Model predictive control calculation method and device of suspension system

Abstract

The invention relates to the technical field of suspension control, in particular to a model prediction control calculation method and device for a suspension system, and the method comprises the steps: selecting at least one group of existing model prediction control calculation results according to the state observation information of a target suspension system and a mathematical model of the target suspension system; calculating a simplified constraint set of model predictive control of the mathematical model by using at least one group of existing model predictive control calculation results and state observation information, and constructing a simplified quadratic programming problem based on the simplified constraint set and the mathematical model; and solving the simplified quadratic programming problem to obtain an optimal solution result of the mathematical model, and determining an actual control quantity of the target suspension system according to the optimal solution result of the mathematical model. According to the method and the device, the redundancy constraint can be deleted through the existing online trajectory calculation result or the partial solution of the offline calculation model prediction control, so that the solution calculation speed of the model prediction control is effectively improved under the condition that the optimal solution is not changed.

Description

Technical Field

[0001] The present application relates to the technical field of suspension control, and in particular to a model predictive control calculation method and device for a suspension system. Background Art

[0002] In the related art, the traditional suspension control method of maglev trains is to linearize at the equilibrium point and use PID and other methods to design the controller. However, the maglev system is a complex nonlinear system, and the traditional control method may cause unstable train operation and even serious safety accidents under high-speed conditions. With the rapid development of modern control theory, more advanced control methods are applied to the suspension control of maglev trains. For example, adaptive control, a control method of this nonlinear model has a certain anti-interference ability; or linear quadratic regulator, this method is a linear control method, but this method is difficult to deal with the nonlinearity of the suspension system.

[0003] In recent years, model predictive control (MPC) has made great progress and has been successfully applied in the fields of unmanned aerial vehicles, rail transportation, power systems, etc., and the application of model predictive control methods in suspension control has also achieved good control effects. The mechanism of model predictive control is: at each sampling moment, based on the state observation information of the current controlled object, combined with the mathematical model of the controlled system, an open-loop optimization problem in a finite time domain is solved online, and the first element of the obtained optimal control sequence is applied to the controlled object. At the next sampling moment, the above process is repeated. Since model predictive control requires the optimization problem to be solved online at each moment, it consumes certain computing and storage resources, which limits the application of this control method. For this reason, a large number of algorithms have been proposed to further accelerate the solution of model predictive control. For example, the active set method is used to iteratively search for the active set of the optimization problem, and the optimization problem is converted into a quadratic programming problem with equality constraints; the interior point method is used to solve the MPC problem, but this method is difficult to hot start; the first-order method is used to gradually search for the optimal solution through the gradient descent method. These methods are

[0004] However, the methods in related technologies are all committed to improving the solution speed of model predictive control by utilizing the special structure of MPC, ignoring the role of historical solution data. For example, for trains on a circular track, if the previous solution results can be utilized, the calculation speed of MPC is expected to be further improved. However, the solution of model predictive control is still difficult and difficult to apply to embedded systems, which needs to be solved urgently. Summary of the invention

[0005] The present application provides a model predictive control calculation method and device for a suspension system to solve the problems that the methods in the related technologies are all committed to improving the solution speed of the model predictive control by utilizing the special structure of MPC, ignoring the role of historical solution data, making it difficult to further improve the calculation speed of MPC, and the solution of model predictive control is still difficult and difficult to apply to embedded systems.

[0006] The first aspect of the present application provides a model predictive control calculation method for a suspension system, characterized in that it includes the following steps: screening at least one group of existing model predictive control calculation results according to state observation information of a target suspension system and a mathematical model of the target suspension system; calculating a simplified constraint set of the model predictive control of the mathematical model using the at least one group of existing model predictive control calculation results and the state observation information, and constructing a simplified quadratic programming problem based on the simplified constraint set and the mathematical model; solving the simplified quadratic programming problem to obtain an optimized solution result of the mathematical model, so as to determine the actual control quantity of the target suspension system according to the optimized solution result of the mathematical model.

[0007] Optionally, in one embodiment of the present application, the screening of at least one group of existing model predictive control calculation results based on the state observation information of the target suspension system and the mathematical model of the target suspension system includes: obtaining online trajectory calculation results and / or offline calculation results of the mathematical model based on the state observation information; and screening at least one group of existing model predictive control calculation results from the online trajectory calculation results and / or the offline calculation results.

[0008] Optionally, in one embodiment of the present application, the method of calculating the simplified constraint set of the model predictive control of the mathematical model using the at least one set of existing model predictive control calculation results and the state observation information includes: when calculating the model predictive control calculation results of the mathematical model online, using the offline calculation results to delete the target constraints to calculate the simplified constraint set of the model predictive control of the mathematical model.

[0009] Optionally, in one embodiment of the present application, the at least one group of existing model predictive control calculation results includes a state, an active set, and an optimal solution corresponding to the at least one group of existing model predictive control calculation results.

[0010] Optionally, in one embodiment of the present application, the expression of the simplified constraint set is:

[0011]

[0012] in, is a simplified set of constraints, is the state, positive set of the selected model predictive control calculation results, It is a collection of predictive control calculation results obtained using the selected model.

[0013] Optionally, in one embodiment of the present application, the simplified quadratic programming problem is expressed as:

[0014]

[0015] stGz≤Sx+W

[0016] in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

[0017] The second aspect of the present application provides a model predictive control calculation device for a suspension system, including: a selection module, used to screen at least one group of existing model predictive control calculation results according to state observation information of a target suspension system and a mathematical model of the target suspension system; a calculation module, used to calculate a simplified constraint set of the model predictive control of the mathematical model using the at least one group of existing model predictive control calculation results and the state observation information, and construct a simplified quadratic programming problem based on the simplified constraint set and the mathematical model; a solution module, used to solve the simplified quadratic programming problem to obtain an optimized solution result of the mathematical model, so as to determine the actual control quantity of the target suspension system according to the optimized solution result of the mathematical model.

[0018] Optionally, in one embodiment of the present application, the selection module includes: an acquisition unit, used to acquire online trajectory calculation results and / or offline calculation results of the mathematical model based on the state observation information; a selection unit, used to screen at least one group of existing model predictive control calculation results from the online trajectory calculation results and / or the offline calculation results.

[0019] Optionally, in one embodiment of the present application, the calculation module includes: a deletion unit, which is used to delete the target constraint using the offline calculation result when calculating the model predictive control calculation result of the mathematical model online, so as to calculate the simplified constraint set of the model predictive control of the mathematical model.

[0020] Optionally, in one embodiment of the present application, the at least one group of existing model predictive control calculation results includes a state, an active set, and an optimal solution corresponding to the at least one group of existing model predictive control calculation results.

[0021] Optionally, in one embodiment of the present application, the expression of the simplified constraint set is:

[0022]

[0023] in, is a simplified set of constraints, is the state, positive set of the selected model predictive control calculation results, It is a collection of predictive control calculation results obtained using the selected model.

[0024] Optionally, in one embodiment of the present application, the simplified quadratic programming problem is expressed as:

[0025]

[0026] stGz≤Sx+W

[0027] in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

[0028] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the model predictive control calculation method for the suspension system as described in the above embodiment.

[0029] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned model predictive control calculation method for the suspension system.

[0030] A fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned model predictive control calculation method for the suspension system.

[0031] The embodiment of the present application can calculate a simplified constraint set based on the existing model predictive control calculation results of the target suspension system, construct a simplified quadratic programming problem, and thus solve the optimal solution result of the mathematical model of the target suspension system to determine the actual control amount of the target suspension system. Thus, it is achieved to adaptively delete redundant constraints through the calculation results of the previous moment, or to delete constraints by using the partial solution of the offline model predictive control during online calculation, thereby effectively improving the solution calculation speed of the model predictive control without changing the optimal solution, and the storage and search time of this offline solution can be adjusted according to the model of the control system itself, which is convenient for practical application in the embedded system. Thus, it solves the problem that the methods in the related art are all committed to improving the solution speed of the model predictive control by using the special structure of MPC, ignoring the role of historical solution data, and it is difficult to further improve the calculation speed of MPC, and the solution of the model predictive control is still difficult, and it is difficult to apply to the embedded system.

[0032] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0034] Figure 1 A flow chart of a model predictive control calculation method for a suspension system provided according to an embodiment of the present application;

[0035] Figure 2 A schematic diagram of the framework of another model predictive control method principle of an embodiment of the present application;

[0036] Figure 3 A schematic diagram of the framework of a suspension system model predictive control method based on online deletion of constraints according to an embodiment of the present application;

[0037] Figure 4 A schematic diagram of redundant constraint definition for one embodiment of the present application;

[0038] Figure 5 A schematic diagram of the principle of deleting redundant constraints according to an embodiment of the present application;

[0039] Figure 6 This is a schematic diagram of the effect of adaptively deleting redundant constraints according to an embodiment of the present application;

[0040] Figure 7 A schematic diagram of the impact of offline data according to an embodiment of the present application;

[0041] Figure 8 A schematic diagram of the structure of a model predictive control computing device for a suspension system provided according to an embodiment of the present application;

[0042] Fig. 9 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application.

[0043] Reference numerals:

[0044] 10-Model predictive control computing device for suspension system: 100-selection module, 200-computation module and 300-solution module; 901-memory, 902-processor and 903-communication interface. DETAILED DESCRIPTION

[0045] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0046] The following describes the model predictive control calculation method and device of the suspension system of the embodiment of the present application with reference to the accompanying drawings. In view of the methods in the related art mentioned in the above background technology, all of them are committed to improving the solution speed of the model predictive control by utilizing the special structure of MPC, ignoring the role of historical solution data, and it is difficult to further improve the calculation speed of MPC. The solution of model predictive control is still difficult, and it is difficult to apply to the problem of embedded system. The present application provides a model predictive control calculation method for a suspension system, in which a simplified constraint set can be calculated according to the existing model predictive control calculation results of the target suspension system, and a simplified quadratic programming problem can be constructed to solve the optimal solution result of the mathematical model of the target suspension system and determine the actual control amount of the target suspension system. Thus, redundant constraints can be adaptively deleted through the calculation results of the previous moment, or the partial solution of the offline model predictive control can be used to delete the constraints during online calculation, thereby effectively improving the solution calculation speed of the model predictive control without changing the optimal solution, and the storage and search time of this offline solution can be adjusted according to the model of the control system itself, which is convenient for embedded system to be applied in practice. This solves the problem that the methods in related technologies are all committed to improving the solution speed of model predictive control by utilizing the special structure of MPC, but ignore the role of historical solution data, making it difficult to further improve the calculation speed of MPC, and the solution of model predictive control is still difficult and difficult to apply to embedded systems.

[0047] Specifically, Figure 1 A flow chart of a model predictive control calculation method for a suspension system provided in an embodiment of the present application.

[0048] like Figure 1 As shown, the model predictive control calculation method of the suspension system includes the following steps:

[0049] In step S101, at least one set of existing model predictive control calculation results is selected according to the state observation information of the target suspension system and the mathematical model of the target suspension system, wherein the at least one set of existing model predictive control calculation results includes the state, active set, and optimal solution corresponding to the at least one set of existing model predictive control calculation results.

[0050] It is understandable that the target suspension system here refers to the suspension system that applies model predictive control, such as the various control systems of maglev trains. Among them, maglev trains are a new mode of transportation that uses contactless electromagnetic suspension, guidance and drive systems, and have high-speed and medium-low speed forms. High-speed maglev trains can reach a maximum speed of more than 500 kilometers per hour, making them the fastest ground passenger transportation in the world; while medium- and low-speed maglev trains have speeds between 100 and 120 kilometers per hour, with advantages such as energy saving, environmental protection, low noise and small turning radius. They are especially suitable for operation in steep areas such as tourist attractions and suburbs, showing great application potential.

[0051] Although maglev trains have initially achieved engineering applications, they still face some challenges in long-term operation, including suspension point dropout and rail smashing. These problems may cause the train to slow down, thus affecting operating efficiency and passenger comfort. The point dropout problem refers to the failure of some suspension points in the suspension frame, causing the train to slow down; and the point dropout problem may cause rail smashing, the skid contacts the ground, causing the carriage to slide against the track, affecting the passenger experience. Rail smashing means that some suspension points in the suspension frame are unstable, the train is adsorbed on the track, and the speed needs to be reduced to ensure safety. Therefore, the development of control technology for maglev trains is one of the keys to solving current challenges and breaking through engineering bottlenecks. The application of model predictive control methods to suspension control has achieved good control effects.

[0052] In some embodiments, in order to improve the solution speed of the model predictive control method, most studies are committed to starting from the structure of MPC, which easily limits the solution speed of the model predictive control. Based on this, the present application can start from the historical solution data of the model predictive control, that is, the existing model predictive control calculation results, to accelerate the solution speed of the model predictive control.

[0053] The mechanism of model predictive control is to solve an open-loop optimization problem in a finite time domain online at each sampling moment based on the current state observation information of the controlled object and the mathematical model of the controlled system, and apply the first element of the obtained optimal control sequence to the controlled object, and repeat the above process at the next sampling moment.

[0054] Therefore, the present application can also select at least one set of existing model predictive control calculation results in combination with the state observation information and mathematical model of the target suspension system to accelerate the solution speed of the model predictive control, thereby improving the accuracy of the solution. Among them, the at least one set of existing model predictive control calculation results can include but is not limited to the state, active set and optimal solution of the model predictive control calculation result.

[0055] When screening at least one set of existing prediction model control calculation results, the present application may, but is not limited to, be performed according to certain standards or principles. For example, the present application may select a set of model prediction control calculation results based on the minimum difference between the state and the current observed state. It should be noted that the specific selection requirements can be selected or adjusted by professional and technical personnel in this field according to actual conditions. This is only an exemplary description and no specific requirements are made.

[0056] Among them, the state here can be understood as the system state corresponding to the model predictive control calculation result at that time; the positive solution here can be understood as a set of constraints that satisfy the equality constraints or inequality constraints on the boundary under a specific solution (equality constraints usually come from the dynamic model of the system, namely the state equation, and inequality constraints may come from the physical limitations, safety requirements or other constraints of the system); the optimal solution here refers to the control strategy that makes the objective function optimal (usually minimized) under the premise of satisfying all constraints (including equality constraints and inequality constraints).

[0057] Optionally, in one embodiment of the present application, at least one group of existing model predictive control calculation results is screened according to the state observation information of the target suspension system and the mathematical model of the target suspension system, including: obtaining online trajectory calculation results and / or offline calculation results of the mathematical model based on the state observation information; and screening at least one group of existing model predictive control calculation results from the online trajectory calculation results and / or offline calculation results.

[0058] Based on the relevant descriptions of other embodiments, it can be understood that the present application can select at least one set of existing model predictive control calculation results based on the state observation information of the target suspension system and the mathematical model of the target suspension system. In this process, the present application can obtain the online trajectory calculation results and offline calculation results of the mathematical model corresponding to the target suspension system based on the state observation information, and then select at least one set of existing model predictive control calculation results from the online trajectory calculation results and the offline calculation results.

[0059] For example, the linear model of the suspension system can be considered as follows: +=Ax+Bu, where A and B are coefficients, x is the state corresponding to the selected model predictive control calculation result, and u is the control constraint corresponding to the selected model predictive control calculation result. The state constraint set can be set to The control constraint set can be set to In the sampling time domain at the kth step, the observed state can be expressed as x k .

[0060] Next, the embodiment of the present application can construct an optimization problem of the model predictive control problem in the N-th step prediction time domain, and the expression can be but not limited to the following:

[0061]

[0062] stx t+1|k =Ax t|k +Bu t|k ,

[0063]

[0064] x 0|k =x k ,

[0065] k=0,1,...,N-1

[0066] in, is a designed terminal constraint set, Q and R are the weight matrices of the stage losses in the optimization objective of the model predictive control problem, and P is the weight matrix of the terminal losses of the optimization objective. It should be noted that Q, R, and P can be manually adjusted and designed by professional and technical personnel in this field so that the control objective is biased between stabilization and energy saving. The embodiments of this application are only for illustrative purposes and are not specifically limited.

[0067] Furthermore, the original optimization problem can be written as a standard quadratic programming problem through the construction method of condense MPC. The expression can be but is not limited to the following:

[0068]

[0069] stGz≤Sx+W

[0070] Then, the Lipsitz constant is calculated offline, and the formula can be, but is not limited to, expressed as follows:

[0071]

[0072] Among them, H and F are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem, and G j is the row vector of the j-th row of the G matrix.

[0073] Finally, a set of calculation results is selected from the existing online trajectories or offline calculation results.

[0074] For example, we can assume that the existing online trajectory calculation result is x i ,i∈[1,k-1], the status of the offline calculation result is From x i ,i∈[1,k-1] and Find a state in Make this state consistent with the current observation state x k The distance is the smallest, that is Then the optimal solution and positive set corresponding to the model predictive control calculation results are recorded as

[0075] In this way, the selection of the model predictive control calculation results can be completed, and the state, active set and optimal solution corresponding to the model predictive control calculation results can be obtained. Moreover, the embodiment of the present application can derive the relationship between the two optimal solutions by analyzing the Lipsitz continuity of the quadratic programming, identify redundant constraints based on this relationship, and delete these redundant constraints, so as not to change the optimal solution.

[0076] Step S102, using at least one set of existing model predictive control calculation results and state observation information to calculate a simplified constraint set of the model predictive control of the mathematical model, and constructing a simplified quadratic programming problem based on the simplified constraint set and the mathematical model.

[0077] The expression of the simplified constraint set may be, but is not limited to,:

[0078]

[0079] in, is a simplified set of constraints, is the state, positive set of the selected model predictive control calculation results, It is a collection of predictive control calculation results obtained using the selected model.

[0080] And, the expression of the simplified quadratic programming problem can be but is not limited to:

[0081]

[0082] stGz≤Sx+W

[0083] in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

[0084] In other embodiments, after obtaining at least one set of existing model predictive control calculation results, the present application can obtain a simplified constraint set of the model predictive control of the mathematical model of the target suspension system by calculation based on the calculation results combined with the previously obtained state observation information, that is, a simplified constraint set. Further, based on the simplified constraint set and the mathematical model of the target suspension system, a simplified quadratic programming problem can be constructed.

[0085] For example, the mathematical model of the target suspension system is still a linear model x + =Ax+Bu; the existing online trajectory calculation result is x i ,i∈[1,k-1], the status of the offline calculation result is Status is The optimal solution and the positive set are

[0086] Next, the simplified constraint set is calculated. First, the embodiment of the present application can calculate:

[0087]

[0088] in, It is a positive set The complement of G j , w j , S j are the row vectors (row values) of the j-th row of the matrices (vectors) G, w, and S respectively.

[0089] Then, we obtain the simplified constraint set

[0090]

[0091] Using the simplified constraint set of calculations Then a simplified quadratic programming problem can be constructed. The problem can be, but is not limited to, expressed as follows:

[0092]

[0093] stGz≤Sx+W

[0094] in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

[0095] Figure 2 A schematic diagram of the framework of another model predictive control method principle of an embodiment of the present application; Figure 3 This is a schematic diagram of the framework of the suspension system model predictive control method principle based on online data deletion constraints in one embodiment of the present application. Figure 2 and Figure 3 As shown, other model predictive control methods mainly solve the optimization problem directly through explicit MPC, and then input the obtained control quantity into the controlled system to control the related system. The embodiment of the present application uses the existing model predictive control calculation results to construct a simplified constraint set to obtain a simplified optimization problem, and then solves the corresponding control quantity and inputs it into the controlled system to control the related system.

[0096] Optionally, in one embodiment of the present application, a simplified constraint set of the model predictive control of the mathematical model is calculated using at least one set of existing model predictive control calculation results and state observation information, including: when calculating the model predictive control calculation results of the mathematical model online, using the offline calculation results to delete the target constraints to calculate the simplified constraint set of the model predictive control of the mathematical model.

[0097] Based on the relevant descriptions of other embodiments, it can be understood that when selecting at least one group of existing model predictive control calculation results, the present application mainly but not limited to selecting from existing online trajectory calculation results and offline calculation results.

[0098] In the actual execution process, when using this group of model predictive control calculation results and state observation information to calculate the simplified constraint set of the model predictive control of the mathematical model of the target suspension system, this application also mainly uses the selected online trajectory calculation results or offline calculation results to delete the target constraints, thereby constructing a simplified constraint set.

[0099] Furthermore, considering that the offline calculation method of model predictive control in related technologies is mainly through storing explicit MPC, that is, offline calculation of explicit expression of model predictive control law, and obtaining the optimal solution by searching the segmented table of the expression during online control, however, the storage amount and search time of this method will increase exponentially with the increase of the scale of the problem.

[0100] Therefore, the embodiments of the present application can calculate some solutions of model predictive control offline, and when calculating the model predictive control calculation results of the mathematical model of the target suspension system online, use these offline solutions to delete the target constraints and calculate a simplified constraint set, thereby improving the calculation speed. Moreover, both the storage and search time of such offline solutions can be adjusted according to the model of the control system itself.

[0101] Figure 4 Schematic diagram of the definition of redundant constraints for an embodiment of the present application; Figure 5 Schematic diagram of the principle of deleting redundant constraints for an embodiment of the present application. As Figure 4 and Figure 5 shown, the embodiments of the present application can use the existing calculation results to delete redundant constraints, thereby accelerating the model predictive control calculation. Among them, the existing calculation results include but are not limited to the calculation results of the online trajectory and the offline calculation results; the redundant constraints here refer to the constraints that will not change the optimal value after being deleted in the considered optimization problem. Thus, the present application can adaptively delete constraints according to the calculation results of the previous moment, and can also use the offline calculation results to delete redundant constraints. Moreover, the redundant constraints deleted by the embodiments of the present application will not change the optimal value of the original optimization problem, but reduce the number of constraints, thereby reducing the complexity of the optimization problem, and further reducing the solution time of the solver and accelerating the online solution process of model predictive control.

[0102] Figure 6 Schematic diagram of the effect of adaptively deleting redundant constraints for an embodiment of the present application, Figure 7 Schematic diagram of the influence of offline data for an embodiment of the present application. As Figure 6 and Figure 7 shown, Figure 6 shows the comparison results of the ratio of the calculation time of the model predictive control calculation method of the suspension system in the embodiments of the present application and other solution methods, as well as the comparison results of the proportion of the remaining constraints in the model predictive control. It can be seen that the model predictive control calculation method of the suspension system in the embodiments of the present application has a better solution speed; as Figure 7 shown, D1, D2, and D3 are offline data sets corresponding to different scales respectively, where the scale size relationship is D1 < D2 < D3. Combining Figure 6 and Figure 7 it can be seen that for the model predictive control calculation method of the suspension system in the embodiments of the present application, the larger the scale of the offline data set, the larger the number of deleted constraints, and the smaller the total calculation time, indicating that the effect of the process of deleting constraints in the embodiments of the present application will gradually improve as the system runs. That is, by designing the model predictive controller, the system state can converge to a stable point. As the system gradually approaches the stable point, the number of deleted constraints gradually increases, and finally the original optimization problem can be simplified to an unconstrained quadratic programming problem.

[0103] Step S103, solving the simplified quadratic programming problem to obtain an optimization solution result of the mathematical model, so as to determine the actual control amount of the target suspension system according to the optimization solution result of the mathematical model.

[0104] As a possible implementation method, after constructing a simplified quadratic programming problem, the embodiment of the present application can obtain the optimal solution of the model predictive control of the mathematical model of the target suspension system by solving the quadratic programming problem, and select the first element as the control quantity and input it into the controlled object.

[0105] Furthermore, after the control quantity is applied to the controlled system, i.e., the target suspension system, the embodiment of the present application can also observe the state at the next sampling moment and repeat the process, thereby achieving real-time solution of the model predictive control of the target suspension system.

[0106] According to the model predictive control calculation method for the suspension system proposed in the embodiment of the present application, a simplified constraint set can be calculated based on the existing model predictive control calculation results of the target suspension system, and a simplified quadratic programming problem can be constructed, thereby solving the optimization solution result of the mathematical model of the target suspension system and determining the actual control amount of the target suspension system. Thus, it is achieved to adaptively delete redundant constraints through the calculation results of the previous moment, or to delete constraints by using the partial solution of the offline model predictive control during online calculation, thereby effectively improving the solution calculation speed of the model predictive control without changing the optimal solution, and the storage and search time of this offline solution can be adjusted according to the model of the control system itself, which is convenient for practical application in the embedded system. Thus, the problems that the methods in the related art are all committed to improving the solution speed of the model predictive control by utilizing the special structure of MPC, ignoring the role of historical solution data, making it difficult to further improve the calculation speed of MPC, and the solution of the model predictive control is still difficult, and it is difficult to apply to the embedded system are solved.

[0107] Next, a model predictive control computing device for a suspension system according to an embodiment of the present application will be described with reference to the accompanying drawings.

[0108] Figure 8 It is a structural schematic diagram of a model predictive control computing device for a suspension system according to an embodiment of the present application.

[0109] like Figure 8 As shown, the model predictive control computing device 10 of the suspension system includes: a selection module 100, a calculation module 200 and a solution module 300.

[0110] The selection module 100 is used to select at least one group of existing model predictive control calculation results according to the state observation information of the target suspension system and the mathematical model of the target suspension system.

[0111] The calculation module 200 is used to calculate a simplified constraint set of the model predictive control of the mathematical model using at least one set of existing model predictive control calculation results and state observation information, and construct a simplified quadratic programming problem based on the simplified constraint set and the mathematical model.

[0112] The solving module 300 is used to solve the simplified quadratic programming problem to obtain an optimized solution result of the mathematical model, so as to determine the actual control amount of the target suspension system according to the optimized solution result of the mathematical model.

[0113] Optionally, in one embodiment of the present application, the selection module 100 includes: an acquisition unit and a selection unit.

[0114] Wherein, the acquisition unit is used to acquire the online trajectory calculation result and / or the offline calculation result of the mathematical model based on the state observation information.

[0115] The selection unit is used to select at least one group of existing model predictive control calculation results from the online trajectory calculation results and / or the offline calculation results.

[0116] Optionally, in one embodiment of the present application, the calculation module 200 includes: a deletion unit.

[0117] The deletion unit is used to delete the target constraint by using the offline calculation result when calculating the model predictive control calculation result of the mathematical model online, so as to calculate the simplified constraint set of the model predictive control of the mathematical model.

[0118] Optionally, in one embodiment of the present application, at least one set of existing model predictive control calculation results includes a state, an active set, and an optimal solution corresponding to at least one set of existing model predictive control calculation results.

[0119] Optionally, in one embodiment of the present application, the expression of the simplified constraint set may be, but is not limited to:

[0120]

[0121] in, is a simplified set of constraints, is the state, positive set of the selected model predictive control calculation results, It is a collection of predictive control calculation results obtained using the selected model.

[0122] Optionally, in one embodiment of the present application, the expression of the simplified quadratic programming problem may be, but is not limited to,:

[0123]

[0124] stGz≤St+W

[0125] in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

[0126] It should be noted that the above explanation of the embodiment of the model predictive control calculation method for the suspension system is also applicable to the model predictive control calculation device for the suspension system of this embodiment, and will not be repeated here.

[0127] According to the model predictive control calculation device for the suspension system proposed in the embodiment of the present application, a simplified constraint set can be calculated based on the existing model predictive control calculation results of the target suspension system, and a simplified quadratic programming problem can be constructed, thereby solving the optimization solution result of the mathematical model of the target suspension system and determining the actual control amount of the target suspension system. Thus, it is achieved to adaptively delete redundant constraints through the calculation results of the previous moment, or to delete constraints by using the partial solution of the offline model predictive control during online calculation, thereby effectively improving the solution calculation speed of the model predictive control without changing the optimal solution, and the storage and search time of this offline solution can be adjusted according to the model of the control system itself, which is convenient for embedding the system for practical application. Thus, the problems that the methods in the related art are all committed to improving the solution speed of the model predictive control by utilizing the special structure of MPC, ignoring the role of historical solution data, making it difficult to further improve the calculation speed of MPC, and the solution of the model predictive control is still difficult, and it is difficult to apply to the embedded system are solved.

[0128] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0129] A memory 901 , a processor 902 , and a computer program stored in the memory 901 and executable on the processor 902 .

[0130] When the processor 902 executes the program, the model predictive control calculation method for the suspension system provided in the above embodiment is implemented.

[0131] Furthermore, the electronic device further comprises:

[0132] The communication interface 903 is used for communication between the memory 901 and the processor 902 .

[0133] The memory 901 is used to store computer programs that can be executed on the processor 902 .

[0134] The memory 901 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0135] If the memory 901, the processor 902 and the communication interface 903 are implemented independently, the communication interface 903, the memory 901 and the processor 902 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0136] Optionally, in a specific implementation, if the memory 901, the processor 902 and the communication interface 903 are integrated on a chip, the memory 901, the processor 902 and the communication interface 903 can communicate with each other through an internal interface.

[0137] The processor 902 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0138] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned model predictive control calculation method for the suspension system.

[0139] The embodiment of the present application also provides a computer program product, including a computer program, which can run computer instructions. When the computer instructions are executed by a processor, the model predictive control calculation method of the suspension system provided in the embodiment of the present application is implemented.

[0140] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0141] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0142] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0143] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0144] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0145] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0146] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0147] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A model predictive control calculation method for a suspension system, characterized in that: The following steps are involved: Selecting at least one set of existing model predictive control calculation results according to state observation information of the target suspension system and a mathematical model of the target suspension system; Calculating a simplified constraint set of the model predictive control of the mathematical model using the at least one set of existing model predictive control calculation results and the state observation information, and constructing a simplified quadratic programming problem based on the simplified constraint set and the mathematical model; The simplified quadratic programming problem is solved to obtain an optimization solution result of the mathematical model, so as to determine the actual control amount of the target suspension system according to the optimization solution result of the mathematical model.

2. The method according to claim 1, characterized in that The step of selecting at least one set of existing model predictive control calculation results according to the state observation information of the target suspension system and the mathematical model of the target suspension system includes: Based on the state observation information, obtaining an online trajectory calculation result and / or an offline calculation result of the mathematical model; At least one group of existing model predictive control calculation results is screened from the online trajectory calculation results and / or the offline calculation results.

3. The method according to claim 2, characterized in that The method of calculating a simplified constraint set of the model predictive control of the mathematical model using the at least one set of existing model predictive control calculation results and the state observation information includes: When the model predictive control calculation result of the mathematical model is calculated online, the target constraint is deleted using the offline calculation result to calculate a simplified constraint set of the model predictive control of the mathematical model.

4. The method according to claim 1, characterized in that: The at least one group of existing model predictive control calculation results includes a state, an active set, and an optimal solution corresponding to the at least one group of existing model predictive control calculation results.

5. The method according to claim 1, characterized in that: The expression of the simplified constraint set is: in, is a simplified set of constraints, is the state, positive set of the selected model predictive control calculation results, It is a collection of predictive control calculation results obtained using the selected model.

6. The method according to claim 1, characterized in that The simplified quadratic programming problem is expressed as: in, G, S, and W correspond to the simplified constraint set. The submatrix of H, F, G, S, and W are the parameters corresponding to the transformation of the model predictive control problem into a standard quadratic programming problem. z is the optimization variable input into the model predictive control problem. x is the initial state of the model predictive control problem. z T , x T are the transposed vectors of z and x respectively.

7. A model predictive control computing device for a suspension system, characterized in that: include: A selection module, used for selecting at least one set of existing model predictive control calculation results according to the state observation information of the target suspension system and the mathematical model of the target suspension system; A calculation module, used to calculate a simplified constraint set of the model predictive control of the mathematical model using the at least one set of existing model predictive control calculation results and the state observation information, and construct a simplified quadratic programming problem based on the simplified constraint set and the mathematical model; A solution module is used to solve the simplified quadratic programming problem to obtain an optimization solution result of the mathematical model, so as to determine the actual control amount of the target suspension system according to the optimization solution result of the mathematical model.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the model predictive control calculation method for the suspension system according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the model predictive control calculation method for the suspension system as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed, it is used to implement the model predictive control calculation method for the suspension system as described in any one of claims 1-6.

Images