Intelligent asymptotic motion control method and system for hydraulic mechanical arm

Abstract

The invention discloses an intelligent asymptotic motion control method and system for a hydraulic mechanical arm. The method comprises the following steps: firstly, establishing a nonlinear mathematical model of a multi-degree-of-freedom series hydraulic mechanical arm system; secondly, estimating an endogenous uncertain item suffered by the system based on a multi-layer feedforward neural network; then, a robust control law for other uncertain items of the system is constructed, and an intelligent asymptotic motion controller of the hydraulic mechanical arm system is designed; designing a multilayer feedforward neural network updating law; and finally, parameters of the intelligent asymptotic motion controller of the hydraulic mechanical arm system are selected. According to the method, asymptotic tracking control of all the joints can be achieved on the premise that rotational inertia information of the hydraulic mechanical arm is not needed, and the tracking performance, adaptability and reliability of the system under complex working conditions are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electromechanical and hydraulic servo control, and in particular to an intelligent asymptotic motion control method and system for a hydraulic manipulator. Background Art

[0002] As high-precision electromechanical and hydraulic integrated equipment, hydraulic manipulators leverage the high power density of hydraulic transmission to achieve millimeter-level positioning accuracy and high-speed response under heavy loads. Their electro-hydraulic systems utilize highly nonlinear servo valves, composite seals, and multi-axis linkage technology for precise control. Consequently, hydraulic manipulators are increasingly being used in industry and engineering. However, the system faces significant nonlinear challenges: servo valve flow-pressure nonlinearity, servo valve hysteresis nonlinearity, oil elastic modulus temperature drift, and friction nonlinearity. Combined with time-varying factors such as changes in load inertia and oil viscosity temperature variations, achieving high-precision control presents significant challenges.

[0003] Control strategies for hydraulic manipulators are primarily categorized into traditional control methods and modern and intelligent control methods. In the traditional control field, PID (proportional-integral-derivative) control plays a prominent role in industrial control due to its simple structure, ease of implementation, and low cost. However, due to the strong nonlinear characteristics of hydraulic manipulator systems, traditional PID control faces limitations such as insufficient bandwidth, parameter sensitivity, and significant overshoot. Therefore, modern and intelligent control plays a crucial role in addressing these control issues. Current modern and intelligent control methods include adaptive robust control, sliding mode control, and artificial intelligence control. However, further research is needed to ensure that hydraulic manipulator systems achieve asymptotic tracking performance (i.e., joint angle tracking errors tend to zero as time approaches infinity) when faced with various model uncertainties.

[0004] In summary, the shortcomings of existing hydraulic manipulator control are mainly as follows:

[0005] 1. Due to the inherently complex nature of hydraulic manipulator systems, high-performance control of hydraulic manipulators faces numerous challenges. First, the strong nonlinear characteristics of hydraulic manipulator systems, such as the flow-pressure, deadband, and friction nonlinearities of the hydraulic servo valves, make accurate modeling extremely difficult. Second, the system parameters of hydraulic manipulators exhibit significant time-varying characteristics. Furthermore, the strong coupling between the hydraulic actuators of the joints in multi-degree-of-freedom manipulators further complicates control. These complexities make it difficult for conventional control algorithms to meet the control requirements of high-precision hydraulic manipulators, particularly to achieve asymptotic motion control performance.

[0006] 2. As the application of hydraulic manipulators continues to expand, the requirements for their control performance are also increasing. In particular, the control system must be able to adapt to uncertain and time-varying objects and environments. In particular, the moment of inertia of the hydraulic manipulator has a decisive influence on its motion control performance. The dynamic relationship between moment of inertia and angular acceleration directly determines the system response speed. However, in practice, it is difficult to accurately obtain the moment of inertia of the hydraulic manipulator. Summary of the Invention

[0007] The object of the present invention is to provide an intelligent asymptotic motion control method and system for a hydraulic manipulator with strong adaptability to working conditions, high reliability and high control accuracy.

[0008] The technical solution adopted by the present invention to solve the above problems is:

[0009] An intelligent asymptotic motion control method for a hydraulic manipulator comprises the following steps:

[0010] Step 1: Establish a nonlinear mathematical model of a multi-degree-of-freedom serial hydraulic manipulator system;

[0011] Step 2: Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system;

[0012] Step 3: Construct a robust control law for other uncertainties in the system and design an intelligent asymptotic motion controller for the hydraulic manipulator system;

[0013] Step 4: Design the update law of the multi-layer feedforward neural network;

[0014] Step 5: Select the parameters of the intelligent asymptotic motion controller for the hydraulic manipulator system.

[0015] Furthermore, the nonlinear mathematical model of the multi-degree-of-freedom serial hydraulic manipulator system is established as described in step 1, specifically as follows:

[0016] Step 1.1, define the state variables of the multi-degree-of-freedom serial hydraulic manipulator system as:

[0017]

[0018] Where, represents the joint angular displacement, n is the degree of freedom of the system, They represent the effective area of the rodless and rod-shaped chambers of the hydraulic cylinder or the displacement of the hydraulic motor. Indicates the pressure of the rodless cavity and rod cavity of the hydraulic cylinder or the pressure of the two cavities of the hydraulic motor;

[0019] The state space form of the nonlinear model of the multi-degree-of-freedom serial hydraulic manipulator system is expressed as:

[0020]

[0021] in

[0022]

[0023] Where, represents the unknown moment of inertia, represents the centripetal-Coriolis matrix, represents the gravity vector, represents a nonlinear function, represents the remaining disturbances that may exist in the system, H β is the elastic modulus of hydraulic oil, and are the volumes of the two chambers of the hydraulic actuator, and Indicates the flow from the hydraulic servo valve to the hydraulic actuator cavity and from the hydraulic actuator cavity to the hydraulic servo valve, is the internal leakage coefficient of the hydraulic actuator, and is the disturbance suffered by the system, and They represent the flow coefficient, orifice area gradient and electrical gain of the hydraulic servo valve respectively. Indicates the density of hydraulic oil, is the displacement of the hydraulic actuator, is the control input voltage of the system, represents the gain matrix, and Respectively represent the oil supply pressure and oil return pressure;

[0024] Step 1.2, control target: for the position command trajectory An intelligent asymptotic motion controller is designed for a multi-degree-of-freedom serial hydraulic manipulator system, which enables the hydraulic manipulator system to achieve asymptotic motion control under working conditions where the system's moment of inertia information is unknown and the system is subject to various modeling uncertainties.

[0025] Step 1.3, Definition: For any vector There is a constant and functions To satisfy:

[0026]

[0027] Setting 1: For any vector All satisfy the following equations:

[0028]

[0029] Setting 2: The position command signal χ that the system expects to track 1d is first-order continuously differentiable, and χ 1d and its first-order derivative are bounded;

[0030] Setting 3: Perturbations to the system and are all bounded;

[0031] Setting 4: Setting represents the estimated value of , represents the estimation error of .

[0032] Furthermore, the design described in step 2 is based on a multi-layer feedforward neural network uncertainty estimator to estimate the endogenous uncertainty in the system, as follows:

[0033] For any continuous unknown nonlinear function and It satisfies:

[0034]

[0035] Where, and represents the ideal weights of a multilayer feedforward neural network; N1, N2, N3, and N4 are the number of neurons; ξ2(·) and ξ3(·) represent activation functions; and represents the input vector; and represents the function reconstruction error;

[0036] Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the intrinsic uncertainty in the system. and Make an estimate:

[0037]

[0038] Furthermore, the robust control law for other uncertainties of the system is constructed as described in step 3, and the intelligent asymptotic motion controller of the hydraulic manipulator system is designed, as follows:

[0039] Define z1 = χ1 - χ 1d The tracking error and control errors z2 and z3 of the system are:

[0040] z2=χ2-σ1,z3=χ3-σ 2c

[0041] Where, represents the virtual control law, is the filtered value of the virtual control law σ2, obtained by the following filter:

[0042]

[0043] Where, is an adjustable positive gain, represents a positive time-varying function and its integral upper limit is a bounded positive constant; β2 is The upper bound of the estimated value Real-time update is performed via:

[0044]

[0045] Where, is a positive design parameter, is the filtering error of σ2;

[0046] Design virtual control laws σ1, σ2 and actual control law u as:

[0047]

[0048] Where, and is an adjustable positive gain, and They all represent positive time-varying functions and their upper limits of integration are all bounded positive constants. and Update it with the following formula:

[0049]

[0050] Where p2 and p3 are positive design parameters.

[0051] Furthermore, the update law of the multi-layer feedforward neural network is designed as described in step 4, which is as follows:

[0052] The weight parameters of the multi-layer feedforward neural network are updated by the following formula:

[0053]

[0054] Where Proj(·) is the continuous projection mapping function, as well as represents the positive diagonal adaptive law matrix, γ W2 (t), γ V2 (t), γ W3 (t) and γ V3 (t) are all positive time-varying functions and the upper bounds of the integrals are all positive constants.

[0055] Furthermore, the parameters of the intelligent asymptotic motion controller for the hydraulic manipulator system are selected in step 5, specifically as follows:

[0056] Select the initial value of the neural network weight parameter, the adaptive law matrix Positive time-varying function γ W2 (t), γ V2 (t), γ W3 (t), γ V3 (t) and adjust other controller parameters including ω c2 ,α2,φ(t),k1,k2,k3,p2,p3, and The value of is used to achieve the asymptotic motion control goal of each joint of the hydraulic manipulator system.

[0057] An intelligent asymptotic motion control system for a hydraulic manipulator, which is used to implement the intelligent asymptotic motion control method for a hydraulic manipulator, includes first to fifth modules, wherein:

[0058] The first module establishes a nonlinear mathematical model of a multi-degree-of-freedom serial hydraulic manipulator system;

[0059] The second module designs an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system;

[0060] The third module builds a robust control law for other uncertainties in the system and designs an intelligent asymptotic motion controller for the hydraulic manipulator system.

[0061] The fourth module is to design the update law of multi-layer feedforward neural network;

[0062] The fifth module selects the parameters of the intelligent asymptotic motion controller of the hydraulic manipulator system.

[0063] A mobile terminal comprises a memory, a processor and a computer program stored in the memory and operable on the processor. When the processor executes the program, the intelligent asymptotic motion control method for a hydraulic mechanical arm is implemented.

[0064] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the intelligent asymptotic motion control method for a hydraulic mechanical arm.

[0065] Compared with the prior art, the present invention has the following significant advantages:

[0066] (1) A multi-layer feedforward neural network is used to estimate and feedforward compensate the intrinsic uncertainty of the hydraulic manipulator system, thereby improving the anti-disturbance capability of the system;

[0067] (2) The system can realize asymptotic tracking control of each joint without requiring information on the moment of inertia of the hydraulic manipulator, thereby improving the adaptability and reliability of the hydraulic manipulator system under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a flow chart of the intelligent asymptotic motion control method and system for hydraulic manipulators of the present invention.

[0069] Figure 2 It is a structural principle diagram of the multi-degree-of-freedom serial hydraulic mechanical arm system driven by a hydraulic actuator in the present invention.

[0070] Figure 3 It is a graph showing the change of the tracking performance of the system over time under the action of the controller in an embodiment of the present invention.

[0071] Figure 4 In the embodiment of the present invention, the parameters under the action of the controller are A graph showing changes over time.

[0072] Figure 5 It is a graph showing how the unknown parameters of the system joint 1 change over time under the action of the controller in an embodiment of the present invention.

[0073] Figure 6 It is a graph showing how the unknown parameters of the system joint 2 change over time under the action of the controller in an embodiment of the present invention.

[0074] Figure 7 is a graph showing how the uncertainty estimation performance of the system joint 1 changes over time under the action of the controller in an embodiment of the present invention.

[0075] Figure 8 is a graph showing how the uncertainty estimation performance of the system joint 2 changes over time under the action of the controller in an embodiment of the present invention.

[0076] Figure 9 is a graph showing how the input voltage of the controller changes over time in an embodiment of the present invention. DETAILED DESCRIPTION

[0077] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0078] Combine Figure 1 The present invention provides an intelligent asymptotic motion control method and system for a hydraulic manipulator, comprising the following steps:

[0079] Step 1: Establish a nonlinear mathematical model of the multi-degree-of-freedom series hydraulic manipulator system, such as Figure 2 As shown, the details are as follows:

[0080] Step 1.1, define the system state variables as in represents the joint angular displacement (n is the degree of freedom of the system), They represent the effective area of the rodless and rod-shaped chambers of the hydraulic cylinder or the displacement of the hydraulic motor. Indicates the pressure of the rodless cavity and rod cavity of the hydraulic cylinder or the pressure of the two cavities of the hydraulic motor;

[0081] Then the state space form of the nonlinear model of the system is:

[0082]

[0083] In formula (1):

[0084]

[0085] In formula (2), represents the unknown moment of inertia, represents the centripetal-Coriolis matrix, represents the gravity vector, represents a nonlinear function, represents the remaining disturbances that may exist in the system, H β is the elastic modulus of hydraulic oil, and are the volumes of the two chambers of the hydraulic actuator, and Indicates the flow from the hydraulic servo valve to the hydraulic actuator cavity and from the hydraulic actuator cavity to the hydraulic servo valve, is the internal leakage coefficient of the hydraulic actuator, and is the disturbance suffered by the system, and They represent the flow coefficient, orifice area gradient and electrical gain of the hydraulic servo valve respectively. Indicates the density of hydraulic oil, is the displacement of the hydraulic actuator, is the control input voltage of the system, represents the gain matrix, and Respectively represent the oil supply pressure and oil return pressure;

[0086] Step 1.2, control target: for the position command trajectory An intelligent asymptotic motion controller is designed for a multi-degree-of-freedom serial hydraulic manipulator system, which enables the hydraulic manipulator system to achieve asymptotic motion control under working conditions where the system's moment of inertia information is unknown and the system is subject to various modeling uncertainties.

[0087] Step 1.3, Setting 1: The position command signal x that the system expects to track 1d is first-order continuously differentiable, and χ 1d and its first-order derivative are bounded;

[0088] Setting 2: The system is subjected to a disturbance and are all bounded;

[0089] Setting 3: Setting represents the estimated value of , represents the estimation error of .

[0090] Step 2: Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system, as follows:

[0091] Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the intrinsic uncertainty in the system. and Make an estimate:

[0092]

[0093] In formula (3), and represents the ideal weights of a multilayer feedforward neural network; N1, N2, N3, and N4 are the number of neurons; ξ2(·) and ξ3(·) represent activation functions; and represents the input vector;

[0094] Step 3: Construct a robust control law for other uncertainties in the system and design an intelligent asymptotic motion controller for the hydraulic manipulator system, as follows:

[0095] Design of virtual control laws and And the actual control law u is:

[0096]

[0097] In formula (4), is an adjustable positive gain, and All of them represent positive time-varying functions and their upper limits of integration are bounded positive constants, and:

[0098]

[0099] In formula (5), z1 is the tracking error of the system, z2 and z3 are the control errors of the system, is the filtered value of the virtual control law σ2, is the filtering error of σ2, is an adjustable positive gain, β2 is The upper bound of , p2 and p3 are both positive design parameters, φ(t) is a positive time-varying function and its integral upper limit is a bounded positive constant.

[0100] Step 4: Design the update law of the multi-layer feedforward neural network as follows:

[0101] The weight parameters of the multi-layer feedforward neural network can be updated by the following formula:

[0102]

[0103] In formula (6), Proj(·) is the continuous projection mapping function, as well as represents the positive diagonal adaptive law matrix, γ W2 (t), γ V2 (t), γ W3 (t) and γ V3 (t) are all positive time-varying functions and their upper bounds of integration are all constants.

[0104] Step 5: Select the initial values of the neural network weight parameters, the adaptive law matrix, and the controller parameters, as follows:

[0105] Properly select the initial values of the neural network weight parameters and the adaptive law matrix Positive time-varying function (γ W2 (t), γ V2 (t), γ W3 (t), γ V3 (t)) and adjust other controller parameters ω c2 ,α2,φ(t),k1,k2,k3,p2,p3, and The value of can achieve the asymptotic motion control goal of each joint of the hydraulic manipulator system.

[0106] An intelligent asymptotic motion control system for a hydraulic manipulator, which is used to implement the intelligent asymptotic motion control method for a hydraulic manipulator, includes first to fifth modules, wherein:

[0107] The first module establishes a nonlinear mathematical model of a multi-degree-of-freedom serial hydraulic manipulator system;

[0108] The second module designs an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system;

[0109] The third module builds a robust control law for other uncertainties in the system and designs an intelligent asymptotic motion controller for the hydraulic manipulator system.

[0110] The fourth module is to design the update law of multi-layer feedforward neural network;

[0111] The fifth module selects the parameters of the intelligent asymptotic motion controller of the hydraulic manipulator system.

[0112] A mobile terminal comprises a memory, a processor and a computer program stored in the memory and operable on the processor. When the processor executes the program, the intelligent asymptotic motion control method for a hydraulic mechanical arm is implemented.

[0113] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the intelligent asymptotic motion control method for a hydraulic mechanical arm.

[0114] Example

[0115] This embodiment uses a two-degree-of-freedom tandem hydraulic manipulator system to test and verify the intelligent asymptotic motion control method provided by the present invention. The first connecting rod of the hydraulic manipulator is driven by a single-rod hydraulic cylinder, and the other connecting rod is driven by a dual-blade hydraulic motor. For simplicity, hinges Z1 and Z2 are assumed to be on the same horizontal line. According to rigid body dynamics analysis, for the tandem hydraulic manipulator, M(θ), and G(θ) can be described as:

[0116]

[0117] In formula (7), the expressions of each part are as follows:

[0118]

[0119] In formula (8), θ j (j=1, 2), S j 、M Ij and M mj represent the angular position, length, moment of inertia and mass of the jth connecting rod respectively; M L Indicates load mass; S cj Represents the distance from the center of gravity of the j-th link to the rotation axis Z j length; g represents the acceleration due to gravity; set H c =[H c11 , H c22 ] T , A a =diag{A a11 , A a22}, A b =diag{A b11 , A b22}, P a =[P a1 , P a2] T , P b =[P b1 , P b2 ] T , χ1=[χ 11 , χ 12 ] T , χ2=[χ 21 , χ 22 ] T , χ3=[χ 31 , χ 32 ] T , F2=[F 21 , F 22 ] T , F3=[F 31 , F 32 ] T , β2=[β 21 , β 22 ] T ,ρ2=[ρ 21 ,ρ 22 ] T and ρ3=[ρ 31 ,ρ 32 ] T ,definition And H Q =diag{H Q11 ,H Q22};

[0120] The parameters of the two-degree-of-freedom series hydraulic manipulator system are: M m1 =10kg, M m2 =8kg, M I1 =3.2kg·m 2 , M I2 =1.8kg·m 2 , M L =20kg, S c1 =0.3m, S c2 =0.2m, S1=0.65m, S2=0.4m, H β =1.275×10 9 Pa, P s =4×10 6 Pa, P r =0Pa, A a1 =1.3×10 -3 m 2 , A b1 =1.131×10 -4 m 2 , H c =diag{2.9×10 -12 ,2.9×10-12}m 3 / s / Pa, A a2 =1.24×10 -4 rad / s,A b2 =1.24×10 -4 rad / s, the initial value of the system is χ1(0)=[χ 11 (0),χ 12 (0)] T =[70°,10°] T , the joint position command that the system expects to track is χ 1d =[χ 11d ,χ 12d ] T =[20sin(1.25t)(1–e –0.5t )+100°,–30cos(2.2t)(1–e –0.5t )–10°] T , and set F f (χ2) = [30tanh(0.82χ 21 )–30tanh(0.48χ 21 )+30tanh(0.78χ 21 )+30χ 21 , 30tanh(0.82χ 22 )–30tanh(0.48χ 22 )+30tanh(0.78χ 22 )+30χ 22 ] T , F d (t,χ1,χ2)=[0.8χ 11 χ 21 +2.5χ 21 cos(χ 12 )+1.8χ 22 cos(χ 12 )+190sin(t),0.8χ 12 χ 22 +2.5χ 21 +1.8χ 22 +160sin(t)] T , Q d1 (t,χ1,χ2,P a ,P b )=[1.1×10 4 χ 11 χ 21 +7×10 -8 (P a1 –P b1 )2 +1.1×10 6 sin(t),9×10 4 χ 11 χ 21 +2×10 -8 (P a1 –P b1 ) 2 +7.8×10 6 sin(t)] T , Q d2 (t,χ1,χ2,P a ,P b )=[1.1×10 4 χ 12 χ 22 +7×10 -8 (P a2 –P b2 ) 2 +1.1×10 6 sin(t),9×10 4 χ 12 χ 22 +2×10 -8 (P a2 –P b2 ) 2 +7.8×10 6 sin(t)] T .

[0121] Controller design parameters:

[0122] After continuous adjustment, the controller design parameters are selected as k1=diag{1000,510}, k2=diag{800,510}, k3=diag{500,310}, ω o =diag{600,600}, ξ2(·)=tanh(·), ξ3(·)=tanh(·), γ W2 (t) = γ V2 (t) = γ W3 (t) = γ V3 (t) = 5e -0.5t ,ω c2 =diag{1×10 -4 ,1×10 -4}, α2=diag{5×10 2 ,5×10 2}, p2 = diag{2×10 3 ,1×10 3}, p3 = diag{5×102 ,5×10 2},

[0123] Controller effect: Figure 3 This is a graph showing how the tracking performance of the two joints of the system changes over time under the action of the controller designed by the present invention. It can be seen from the graph that under the action of the controller designed by the present invention, the system achieves a higher tracking accuracy, thereby verifying the effectiveness of the controller designed by the present invention.

[0124] Figure 4 、 Figure 5 and Figure 6 It is the unknown parameter of the system under the action of the controller designed by the present invention The curve graph changes with time. It can be seen from the figure that under the action of the controller designed by the present invention, they are all bounded and eventually fluctuate around a certain value.

[0125] Figure 7 and Figure 8 This is a graph showing how the estimation performance of the system uncertainty under the controller designed by the present invention changes over time. It can be seen from the graph that they are all bounded and eventually fluctuate around a certain value, thereby being able to effectively estimate the intrinsic uncertainty terms in the system.

[0126] Figure 9 This is a graph showing the change of the control input voltage of the controller designed by the present invention over time. It can be seen from the graph that the control input signal obtained by the present invention is continuous, diffusible and bounded, which is conducive to application in practical engineering.

[0127] The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent asymptotic motion control method for a hydraulic manipulator, characterized in that: The following steps are involved: Step 1: Establish a nonlinear mathematical model of a multi-degree-of-freedom serial hydraulic manipulator system; Step 2: Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system; Step 3: Construct a robust control law for other uncertainties in the system and design an intelligent asymptotic motion controller for the hydraulic manipulator system; Step 4: Design the update law of the multi-layer feedforward neural network; Step 5: Select the parameters of the intelligent asymptotic motion controller for the hydraulic manipulator system.

2. The intelligent asymptotic motion control method for a hydraulic manipulator according to claim 1, characterized in that: The nonlinear mathematical model of the multi-degree-of-freedom serial hydraulic manipulator system described in step 1 is as follows: Step 1.1, define the state variables of the multi-degree-of-freedom serial hydraulic manipulator system as: Where, represents the joint angular displacement, n is the degree of freedom of the system, They represent the effective area of the rodless and rod-shaped chambers of the hydraulic cylinder or the displacement of the hydraulic motor. Indicates the pressure of the rodless cavity and rod cavity of the hydraulic cylinder or the pressure of the two cavities of the hydraulic motor; The state space form of the nonlinear model of the multi-degree-of-freedom serial hydraulic manipulator system is expressed as: in: Where, represents the unknown moment of inertia, represents the centripetal-Coriolis matrix, represents the gravity vector, represents a nonlinear function, represents the remaining disturbances that may exist in the system, H β is the elastic modulus of hydraulic oil, and are the volumes of the two chambers of the hydraulic actuator, and Indicates the flow from the hydraulic servo valve to the hydraulic actuator cavity and from the hydraulic actuator cavity to the hydraulic servo valve, is the internal leakage coefficient of the hydraulic actuator, and is the disturbance suffered by the system, and They represent the flow coefficient, orifice area gradient and electrical gain of the hydraulic servo valve respectively. Indicates the density of hydraulic oil, is the displacement of the hydraulic actuator, is the control input voltage of the system, represents the gain matrix, and Respectively represent the oil supply pressure and oil return pressure; Step 1.2, control target: for the position command trajectory An intelligent asymptotic motion controller is designed for a multi-degree-of-freedom serial hydraulic manipulator system, which enables the hydraulic manipulator system to achieve asymptotic motion control under working conditions where the system's moment of inertia information is unknown and the system is subject to various modeling uncertainties. Step 1.3, Definition: For any vector There is a constant and functions To satisfy: Setting 1: For any vector All satisfy the following equations: Setting 2: The position command signal χ that the system expects to track 1d is first-order continuously differentiable, and χ 1d and its first-order derivative are bounded; Setting 3: Perturbations to the system and are all bounded; Setting 4: Setting represents the estimated value of , represents the estimation error of .

3. The intelligent asymptotic motion control method for a hydraulic manipulator according to claim 1, characterized in that: The design described in step 2 is based on a multi-layer feedforward neural network uncertainty estimator to estimate the endogenous uncertainty in the system, as follows: For any continuous unknown nonlinear function and It satisfies: Where, and represents the ideal weights of a multilayer feedforward neural network; N1, N2, N3, and N4 are the number of neurons; ξ2(·) and ξ3(·) represent activation functions; and represents the input vector; and represents the function reconstruction error; Design an uncertainty estimator based on a multi-layer feedforward neural network to estimate the intrinsic uncertainty in the system. and Make an estimate:

4. The intelligent asymptotic motion control method for a hydraulic manipulator according to claim 1, characterized in that: Step 3 describes the construction of a robust control law for other uncertainties in the system and the design of an intelligent asymptotic motion controller for the hydraulic manipulator system, as follows: Define z1 = χ1 - χ 1d The tracking error and control errors z2 and z3 of the system are: z2=χ2-σ1,z3=χ3-σ 2c Where, represents the virtual control law, is the filtered value of the virtual control law σ2, obtained by the following filter: Where, is the adjustable positive gain, represents a positive time-varying function whose upper limit of integration is a bounded positive constant; β2 is The upper bound of the estimated value Real-time update is performed via: Where, is a positive design parameter, is the filtering error of σ2; Design virtual control laws σ1, σ2 and actual control law u as: Where, and is the adjustable positive gain, and They all represent positive time-varying functions and their upper limits of integration are all bounded positive constants. and Update it with the following formula: Where p2 and p3 are positive design parameters.

5. The intelligent asymptotic motion control method for a hydraulic manipulator according to claim 1, characterized in that: The update law of the multi-layer feedforward neural network described in step 4 is designed as follows: The weight parameters of the multi-layer feedforward neural network are updated by the following formula: Where Proj(·) is the continuous projection mapping function, γ W2 , Υ V2 , Υ W3 and Y V3 represents the positive diagonal adaptive law matrix, γ W2 (t), γ V2 (t), γ W3 (t) and γ V3 (t) are all positive time-varying functions and the upper bounds of the integrals are all positive constants.

6. The intelligent asymptotic motion control method for a hydraulic manipulator according to claim 1, characterized in that: The parameters of the intelligent asymptotic motion controller for the hydraulic manipulator system described in step 5 are as follows: Select the initial value of the neural network weight parameter, the adaptive law matrix Y W2 , Υ V2 , γ W3 , γ V3 , positive time-varying function γ W2 (t), γ V2 (t), γ W3 (t), γ V3 (t) and adjust other controller parameters including ω c2 ,α2,φ(t),k1,k2,k3,p2,p3, and The value of is used to achieve the asymptotic motion control goal of each joint of the hydraulic manipulator system.

7. An intelligent asymptotic motion control system for a hydraulic manipulator, characterized in that: The system is used to implement the intelligent asymptotic motion control method for a hydraulic manipulator according to any one of claims 1 to 6, and the system includes a first module to a fifth module, wherein: The first module establishes a nonlinear mathematical model of a multi-degree-of-freedom serial hydraulic manipulator system; The second module designs an uncertainty estimator based on a multi-layer feedforward neural network to estimate the endogenous uncertainty in the system; The third module builds a robust control law for other uncertainties in the system and designs an intelligent asymptotic motion controller for the hydraulic manipulator system. The fourth module is to design the update law of multi-layer feedforward neural network; The fifth module selects the parameters of the intelligent asymptotic motion controller of the hydraulic manipulator system.

8. A mobile terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the intelligent asymptotic motion control method for a hydraulic manipulator according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the intelligent asymptotic motion control method for a hydraulic manipulator according to any one of claims 1 to 6 are implemented.

Images