figure 2. A schematic diagram of the V-type cycle operation of the loader.

Figure 6 .
Figure 6.Efficiency curve of torque converter at full throttle opening in economy mode.

Figure 6 .
Figure 6.Efficiency curve of torque converter at full throttle opening in economy mode.

Figure 7 .
Figure 7. Curve of optimal economy shift pattern in economy mode.

Figure 7 .
Figure 7. Curve of optimal economy shift pattern in economy mode.

Figure 8 .
Figure 8.First set of vehicle speed vs. engine speed data curves.Figure 8. First set of vehicle speed vs. engine speed data curves.

Figure 8 . 19 Figure 9 .
Figure 8.First set of vehicle speed vs. engine speed data curves.Figure 8. First set of vehicle speed vs. engine speed data curves.Actuators 2024, 13, x FOR PEER REVIEW 13 of 19

Figure 9 .
Figure 9. Second set of vehicle speed vs. engine speed data curves.

Actuators 2024 , 19 Figure 9 .
Figure 9. Second set of vehicle speed vs. engine speed data curves.

Figure 12 .
Figure 12.Control and acquisition section.

Figure 13 .
Figure 13.Two-parameter automatic speed change program.

Figure 12 .
Figure 12.Control and acquisition section.

Figure 13 .
Figure 13.Two-parameter automatic speed change program.

Figure 13 .
Figure 13.Two-parameter automatic speed change program.

Figure 19 .
Figure 19.R1-2 Clutch oil pressure curve comparison.By comparing with the simulation results, it is found that the test and simulation curves of clutch shift oil pressure are very close, and the oil pressure in each stage of the clutch is basically the same.Because some factors, such as oil viscosity change and flow

Figure 19 .
Figure 19.R1-2 Clutch oil pressure curve comparison.By comparing with the simulation results, it is found that the test and simulation curves of clutch shift oil pressure are very close, and the oil pressure in each stage of the clutch is basically the same.Because some factors, such as oil viscosity change and flow

Figure 19 .
Figure 19.R1-2 Clutch oil pressure curve comparison.By comparing with the simulation results, it is found that the test and simulation curves of clutch shift oil pressure are very close, and the oil pressure in each stage of the clutch is basically the same.Because some factors, such as oil viscosity change and flow

Table 1 .
Parameters of a certain type of wheel loader of 7 tons.

Table 2 .
Gear ratios and transmission efficiency of gears of the transmission.

Table 4 .
Full throttle/full load common output characteristics in loader power mode.

Table 5 .
Full throttle/full load traction characteristics in loader power mode.

Table 6 .
Loader power mode full load condition shift points.

Table 6 .
Loader power mode full load condition shift points.

Table 7 .
Optimal economy shift points in economy mode.

Table 8 .
Parameters of hydrodynamic transmission test stand.

Table 8 .
Parameters of hydrodynamic transmission test stand.