Digital exhibition and display intelligent control method and system based on internet of things

Abstract

The application provides a digital exhibition intelligent control method and system based on the Internet of Things, obtains a dynamic system model and a current state deviation of a target controlled object in a digital exhibition environment; based on a system topological structure and historical state data of the digital exhibition environment, a graph convolution gating recurrent unit network is used to predict a state deviation of the target controlled object at a certain time in the future, and a predicted deviation vector is obtained; a nonlinear integral sliding surface of fusion prediction information is obtained by using the current state deviation and the predicted deviation vector; an estimated value of lumped uncertainty is obtained by using a nonlinear disturbance observer to estimate the lumped uncertainty received during system operation; a total control amount is composed of an equivalent control term and a robust approaching term; and the target controlled object is controlled by using the total control amount.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of intelligent control, and particularly relates to a digital exhibition intelligent control method and system based on Internet of Things. BACKGROUND

[0002] Digital exhibition is a new type of display form. In order to ensure the smooth operation of the digital exhibition system such as intelligent lighting, interactive equipment, environment simulator, etc. and provide high-quality user experience, it is crucial to accurately control the core controlled object. Sliding mode control has the advantages of being insensitive to external disturbance and fast response, and has been widely used in the field of nonlinear system control. Sliding mode control uses a predetermined sliding surface and applies a switching control law to make the system state trajectory reach and move along the sliding surface in a limited time. However, the discontinuous sign function switching term in the standard sliding mode control can easily cause the system to produce chattering phenomenon. This high-frequency oscillation can cause waste of control energy and even cause physical wear and damage to the actuator. To suppress chattering, existing methods often use boundary layer or saturation function for continuous processing, but this sacrifices tracking accuracy. Moreover, the traditional sliding mode controller is usually a post-compensation mechanism that only relies on the feedback adjustment of the system state deviation at the current time, lacking the ability to predict the future of the system. When dealing with systems that change rapidly or have obvious time delay, the control response lag and overshoot problems are prone to occur. How to accurately estimate and feed forward compensation online to improve the anti-disturbance performance of the digital exhibition control system is a difficult problem to be solved in the prior art. SUMMARY

[0003] The purpose of the present application is to provide a digital exhibition intelligent control method and system based on Internet of Things, which can solve the problems of chattering and time delay when controlling the digital exhibition control object, comprising the following steps:

[0004] Obtain the dynamic system model and the current state deviation of the target controlled object in the digital exhibition environment; based on the system topology and historical state data of the digital exhibition environment, use the graph convolution gating recurrent unit network to predict the state deviation of the target controlled object at a certain time in the future, and obtain the predicted deviation vector;

[0005] Use the current state deviation and the predicted deviation vector to obtain a nonlinear integral sliding surface with fused prediction information; use a nonlinear disturbance observer to estimate the lumped uncertainty received during system operation and obtain an estimated value of the lumped uncertainty;

[0006] An equivalent control term is derived based on a dynamic system model and a sliding mode condition, a robust approaching term is obtained according to a disturbance compensation term of an estimated value of lumped uncertainty and a fuzzy logic system taking a sliding mode variable as input; a total control quantity is constituted by the equivalent control term and the robust approaching term; and the target controlled object is controlled by using the total control quantity.

[0007] The application further provides a digital exhibition intelligent control system based on the Internet of Things, comprising the following modules.

[0008] A prediction deviation calculation module is configured to acquire a dynamic system model and a current state deviation of a target controlled object in a digital exhibition environment, predict a state deviation of the target controlled object at a future time based on a system topology and historical state data of the digital exhibition environment, and obtain a prediction deviation vector by using a graph convolution gate recurrent unit network.

[0009] An estimation module is configured to obtain a nonlinear integral sliding mode surface of fusion prediction information by using the current state deviation and the prediction deviation vector, and obtain an estimated value of lumped uncertainty by using a nonlinear disturbance observer to estimate the lumped uncertainty in the system operation process.

[0010] A control module is configured to derive an equivalent control term based on a dynamic system model and a sliding mode condition, obtain a robust approaching term according to a disturbance compensation term of an estimated value of lumped uncertainty and a fuzzy logic system taking a sliding mode variable as input, constitute a total control quantity by the equivalent control term and the robust approaching term, and control the target controlled object by using the total control quantity.

[0011] By introducing a graph convolution gate recurrent unit network, the application can predict the future state deviation of the system, thereby realizing the pre-adjustment of the control quantity and significantly reducing the overshoot in the control process. The nonlinear sliding mode surface constructed by fusing the prediction information and the integral term not only ensures the accurate tracking of the system state, but also effectively eliminates the steady-state error. Meanwhile, the nonlinear disturbance observer can accurately estimate and compensate the lumped uncertainty encountered in the system operation process, thereby enhancing the anti-disturbance ability and robustness of the system. The fuzzy logic system in the approaching law can smoothly adjust the approaching action according to the sliding mode variable, thereby fundamentally suppressing the chattering phenomenon of the traditional sliding mode control and ensuring the smoothness of the control output, and comprehensively improving the control performance and operation reliability of the digital exhibition system. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 The flowchart is the first embodiment;

[0013] Figure 2 The schematic diagram of input s fuzzification;

[0014] Figure 3 is a schematic diagram of a fuzzy rule base. DETAILED DESCRIPTION

[0015] For the purposes of the present application, the technical solutions and advantages thereof will be more apparent from the following detailed description of the embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application. It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of the related data need to comply with relevant laws, regulations and standards, and provide corresponding operation portal for the user to choose authorization or rejection.

[0016] In a first embodiment, the present application provides a digital exhibition intelligent control method based on Internet of Things, as shown in the following steps: Figure 1

[0017] S1, obtaining a dynamic system model and a current state deviation of a target controlled object in a digital exhibition environment; based on a system topology structure and historical state data of the digital exhibition environment, a graph convolution gate recurrent unit network is used to predict a state deviation of the target controlled object at a future time, to obtain a predicted deviation vector;

[0018] The digital exhibition includes a plurality of controlled objects, such as track lights, interactive mechanical arms, screen rotation controllers, etc. The dynamic system model of the target controlled object, such as the intelligent track light in the exhibition hall, can be represented as a second-order nonlinear system. The actual state, such as the angle or position, is collected in real time through an encoder or a visual sensor, and compared with the preset expected state to obtain the state deviation e. The rate of change of the deviation de / dt is calculated by differencing the state deviations e at two consecutive sampling times and then dividing by the sampling time. Each device in the exhibition environment is taken as a node of a graph, and the communication or physical association between the devices is taken as an edge, to construct a graph structure. A graph convolution network layer is used to extract the spatial dependence features between nodes, and the output thereof is fed into a gate recurrent unit layer, which is responsible for learning the dynamic pattern of the system state evolution over time. Using historical operation data containing device states and control instructions, the network is trained offline through a back propagation algorithm, so that it can output state deviation prediction values for a plurality of time steps in the future according to the current and historical states .

[0019] ​S2, obtaining a nonlinear integral sliding surface of fusion prediction information by using the current state deviation and the measurement deviation vector; obtaining an estimated value of the lumped uncertainty by using a nonlinear disturbance observer to estimate the lumped uncertainty received by the system during operation on line;

[0020] A nonlinear integral sliding surface s of fusion prediction information is constructed, preferably, the expression of the nonlinear integral sliding surface s, i.e., the sliding variable s, is: wherein c1, c2, c3 and γ are preset normal number control gains. The control gains c1 and c2 are selected according to the expected system closed-loop pole configuration to ensure the stability and convergence speed of the sliding mode dynamics. The integral term The numerical calculation is performed by using the cumulative summation to eliminate the steady-state error. The prediction deviation The prediction value at the most future time is introduced and smoothed and amplitude-limited by using the hyperbolic tangent function tanh, and the gains c3 and γ are obtained through simulation experiments to balance the strength and stability of the prediction compensation.

[0021] A finite-time disturbance observer is used. The observer constructs an auxiliary variable, and a nonlinear function term based on the system state and the control input is included in the dynamic equation of the auxiliary variable. By designing the nonlinear function term, for example, using the p-norm or the high-order power term based on the sign function, the estimation error of the observer to the lumped uncertainty can converge to zero in a finite time, so that the disturbance estimation value .

[0022] S3, an equivalent control term is obtained by deduction based on the dynamic system model and the sliding mode condition, a robust approaching component is obtained by using a fuzzy logic system taking the sliding variable as the input and a disturbance compensation term based on the estimated value of the lumped uncertainty, a total control quantity is constituted by the equivalent control term and the robust approaching component, and the target controlled object is controlled by using the total control quantity.

[0023] The time derivative of the expression of the sliding surface s is taken, and is obtained, and the system dynamic model is substituted to solve the equivalent control term containing the system model parameters. The robust approaching component is composed of the disturbance compensation term and the fuzzy adaptive term. The fuzzy logic system takes the sliding variable s as the only input, and the input membership function is divided into five fuzzy subsets, i.e., negative large, negative small, zero, positive small and positive large. Based on the set fuzzy rules, if s is positive large, the output is large, and if s is zero, the output is small. The defuzzification is performed by using the center of gravity method to obtain an adaptive gain varying with s, and the gain is used to adjust the approaching speed to realize fast approaching at large deviation and smooth transition at small deviation, so as to suppress the chattering.

[0024] In each control cycle, the calculated equivalent control term is added to the robust approaching term to obtain the final total control amount u. The digital control signal is converted into an analog voltage or current signal by a digital-to-analog converter and applied to the actuator of the target controlled object, such as a driver for driving a DC motor or a power module for controlling the brightness of an LED light, thereby completing the precise control of the exhibition object.

[0025] In one possible implementation, the prediction of the state deviation of the target controlled object at a future time by using the graph convolution gated recurrent unit network includes:

[0026] Each sensor and actuator in the digitized exhibition environment is abstracted as a network node, and the communication relationship and physical connection between nodes are taken as edges to construct an adjacency matrix of the system topology.

[0027] The historical state data sequence of each node at the past N consecutive sampling times is taken as the feature matrix of the node.

[0028] The adjacency matrix and the node feature matrix are input into a graph convolution layer, neighbor node information is aggregated by graph convolution operation, and the spatial correlation features of the entire system at the current time are obtained.

[0029] The spatial feature vector output by the graph convolution layer is sequentially sent to the subsequent gated recurrent unit layer, the dynamic evolution law of the state data in the time dimension is learned, and the state deviation prediction value at the future time is output to obtain the prediction deviation vector.

[0030] In an interactive art device composed of multiple collaborative robotic arms and environment perception cameras, the joint motors, end grippers of each robotic arm, and each camera can be regarded as network nodes. The physical connections between the joints of the robotic arm, the communication bus between the motor controllers and the camera data processing units constitute the edges of the graph, thereby an adjacency matrix A describing the information flow and physical constraints of the entire system can be constructed. For example, if joint motor 1 can receive the position information processed by camera 3, the value of element A13 in matrix A is 1, and vice versa.

[0031] For each node, such as a joint motor, its angular position, angular velocity and output torque data every 10 milliseconds in the past 1 second are collected to form a data sequence containing 100 time steps, each step having 3 features, which constitutes the feature matrix of the node. The entire system, such as an adjacency matrix containing 20 nodes, and the feature matrix of all nodes are input into the graph convolution network together. The graph convolution operation can integrate the information of a node itself and all its neighbor nodes, such as the motion of a joint not only depends on its own historical state, but also is affected by other joints connected to it and the vision system working together, thereby effectively extracting the complex spatial coupling characteristics of the entire exhibition system at the current time.

[0032] The system-level spatial feature vector output by the graph convolution network at each time step is sent into a gated recurrent unit network in chronological order. For example, 5 spatial feature vectors extracted in the last 5 time steps are taken as input sequences. The gated recurrent unit network can learn the evolution pattern of the system state in the time dimension by virtue of its unique gating mechanism, such as learning the dynamic law of the process from visual capture to audience approach to the corresponding welcome action of the mechanical arm. The network finally outputs the prediction value of the deviation of the angles of each joint of the mechanical arm from the expected angle in the future, for example, 200 milliseconds later, i.e., the predicted deviation vector .

[0033] In one possible implementation, the nonlinear disturbance observer is used to estimate the lumped uncertainty received by the system during operation online to obtain an estimated value of the lumped uncertainty, comprising:

[0034] A nonlinear state observer of the following form is constructed to obtain an estimated value of the lumped uncertainty :

[0035]

[0036]

[0037] wherein p is an internal state variable of the observer, L is a positive definite observer gain matrix, q and dq / dt are the current state and its rate of change of the target controlled object, M(q), C(q, dq / dt) and G(q) are the inertia matrix, Coriolis and centrifugal force matrix and gravity term in the dynamic system model respectively, and u is the total control amount.

[0038] The core role of the nonlinear disturbance observer is to estimate and compensate various unmodeled dynamics and external disturbances in real time. In a scenario of a 6-DOF manipulator interacting with a human, these uncertainties can include variations in static and dynamic friction of the joints, backlash in the gearboxes, slight load variations, and even air resistance, which collectively form the lumped uncertainty d. The observer can generate an estimate of the disturbance that is very close to the true disturbance for feedforward compensation in the control to improve control accuracy.

[0039] In the implementation of the observer, the observer gain matrix L is set. For a 6-DOF manipulator, L is designed as a 6x6 diagonal positive definite matrix, for example, its diagonal elements can be set to [20, 20, 20, 10, 10, 10]. The size of these gain values determines the convergence speed of the observer estimate and the sensitivity to measurement noise, larger gain makes faster convergence, but may amplify noise. The observer obtains the current state of the manipulator in real time, i.e. the angles q and angular velocities dq / dt of each joint, and uses known dynamic model parameters such as the inertia matrix M(q), the Coriolis matrix C(q, dq / dt), and the gravity term G(q) to calculate the theoretical system behavior.

[0040] The observer continuously updates its internal state variable p through the first differential equation. The first differential equation calculates the difference between the theoretical dynamics generated by the known model and the control input u and the actual system dynamics corrected by the gain matrix L, which is mainly due to the unmodeled disturbance d. Through the second algebraic equation, the internal state p is combined with the system generalized momentum M(q) multiplied by dq / dt, and then scaled by the gain matrix L, to calculate the real-time estimate of the lumped uncertainty . For example, if the second component of is calculated as -0.8 at a certain time, it means that the observer estimates that the second joint of the manipulator is being subjected to an additional resistance torque of 0.8 Nm.

[0041] In a possible implementation, the equivalent control term is derived based on the dynamic system model and the sliding mode condition, specifically:

[0042] Based on the expected state of the controlled object and its derivatives, combined with the sliding mode condition and the dynamic system model of the controlled object ;

[0043] The error acceleration desired by the sliding mode control is solved from ;

[0044] combined with the error definition the actual acceleration of the plant is expressed as ;

[0045] Substitute the actual acceleration into the dynamic model of the system and ignore the lumped uncertainty term d, the expression of the equivalent control term is solved as

[0046] where sech denotes the hyperbolic secant function.

[0047] The equivalent control term is a core component in the sliding mode control law, representing the precise control force that is required to drive the plant to follow the desired trajectory perfectly in the ideal case, i.e. without any uncertainty and external disturbance. The derivation logic is based on a core assumption that the system state is already on and needs to be maintained on a predefined sliding surface, mathematically expressed as the time derivative of the sliding variable being identically zero. The whole derivation process is around this condition, combined with the system dynamics equation to back-solve the corresponding control quantity.

[0048] The first step in the derivation is to solve the error acceleration that the system must have from the differential equation of the sliding surface using the condition This desired error acceleration is a comprehensive term, which not only depends on the current position error e and velocity error but also incorporates the neural network predicted future error and its rate of change This step determines an ideal error convergence dynamic. For example, if the current error is large, the calculated desired error acceleration will be adjusted accordingly to drive the error to decrease faster.

[0049] Then, the relationship between the error dynamic and the actual motion of the system can be established by the error e being equal to the desired position minus the actual position q. Taking the second time derivative of this definition, the actual acceleration of the system can be expressed in terms of the acceleration of the desired trajectory and the desired error acceleration that has just been solved. Substitute this actual acceleration expression into the dynamic model equation of the robot arm, and temporarily ignore the disturbance term d, the equation can be rearranged by algebraic manipulation to solve for For example, in a certain control period, substitute the calculated error values and system state parameters into the final expression to obtain a six-dimensional torque vector, such as [10.5, 8.2, -5.1, 1.0, 0.5, -0.2] Nm, which is the basic control torque required to maintain the ideal motion.

[0050] In one possible implementation, the robust reaching term is generated according to a disturbance compensation term and an adaptive reaching component using a fuzzy logic system with the sliding variable as input, specifically:

[0051] The robust reaching term is composed of the disturbance compensation term and the adaptive reaching component , and the expression is:

[0052]

[0053] where the disturbance compensation term is used to offset the lumped uncertainty estimated by the nonlinear disturbance observer; and the adaptive reaching component is generated by a fuzzy logic system with the sliding variable s as input, to adjust the reaching speed and suppress chattering.

[0054] The robust reaching term is a key part in the sliding mode control that ensures the system state to converge to the sliding surface quickly and stably from any initial position. It is mainly responsible for handling the uncertainty in the system and ensuring the robustness of the control. The term is composed of two components with distinct functions, one is a direct disturbance feedforward compensation, and the other is an intelligent adaptive reaching law. The two components work together to enable the system to maintain excellent performance even in a complex and variable environment.

[0055] The disturbance compensation term is used to counteract the known disturbance. It uses the lumped uncertainty estimate output by the aforementioned nonlinear disturbance observer to introduce an equal and opposite force or torque in the control signal. For example, if the observer estimates that an unexpected increase in the load of the robot arm has resulted in a 1.2 Nm downward torque at the third joint, i.e. the third component in is -1.2, then the disturbance compensation term will actively increase a positive 1.2 Nm torque in the control output of the joint. This direct offsetting method can reduce the tracking error caused by model inaccuracies and external disturbances, greatly improving the control accuracy of the system.

[0056] The adaptive fuzzy reaching component imitates the control experience of human experts, dynamically adjusting the strength of the reaching control action according to the distance of the system state from the sliding surface, i.e., the size and sign of the sliding variable s. When the absolute value of s is large, it means that the system is far from the ideal state, a larger control quantity will be output to achieve fast reaching. When the absolute value of s is small, it means that the system is already very close to the sliding surface, will significantly reduce its output, as Figure 2 shown, thus avoiding the severe chattering problem caused by sign function switching in traditional sliding mode control, making the control process smoother.

[0057] In one possible implementation, the fuzzy logic system with the sliding variable as input is used to generate an adaptive reaching component, specifically:

[0058] The sliding variable s is taken as input, and the universe of discourse is fuzzified into five fuzzy sets: negative big NB, negative small NS, zero ZE, positive small PS, and positive big PB, which are described using triangular or Gaussian membership functions.

[0059] The following fuzzy rule base is established:

[0060] If s is NB, then is positive big PB;

[0061] If s is NS, then is positive small PS;

[0062] If s is ZE, then is zero ZE;

[0063] If s is PS, then is negative small NS;

[0064] If s is PB, then is negative big NB;

[0065] The Mamdani-type fuzzy inference engine and the center-of-gravity method are used to defuzzify, obtaining the output value of .

[0066] The fuzzy logic system converts the precise sliding variable s value into a fuzzy statement and determines the size of the output control variable through a series of experience-based rules, Figure 3 is a possible fuzzy rule base. For example, for a certain joint of a robotic arm, the actual value range of the sliding variable s may be between negative 5 and positive 5. On this universe of discourse, five fuzzy sets are defined, each described by a membership function, such as a triangle with a vertex at zero and a base range from negative 1 to positive 1 to describe the zero ZE set. When the precise value of s measured at a certain time is 0.8, this value may belong to the zero ZE set with a membership degree of 0.2, and belong to the positive small PS set with a membership degree of 0.8.

[0067] The fuzzy inference engine calculates the output based on this input and the pre-set rule base. Since the value of s 0.8 simultaneously activates the rule "if s is ZE, then is ZE" and the rule "if s is PS, then For the two rules of NS, the system combines the results of the two rules. According to the Mamdani method, the first rule fires with a strength of 0.2, resulting in a zero ZE output fuzzy set that is clipped to a height of 0.2; the second rule fires with a strength of 0.8, resulting in a negative small NS output fuzzy set that is clipped to a height of 0.8. These two clipped output fuzzy sets are then merged into a single, irregularly shaped final output fuzzy set.

[0068] To obtain an exact control value that can be used by the actuator, the merged fuzzy set is defuzzified. The centroid method is used, which calculates the horizontal coordinate value of the geometric center or centroid of the irregularly shaped region. For example, assume that the universe of discourse for the output ranges from -10 to +10 newtons, and that the centroid horizontal coordinate of the merged fuzzy set is calculated to be -3.5. Then, -3.5 newtons is the final exact value that the fuzzy logic system outputs at the current time This value is used to construct a robust reaching term that pushes the system state toward the sliding surface with an appropriate amount of force, achieving smooth and adaptive adjustment of the reaching process.

[0069] In a second embodiment, the present application provides an Internet of Things-based digital exhibition intelligent control system, comprising the following modules:

[0070] A prediction deviation calculation module is configured to obtain a dynamic system model of a target controlled object in a digital exhibition environment and a current state deviation; based on a system topology and historical state data of the digital exhibition environment, a graph convolution gate recurrent unit network is used to predict a state deviation of the target controlled object at a future time to obtain a prediction deviation vector;

[0071] An estimation module is configured to obtain a nonlinear integral sliding mode surface of fusion prediction information using the current state deviation and the prediction deviation vector; and a nonlinear disturbance observer is used to estimate an estimated value of lumped uncertainty in system operation.

[0072] A control module is configured to derive an equivalent control term based on a dynamic system model and a sliding mode condition, obtain a robust reaching term from a disturbance compensation term of the estimated value of lumped uncertainty and an adaptive reaching component generated by a fuzzy logic system with a sliding variable as input, and obtain a total control amount from the equivalent control term and the robust reaching term; and the target controlled object is controlled using the total control amount.

[0073] In one possible implementation, the nonlinear integral sliding mode surface of fusion prediction information is obtained using the current state deviation and the prediction deviation vector, and specifically:

[0074] The expression of the nonlinear integral sliding mode surface s is: ;

[0075] wherein e is the current state deviation, is a predicted deviation vector, c1, c2, c3, γ are preset normal number control gains.

[0076] In one possible implementation, the predicting, by the graph convolution gated recurrent unit network, the state deviation of the target controlled object at a future time, to obtain a predicted deviation vector, comprises:

[0077] The adjacency matrix of the system topology is constructed by abstracting each sensor and actuator in the digitized exhibition environment as a network node and taking the communication relationship and physical connection between nodes as an edge;

[0078] The historical state data sequence of each node at the past N consecutive sampling times is taken as the feature matrix of the node;

[0079] The adjacency matrix and the node feature matrix are input into a graph convolution layer, neighbor node information is aggregated by graph convolution operation, and the spatial correlation features of the entire system at the current time are obtained;

[0080] The spatial feature vector output by the graph convolution layer is sequentially sent to the subsequent gated recurrent unit layer, the dynamic evolution law of the state data in the time dimension is learned, and the state deviation prediction value at the future time is output to obtain the predicted deviation vector.

[0081] In one possible implementation, the estimating, by the nonlinear disturbance observer, the lumped uncertainty received by the system during operation, comprises:

[0082] A nonlinear state observer of the following form is constructed to obtain the estimated value of the lumped uncertainty :

[0083]

[0084]

[0085] wherein p is an internal state variable of the observer, L is a positive definite observer gain matrix, q and dq / dt are the current state and its rate of change of the target controlled object, M(q), C(q, dq / dt), and G(q) are the inertia matrix, the Coriolis and centrifugal force matrix, and the gravity term in the dynamic system model, respectively, and u is the total control amount.

[0086] In one possible implementation, the equivalent control term is derived based on the dynamic system model and the sliding mode condition, and specifically:

[0087] based on the expected state of the controlled object and its derivatives, in conjunction with the sliding mode condition and the dynamic model of the plant are derived;

[0088] ;

[0089] ;

[0090]

[0091] .

[0092] In one possible implementation, the robust reaching term is generated according to a disturbance compensation term based on the estimate of the lumped uncertainty and an adaptive reaching component generated by a fuzzy logic system with the sliding variable as input, in particular:

[0093] The robust reaching term is composed of the disturbance compensation term and the adaptive reaching component , and is expressed as:

[0094]

[0095] where the disturbance compensation term is used to counteract the lumped uncertainty estimated by the nonlinear disturbance observer; the adaptive reaching component is generated by a fuzzy logic system with the sliding variable s as input, to adjust the reaching speed and suppress chattering.

[0096] In one possible implementation, the adaptive reaching component is generated by a fuzzy logic system with the sliding variable as input, in particular:

[0097] The sliding variable s is taken as input, and the universe of discourse is fuzzified into five fuzzy sets: negative big NB, negative small NS, zero ZE, positive small PS, and positive big PB, and is described by a triangular or Gaussian membership function;

[0098] The following fuzzy rule base is established:

[0099] If s is NB, then is positive big PB;

[0100] If s is NS, then is positive small PS;​​​​​​​

[0101] If s is ZE, then is zero ZE;

[0102] If s is PS, then is negative small NS;

[0103] If s is PB, then is negative big NB;

[0104] Using a Mamdani type fuzzy inference engine and the center of gravity method to solve fuzzification, the output value of is obtained.

[0105] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and the necessary general hardware platform. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in the various embodiments or some parts of the embodiments.

[0106] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the system or system embodiments, since it is basically similar to the method embodiments, it is described more simply, and the relevant parts can be referred to the part of the method embodiments. The above-described system and system embodiments are only illustrative, and the units described as separate components can be or can not be physically separated, and the components displayed as units can be or can not be physical units, i.e., they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to the actual needs. Those skilled in the art can understand and implement it without creative labor.

[0107] The method for providing commodity object information and the electronic device provided by the present application are described in detail above, and the principles and implementation manners of the present application are described by applying specific examples. The above embodiment is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A digital exhibition intelligent control method based on the Internet of Things, characterized in that, Includes the following steps: Acquire the dynamic system model and current state deviation of the target controlled object in the digital exhibition environment; Based on the system topology and historical state data of the digital exhibition environment, a graph convolutional gated recurrent unit network is used to predict the state deviation of the target controlled object at a certain future moment, and a prediction deviation vector is obtained. The nonlinear integral sliding surface for fusing prediction information is obtained using the current state deviation and the measured deviation vector; the lumped uncertainty experienced by the system during operation is estimated online using a nonlinear disturbance observer to obtain an estimate of the lumped uncertainty. The equivalent control term is derived based on the dynamic system model and sliding mode conditions. The robust approach term is obtained by using the disturbance compensation term for the lumped uncertainty estimate and the adaptive approach component generated by the fuzzy logic system with the sliding mode variable as input. The total control quantity is composed of the equivalent control term and the robust approach term. The target controlled object is controlled using the total control quantity; The method of using a nonlinear perturbation observer to estimate the lumped uncertainty experienced by the system during operation to obtain an estimate of the lumped uncertainty includes: Construct a nonlinear state observer of the following form to obtain an estimate of the lumped uncertainty. : Where p is the internal state variable of the observer, L is the positive definite observer gain matrix, q and dq / dt are the current state of the target controlled object and its rate of change, M(q), C(q,dq / dt), and G(q) are the inertia matrix, Coriolis force and centrifugal force matrix, and gravity term in the dynamic system model, respectively, and u is the total control variable; The equivalent control term derived based on the dynamic system model and sliding mode conditions is as follows: Based on the desired state of the controlled object and its derivatives, combined with sliding mode conditions Dynamic system model of the controlled object Perform the derivation; Depend on The desired error acceleration for sliding mode control can be obtained by solving the problem. ; Combining the definition of error The actual acceleration of the controlled object Represented as ; Substituting the actual acceleration into the dynamic system model and neglecting the lumped uncertainty term d, the equivalent control term is obtained by solving. The expression: 。 2. The method according to claim 1, characterized in that, The nonlinear integral sliding surface for obtaining fused prediction information using the current state deviation and the measured deviation vector is specifically as follows: The expression for the nonlinear integral sliding surface s is: ; Where e is the current state deviation. The prediction bias vector is defined by c1, c2, c3, and γ, which are preset constant control gains.

3. The method according to claim 1, characterized in that, The method of using a graph convolutional gated recurrent unit network to predict the state deviation of the target controlled object at a future time, and obtaining a predicted deviation vector, includes: The sensors and actuators in the digital exhibition environment are abstracted as network nodes, and the communication relationships and physical connections between nodes are used as edges to construct the adjacency matrix of the system topology. The historical state data sequence of each node over the past N consecutive sampling times is used as the feature matrix of the node; The adjacency matrix and node feature matrix are input into the graph convolutional layer, and the neighbor node information is aggregated through graph convolution operation to obtain the spatial association features of the entire system at the current time. The spatial feature vector output by the graph convolutional layer is fed into the subsequent gated recurrent unit layer in chronological order to learn the dynamic evolution law of the state data in the time dimension, and output the predicted value of the state deviation at future time, thus obtaining the prediction deviation vector.

4. The method according to claim 1, characterized in that, The robust approach term is obtained by using the disturbance compensation term for the lumped uncertainty estimate and the adaptive approach component generated by the fuzzy logic system with sliding mode variables as input. Specifically: The robust approach term Composed of disturbance compensation term and adaptive approach component The composition, expressed as: Among them, the disturbance compensation item is This is used to offset the lumped uncertainty estimated by the nonlinear perturbation observer; the adaptive approaching component It is generated by a fuzzy logic system with sliding mode variable s as input, which adjusts the approach speed and suppresses chattering.

5. The method according to claim 4, characterized in that, The method of generating adaptive approaching components using a fuzzy logic system with sliding mode variables as input specifically involves: Using the sliding mode variable s as input, the universe of discourse is fuzzified into five fuzzy sets: negative large NB, negative small NS, zero ZE, positive small PS, and positive large PB, and described by triangular or Gaussian membership functions. Establish the following fuzzy rule base: If s is NB, then For Zhengda PB; If s is NS, then For the positive PS; If s is ZE, then Zero ZE; If s is PS, then For negative NS; If s is PB, then For negative NB; The fuzzification was defuzzified using a Mamdani-type fuzzy inference engine and the centroid method, resulting in... The output value.

6. A digital exhibition intelligent control system based on the Internet of Things, characterized in that, Includes the following modules: The prediction deviation calculation module is used to obtain the dynamic system model and current state deviation of the target controlled object in the digital exhibition environment; based on the system topology and historical state data of the digital exhibition environment, the module uses a graph convolutional gated recurrent unit network to predict the state deviation of the target controlled object at a certain future moment, and obtains the prediction deviation vector. The estimation module is used to obtain a nonlinear integral sliding surface for fusing prediction information using the current state deviation and the measured deviation vector; and to obtain an estimate of the lumped uncertainty during system operation by using a nonlinear perturbation observer online. The control module is used to derive the equivalent control term based on the dynamic system model and sliding mode conditions, and to obtain the robust approach term based on the disturbance compensation term for the lumped uncertainty estimate and the adaptive approach component generated by the fuzzy logic system with the sliding mode variable as input; the total control quantity is composed of the equivalent control term and the robust approach term. The target controlled object is controlled using the total control quantity; The method of using a nonlinear perturbation observer to estimate the lumped uncertainty experienced by the system during operation to obtain an estimate of the lumped uncertainty includes: Construct a nonlinear state observer of the following form to obtain an estimate of the lumped uncertainty. : Where p is the internal state variable of the observer, L is the positive definite observer gain matrix, q and dq / dt are the current state of the target controlled object and its rate of change, M(q), C(q,dq / dt), and G(q) are the inertia matrix, Coriolis force and centrifugal force matrix, and gravity term in the dynamic system model, respectively, and u is the total control variable; The equivalent control term derived based on the dynamic system model and sliding mode conditions is as follows: Based on the desired state of the controlled object and its derivatives, combined with sliding mode conditions Dynamic system model of the controlled object Perform the derivation; Depend on The desired error acceleration for sliding mode control can be obtained by solving the problem. ; Combining the definition of error The actual acceleration of the controlled object Represented as ; Substituting the actual acceleration into the dynamic system model and neglecting the lumped uncertainty term d, the equivalent control term is obtained by solving. The expression: 。 7. The system according to claim 6, characterized in that, The nonlinear integral sliding surface for obtaining fused prediction information using the current state deviation and the measured deviation vector is specifically as follows: The expression for the nonlinear integral sliding surface s is: ; Where e is the current state deviation. The prediction bias vector is defined by c1, c2, c3, and γ, which are preset constant control gains.

8. The system according to claim 6, characterized in that, The method of using a graph convolutional gated recurrent unit network to predict the state deviation of the target controlled object at a future time, and obtaining a predicted deviation vector, includes: The sensors and actuators in the digital exhibition environment are abstracted as network nodes, and the communication relationships and physical connections between nodes are used as edges to construct the adjacency matrix of the system topology. The historical state data sequence of each node over the past N consecutive sampling times is used as the feature matrix of the node; The adjacency matrix and node feature matrix are input into the graph convolutional layer, and the neighbor node information is aggregated through graph convolution operation to obtain the spatial association features of the entire system at the current time. The spatial feature vector output by the graph convolutional layer is fed into the subsequent gated recurrent unit layer in chronological order to learn the dynamic evolution law of the state data in the time dimension, and output the predicted value of the state deviation at future time, thus obtaining the prediction deviation vector.

Images