Event triggering model predictive control elevator speed tracking method based on learning

Abstract

The invention discloses a learning-based event triggering model predictive control elevator speed tracking method. The method comprises the steps of establishing a state model of a system, measuring the model, designing speed curves of different heights, initializing a control law, designing an event triggering strategy and designing a performance index. A neural network learning model is utilized to initialize a control law, corresponding performance indexes are designed, a better control law is obtained, and better tracking is realized. Compared with a standard MPC control method, the method has the advantages that the precision is guaranteed, the calculated amount is reduced, the flexibility is higher, and the applicability is wider.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent transportation system and building automation, and in particular relates to a learning-based event-triggered model predictive control elevator speed tracking method. Background Art

[0002] The advent of industrialization and modernization has led to the widespread installation of elevator systems in multi-story buildings. In the context of personnel transportation, the ride comfort of elevators has become a key issue. Speed ​​tracking curve control is an important link to ensure passenger comfort and system safety. Traditional elevator control systems often rely on pre-set speed curves, which is difficult to cope with nonlinear and complex dynamic environments in practice. Therefore, it is of great significance to study intelligent and adaptive control strategies. Model predictive control is widely used in the industrial field because it can handle multi-variable and multi-constraint problems and predict future actions through online optimization. However, traditional model predictive control (MPC) has challenges in computational complexity and real-time performance, especially in fast-response elevator systems. Supervised learning technology uses sample data to approximate MPC, which is used to intelligently initialize the control law and use historical data or online learning to provide a reasonable initialization control law, thereby reducing the computational burden in each optimization process and improving prediction accuracy and control effect. At the same time, event-triggered control is a control strategy that saves computing resources. Unlike time triggering, event triggering only updates the control signal when needed, thereby reducing the computational burden. Therefore, the event-triggered model predictive control based on learning initialization not only meets the growing demand for elevator riding comfort in modern cities, but also reflects the great potential of intelligent technology in improving the comfort of urban transportation and building facilities. Summary of the invention

[0003] In order to overcome the problems of scenario inapplicability, low precision and large amount of calculation of existing elevator comfort control strategies, and aiming at the problem of passenger riding comfort, the present invention proposes a learning-based event-triggered model predictive control elevator speed tracking method.

[0004] The technical solution adopted by the present invention to solve its technical problem is: According to a first aspect of an embodiment of the present application, a learning-based event-triggered model predictive control elevator speed tracking method is provided, comprising the following steps: (1) establishing dynamic mechanical equations on both sides of the elevator system, where the two sides of the elevator system include a counterweight side and a car side; (2) Based on the dynamic mechanical equation established in step (1), a state space model and a measurement model of the elevator system are established, and discretization operations are performed on the elevator system; (3) Designing speed curves for different floors; the speed curves include three stages: acceleration, constant speed, and deceleration; (4) Sampling data for different floors with different load-bearing conditions and initializing the control law through a neural network; (5) Based on the state space model of step (2), the speed curve of step (3), and the control law of step (4), design corresponding control performance indicators in the initialization phase and the online optimization phase to accurately control the speed of the elevator.

[0005] Furthermore, in step (1), the dynamic mechanical equation of the car side is expressed as: ; ; in, Indicates the quality of the car, represents the acceleration of the car, Indicates the speed of the rope on the car side, is the displacement of the rope on the car side, and Respectively represent the speed of the car and the traction sheave; and They represent the elastic coefficient and damping coefficient of the rope on the car side respectively; The dynamic mechanical equation on the counterweight side is expressed as: ; ; in, represents the mass of the counterweight, is the acceleration of the counterweight, Indicates the speed of the rope on the counterweight side, is the displacement of the rope on the counterweight side, and Represent the speed of the counterweight and traction sheave respectively; and represent the elastic coefficient and damping coefficient of the rope on the counterweight side respectively.

[0006] Furthermore, in step (2), the state space model of the elevator system is expressed as: ; in, Indicates the state of the elevator system at the next moment, Indicates the current state of the elevator system. represents the process noise, and is the moment of inertia and radius of the traction machine, It is the control input of the controlled system.

[0007] Furthermore, in step (3), the curve acceleration segment formula is expressed as: ; in, is a rate constant, Indicates time, Indicates the end time of the first stage of elevator operation. Indicates the running time of the second stage of elevator operation. Indicates the running time of the third stage of elevator operation. Indicates the running time of the fourth stage of elevator operation. is the speed of the elevator at the end of the first segment, is the acceleration constant of the second speed segment, It is the maximum operating speed of the elevator during the entire stage.

[0008] Furthermore, in step (4), data sampling is performed on different floors with different load bearings, and the control law is controlled by a neural network. Initialize; build a model for the optimal solution, where the input is the target position and the quality of the car , the output is the optimal solution , the model G is recorded as , then: ; Where M represents the total number of parameters. Indicates training status, Indicates that the input is the target position and car quality Model, Indicates the reference value, Represents the loss function.

[0009] Furthermore, the step (5) includes the following sub-steps: (5.1) In the initialization phase, the optimal control problem for tracking the entire speed curve is formulated: ; in, is the prediction time domain, It is The predicted state of the step, Indicates The status of the step, It is The predictive control input of the step, are the affine parameters of the closed-loop strategy, It is the first The speed of the step, represents the reference sequence, and R is a known matrix, represents the feedback gain matrix, represents the initial state, represents the system matrix; (5.2) An event-triggered strategy is proposed, using the initial control law to The state of the step is predicted, and the difference between the actual state and the predicted state is used to determine whether the control law needs to be transmitted. If it is triggered, the control law is re-performed with model predictive control. If it is not triggered, the initialized control law is used. The event triggering mechanism is as follows: ; in, Represents a binary number. When its value is 0, it means no triggering, and when it is 1, it means triggering. is a predefined threshold, represents the absolute tracking error, It is The actual status of the step, It is The predicted state of the step; (5.3) In the online optimization phase, according to the time step , two online model predictive control problems are introduced. When given the current state In the case of , solve the following optimization problem: ; in, is the decision variable, represents the prediction time domain, is the prediction interval in the online optimization phase, It is The predicted state of the step, It is The predictive control input of the step, It is the first The speed of the step, Representation control law The parameters after parameterization in the k+ith step; when , solve the following optimization problem: .

[0010] The second aspect of an embodiment of the present invention provides an elevator speed tracking device based on event-triggered model predictive control based on learning, including a memory and a processor, wherein the memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the above-mentioned elevator speed tracking method based on event-triggered model predictive control based on learning.

[0011] A third aspect of an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the program implements the above-mentioned learning-based event-triggered model predictive control elevator speed tracking method.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention firstly adopts a simplified experimental model to model the elevator system; then designs an ideal speed curve to ensure comfort, speed and safety that meet industry standards. The present invention proposes an event-triggered MPC control method, which is divided into two stages: initialization and online optimization; in the initialization stage, firstly, an MPC optimization problem with a long prediction range is designed, and different positions and different lift vehicle weights are sampled as features, and their optimal solutions are included in the network as labels; these data pairs are used to train a network with a designed loss function; in the online optimization stage, the proposed optimization problem has a shorter prediction period and a designed event triggering mechanism, thereby reducing the calculation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the elevator model in the present invention; Figure 2 It is the flow chart of the elevator operation in the present invention; Figure 3 is an ideal operating speed curve diagram of the elevator system in the present invention; Figure 4 Schematic diagram of the event triggering mechanism in the present invention; Figure 5 This is a speed tracking diagram of a 3-story elevator in an embodiment of the present invention; Figure 6 This is a diagram showing the position change of the elevator on the third floor in an embodiment of the present invention; Figure 7 Schematic diagram of the speed curve control input of a 3-story elevator in an embodiment of the present invention; Figure 8 This is a speed tracking diagram of a 20-story elevator in an embodiment of the present invention; Fig. 9 This is a diagram showing the position change of the 20-story elevator in an embodiment of the present invention; Fig.10 Schematic diagram of speed curve control input for a 20-story elevator in an embodiment of the present invention; Fig.11This is a calculation time comparison diagram in an embodiment of the present invention; Fig.12 Schematic diagram of the device of the present invention. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical scheme and advantages of the present invention more clear, the present invention is further described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0015] The present invention will be further described below in conjunction with the accompanying drawings.

[0016] Reference Figure 1 The present invention provides a learning-based event-triggered model predictive control elevator speed tracking method, comprising the following steps: The present invention mainly divides the operation control of the elevator into two stages, see Figure 2 The first stage is the process of passengers entering the elevator when the elevator is stopped, and the second stage is the online control stage when the elevator starts running, including intermediate acceleration and deceleration until it reaches the target floor. Among them, the elevator speed needs to be restricted to a certain extent in the online stage.

[0017] Step 1: In order to achieve precise control of the elevator, it is necessary to establish the dynamic mechanical equations on both sides of the elevator system and perform sampling and discretization operations. The two sides of the elevator system include the counterweight side and the car side.

[0018] The dynamic mechanical equation on the car side can be expressed as: (1) (2) in, Indicates the quality of the car, represents the acceleration of the car, Indicates the speed of the rope on the car side, is the displacement of the rope on the car side, and Represent the speed of the car and traction sheave respectively. and They represent the elastic coefficient and damping coefficient of the rope on the car side respectively.

[0019] The dynamic mechanical equation on the counterweight side can be expressed as: (3) (4) in, represents the mass of the counterweight, is the acceleration of the counterweight, Indicates the speed of the rope on the counterweight side, is the displacement of the rope on the counterweight side, and Represent the speed of the counterweight and traction sheave respectively. and represent the elastic coefficient and damping coefficient of the rope on the counterweight side respectively.

[0020] Step 2: Based on the dynamic mechanical equations on both sides of the elevator system obtained in step 2, further establish the state space model and measurement model of the elevator system.

[0021] (2.1) The state space model of the elevator system is established based on the dynamic mechanical equations of the elevator system: (5) in, Indicates the state of the elevator system at the next moment, Indicates the current state of the elevator system. represents the process noise, is the system state, p represents the displacement of the car, and Represent the parameters of the car and counterweight sides respectively. and is the displacement of the rope, , and Represents the speed of the car, counterweight and traction sheave respectively. and They represent the elastic coefficient and damping coefficient of the rope on the car and counterweight sides respectively. and are the rotational inertia and radius of the traction machine. is the electromagnetic torque of the traction machine, It is the control input of the controlled system.

[0022] (2.2) From the state model of the elevator system, it can be seen that the speed of the car is related to the speed of the traction wheel. In order to keep the running speed of the car within the comfort range of the human body, the present invention controls the speed of the elevator traction wheel. Therefore, the measurement model of the elevator system is established: (6) in, Indicates the measurement output, Indicates the elevator system status, represents the measurement noise.

[0023] (2.3) According to the above analysis, the system state space model is obtained as: (5) in, is the state of the controlled system, is the control input of the controlled system, To control the output of the controlled system, is the disturbance noise of the controlled system. is a known matrix of finite dimension.

[0024] (2.4) The control method proposed in this invention is aimed at discrete-time systems, so the system needs to be discretized.

[0025] Define the state matrix, control matrix and output matrix of the above continuous system as , the discretized system matrix is ​​obtained by the following method: (7) in, represents the state matrix of the discrete system, represents the control matrix of the discretized system, represents the discrete time constant, Indicates the sampling time.

[0026] Step 3: According to the established state space model, the purpose of the present invention is to control the speed of the elevator to meet the comfort conditions of the human body. It is necessary to control the car speed in the model to meet the appropriate conditions. Because the required running speed of the elevator is different at different floors, only under such conditions can the comfort requirements of the human body be met. Therefore, it is necessary to design speed curves for different floors, and use the established state space model to control the speed of the elevator to meet the speed curve of the corresponding floor, so as to meet the comfort requirements of passengers and ensure the comfort of riding. The process is as follows: See also Figure 3 Generally speaking, the ideal speed curve generally meets the three stages of acceleration, uniform speed and deceleration. Among them, the lower floors may not reach the maximum speed. At the same time, the acceleration of the elevator acceleration and deceleration stage should not be too large, so the speed shape of the initial acceleration and deceleration stage is roughly a parabola, the second section is a uniform acceleration straight line stage with a smaller acceleration, and the third section is a uniform speed stage. The subsequent deceleration stage is symmetrical with the acceleration stage. Therefore, according to this design concept, the present invention designs speed curves corresponding to different floors to meet the human comfort problem. According to the different elevator running distances, the ideal elevator running speed curve is proposed. , the curve acceleration segment formula is as follows: (8) in, is a rate constant, Indicates time, Indicates the end time of the first stage of elevator operation. Indicates the running time of the second stage of elevator operation. Indicates the running time of the third stage of elevator operation. Indicates the running time of the fourth stage of elevator operation. is the speed of the elevator at the end of the first segment, is the acceleration constant of the second speed segment, It is the maximum operating speed of the elevator during the entire stage.

[0027] For safety and comfort reasons, elevator acceleration and acceleration rate are strictly limited: (a) The maximum acceleration cannot exceed ; (b) The maximum acceleration cannot exceed .

[0028] The constraints of the elevator system state and the maximum output power of the motor are considered as follows: (9) in, Indicates the system status. Indicates system control, is within two bounded sets.

[0029] Step 4: After designing a reasonable speed curve, the purpose of the present invention is mainly to control the speed of the elevator to meet this speed curve, and the key to control lies in solving the control law. However, in actual process, elevator operation is a relatively fast process, so it is necessary to quickly calculate the control law at each moment. Therefore, considering the timeliness of elevator operation, it is necessary to first process the control law. The present invention uses the idea of ​​neural network to optimize control, specifically: sampling data at different floors with different load-bearing times, and using neural network to optimize the initial control law. Initialization is performed. Since the prediction horizon may be very long (especially when the target position of the car is high), it is expensive to repeatedly solve this optimization problem. Therefore, the present invention attempts to build a model for the optimal solution, where the input is the target position and the mass of the car (i.e. the mass of the car at this time ), the output is the optimal solution , the model G is recorded as : (10) Where M represents the total number of parameters. Indicates training status, Indicates that the input is the target position and car quality Model, Indicates the reference value, Represents the loss function.

[0030] In this model, given a pair , the optimal approximate solution can be calculated efficiently.

[0031] Step 5: By completing the establishment of the elevator model, the design of the speed curve, and the initialization of the control law, the speed of the elevator can be accurately controlled based on this information. The present invention completes the control by designing the corresponding control performance indicators in the initialization stage and the online optimization stage.

[0032] (5.1) In the initialization phase, the optimal control problem for tracking the entire speed curve is formulated: (11) in, is the prediction time domain, It is The predicted state of the step, Indicates The predicted state of the step, It is The predictive control input of the step, are the affine parameters of the closed-loop strategy, It is the first The speed of the step, represents the reference sequence, and R is a known matrix, represents the feedback gain matrix, represents the initial state, represents the system matrix.

[0033] (5.2) From the above-mentioned whole control process, it can be seen that in this project, a lot of resources may be consumed. In order to reduce the computing cost and alleviate the computing burden, the present invention proposes an event triggering strategy, see Figure 4 , using the control law initialized in step 4, The state of the step is predicted, and the difference between the actual state and the predicted state is used to determine whether the control law needs to be transferred. If it is triggered, the Model Predictive Control (MPC) operation is performed again. If it is not triggered, the initialized control law is used. Figure 4 middle, is the initialization control law, is the control law obtained by MPC calculation.

[0034] The specific event triggering mechanism is as follows: (12) in, Represents a binary number. It will not be triggered when it is equal to 0, and it will be triggered when it is equal to 1. is a predefined threshold (0.1-0.5 in this embodiment), represents the absolute tracking error, It is The actual status of the step, It is The predicted state of the step is used .

[0035] (5.3) Then the design of the online optimization phase is as follows: according to the time step , we introduce two online MPC problems. First, when When given the current state In the case of , the following optimization problem is called: (13) in, is the decision variable, Indicates the prediction time domain, upper right corner represents the transpose operation, is the prediction interval of the online optimization phase; It is The predicted state of the step, It is Step predictive control input; It is the first The speed of the step, represents the reference sequence, and R is a known constant matrix, represents the feedback gain matrix, represents the initial state, represents the system matrix; Represents the initialization control law The parameters after parameterization in the k+i step.

[0036] when , solve the following optimization problem: (14) In order to verify the effectiveness of the method designed by the present invention, a simulation was performed in an elevator system for verification.

[0037] The parameters used in the experiment are shown in Table 1: Table 1: Elevator model parameters parameter value Car quality 1600-3200[kg] Counterweight mass 2400[kg] Traction sheave radius 0.25[m] Rope stiffness coefficient on car side <![CDATA[5.03×10 3 [N / m]]]> Car side rope damping coefficient 844[Ns / m] Counterweight side rope stiffness coefficient <![CDATA[10×10 3 [N / m]]]> Damping coefficient of rope on counterweight side 2700[Ns / m] Maximum electromagnetic torque 800[N / m] Moment of Inertia <![CDATA[171.134[kg×m 2 ]]]> The other simulation parameters are set as follows: Sampling period , prediction time domain , , , , , Calculated via the algebraic Ricardi equation.

[0038] The results are as follows Figure 5-Figure 10 As shown, Figure 5 is the reference speed of the elevator at 3 floors and the actual speed of the elevator under standard MPC control and the method proposed in the present invention, Figure 6 is the change of car position, Figure 7 is the control law of the system. Figure 8 is the reference speed of the elevator at a height of 20 floors and the actual speed of the elevator under standard MPC control and the method proposed in the present invention, Fig. 9 is the change of car position, Fig.10 is the control law of the system, Fig.11 The figure shows the comparison of the calculation time of the proposed method and the standard MPC. It can be seen from the results that the speed of the proposed method can track the given reference value better at 3 floors and 20 floors, and compared with the standard MPC algorithm, the performance is also improved to a certain extent, and the required calculation time is greatly reduced. Therefore, the event-triggered model predictive control elevator speed tracking method proposed by the present invention has good tracking accuracy for the target and meets the comfort requirements of passengers.

[0039] See also Fig.12 An embodiment of the present invention provides an elevator speed tracking device based on learning event-triggered model predictive control, including one or more processors, which are used to implement an elevator speed tracking method based on learning event-triggered model predictive control in the above embodiment.

[0040] The embodiment of the elevator speed tracking device based on the learning event-triggered model predictive control of the present invention can be applied to any device with data processing capabilities, and the device with data processing capabilities can be a device or apparatus such as a computer. The device embodiment can be implemented through software, or through hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capabilities in which it is located reading the corresponding computer program instructions in the non-volatile memory into the internal memory for execution. From the hardware level, if Fig.12 As shown in the figure, it is a hardware structure diagram of any device with data processing capability where the event-triggered model prediction control elevator speed tracking device based on learning of the present invention is located, except Fig.12In addition to the processor, memory, network interface, and non-volatile memory shown, any device with data processing capabilities in which the apparatus in the embodiments is located may also include other hardware, generally based on the actual functions of the device with data processing capabilities, which will not be described in detail.

[0041] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0042] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0043] Correspondingly, an embodiment of the present invention further provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the learning-based event-triggered model predictive control elevator speed tracking method in the above embodiment is implemented.

[0044] The computer-readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be any device with data processing capability, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of any device with data processing capability and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0045] Those skilled in the art will readily appreciate other embodiments of the present application after considering the description and practicing the contents disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application. The description and examples are intended to be exemplary only.

[0046] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0047] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A learning-based event-triggered model predictive control elevator speed tracking method, characterized in that: The following steps are involved: (1) establishing dynamic mechanical equations on both sides of the elevator system, where the two sides of the elevator system include a counterweight side and a car side; (2) Based on the dynamic mechanical equation established in step (1), a state space model and a measurement model of the elevator system are established, and discretization operations are performed on the elevator system; (3) Designing speed curves for different floors; the speed curves include three stages: acceleration, constant speed, and deceleration; (4) Sampling data for different floors with different load-bearing conditions and initializing the control law through a neural network; (5) Based on the state space model of step (2), the speed curve of step (3), and the control law of step (4), design corresponding control performance indicators in the initialization phase and the online optimization phase to accurately control the speed of the elevator.

2. The method for elevator speed tracking based on event-triggered model predictive control according to claim 1, characterized in that: In step (1), the dynamic mechanical equation of the car side is expressed as: ； ； in, Indicates the quality of the car, represents the acceleration of the car, Indicates the speed of the rope on the car side, is the displacement of the rope on the car side, and Respectively represent the speed of the car and the traction sheave; and They represent the elastic coefficient and damping coefficient of the rope on the car side respectively; The dynamic mechanical equation on the counterweight side is expressed as: ； ； in, represents the mass of the counterweight, is the acceleration of the counterweight, Indicates the speed of the rope on the counterweight side, is the displacement of the rope on the counterweight side, and Represent the speed of the counterweight and traction sheave respectively; and represent the elastic coefficient and damping coefficient of the rope on the counterweight side respectively.

3. The method for elevator speed tracking based on event-triggered model predictive control according to claim 2, characterized in that: In step (2), the state space model of the elevator system is expressed as: ； in, Indicates the state of the elevator system at the next moment, Indicates the current state of the elevator system. represents the process noise, and is the moment of inertia and radius of the traction machine, It is the control input of the controlled system.

4. The method for elevator speed tracking based on event-triggered model predictive control according to claim 3, characterized in that: In step (3), the curve acceleration segment formula is expressed as: ； in, is a rate constant, Indicates time, Indicates the end time of the first stage of elevator operation. Indicates the running time of the second stage of elevator operation. Indicates the running time of the third stage of elevator operation. Indicates the running time of the fourth stage of elevator operation. is the speed of the elevator at the end of the first segment, is the acceleration constant of the second speed segment, It is the maximum operating speed of the elevator during the entire stage.

5. The method for elevator speed tracking based on event-triggered model predictive control according to claim 4, characterized in that: In step (4), data sampling is performed when the load-bearing capacity of different floors is different, and the control law is adjusted by a neural network. Initialize; build a model for the optimal solution, where the input is the target position and the quality of the car , the output is the optimal solution , the model G is recorded as , then: ； Where M represents the total number of parameters. Indicates training status, Indicates that the input is the target position and car quality Model, Indicates the reference value, represents the loss function.

6. The method for elevator speed tracking based on event-triggered model predictive control according to claim 5, characterized in that: The step (5) comprises the following sub-steps: (5.1) In the initialization phase, the optimal control problem for tracking the entire speed curve is formulated: ； in, is the prediction time domain, It is The predicted state of the step, Indicates The state of the step, It is The predictive control input of the step, are the affine parameters of the closed-loop strategy, It is the first The speed of the step, represents the reference sequence, and R is a known matrix, represents the feedback gain matrix, represents the initial state, represents the system matrix; (5.2) An event-triggered strategy is proposed, using the initial control law to The state of the step is predicted, and the difference between the actual state and the predicted state is used to determine whether the control law needs to be transmitted. If it is triggered, the control law is re-performed with model predictive control. If it is not triggered, the initialized control law is used. The event triggering mechanism is as follows: ； in, Represents a binary number. When its value is 0, it means no triggering, and when it is 1, it means triggering. is a predefined threshold, represents the absolute tracking error, It is The actual status of the step, It is The predicted state of the step; (5.3) In the online optimization phase, according to the time step , two online model predictive control problems are introduced. When given the current state In the case of , solve the following optimization problem: ； in, is the decision variable, represents the prediction time domain, is the prediction interval of the online optimization phase, It is The predicted state of the step, It is The predictive control input of the step, It is the first The speed of the step, Representation of control law The parameters after parameterization in the k+ith step; when , solve the following optimization problem: 。 7. A learning-based event-triggered model predictive control elevator speed tracking device, comprising a memory and a processor, characterized in that: The memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the learning-based event-triggered model predictive control elevator speed tracking method as described in any one of claims 1 to 6 above.

8. A computer-readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, the learning-based event-triggered model predictive control elevator speed tracking method described in any one of claims 1 to 6 is implemented.

Images