Intelligent perception and communication collaborative design method and system of internet of things control system

Abstract

The application provides a kind of intelligent perception and communication collaborative design method and system of Internet of Things control system, comprising: step S1: deploying multiple sensors in each control loop and classifying sensors, offline design sensing mode table;Step S2: carry out sensor transmission scheduling based on double time scale;Step S3: intelligent perception based on multiple sensors and Kalman filter;Step S4: execute sensing mode switching;Step S5: network reconstruction is realized by wireless network coordinator.The application solves the sensor and network interference problem in wireless control with a perception-transmission collaborative design framework, and finally optimizes the overall performance of control system and wireless network.

Description

Technical Field

[0001] This invention relates to the field of wireless network control, and more specifically, to a method and system for intelligent sensing and communication collaborative design of an Internet of Things (IoT) control system. Background Technology

[0002] With the development of embedded intelligence and the Industrial Internet of Things (IIoT), IoT-based industrial control is playing a leading role in the Fourth Industrial Revolution. Recent advances in sensing, wireless networking, computing, and control are transforming how control systems interact with networked and physical processes. IoT-based industrial wireless network control systems integrate computing, communication, and control system design, enabling more reliable, efficient, and collaborative systems with broad application prospects. Sensing and communication are two key processes in networked control systems. Perceiving and transmitting state values ​​is a prerequisite for controlling and driving industrial equipment. The reliability and efficiency of sensing and communication directly affect control performance. However, IoT-based wireless control can be affected by sensor interference (electromagnetic interference, temperature and humidity changes, etc.) and network interference (background noise, metallic obstacles, etc.). Constrained by limited sensing and communication resources in industrial settings, these interferences reduce the reliability and efficiency of sensing and communication, thus degrading control performance.

[0003] Transmitting measurements from a set of fixed sensors to a state observer via wireless sensor-actuator networks is a common approach in Industrial Internet of Things (IIoT) control. However, interference with critical sensors or network links can directly impact state observation performance, thereby compromising control performance. While these adverse effects can be mitigated by adding redundant sensors and optimizing wireless routing, arbitrary redundancy of industrial sensors and their corresponding wireless configurations can lead to resource waste and deployment difficulties, given the often limited sensor and communication resources within a system.

[0004] Regarding interference in sensors and networks, most current research findings and technical implementations tend to design solutions for specific types of interference from the perspective of state estimation or network. For example, outlier-sensitive Kalman filters (H. Fang, MA Haile, and Y. Wang, “Robust extended kalman filtering for systems with measurement outliers,” IEEE Trans. Control Syst. Technol., vol. 30, no. 2, pp. 795–802, 2022) can detect and adapt to abnormal measurements, but are not suitable for wireless control scenarios with packet loss. Dynamic adaptive routing protocols (N. Ma, H. Zhang, H. Hu, and Y. Qin, “Escvad: An energy-saving routing protocol based on voronoi adaptive clustering for wireless sensor networks,” IEEE Internet Things J., vol. 9, no. 11, pp. 9071–9085, 2021) guarantee communication performance when wireless networks change dynamically, but do not consider the dynamic demands of system runtime control performance on communication performance, and so on. These separate designs targeting sensor or network interference do not consider the deep integration of sensing and communication. For example, if a sensor malfunctions, continuing to reliably transmit abnormal measurements is a waste of network resources; simultaneously, because sensor data is transmitted wirelessly using binary encoding, variations in network noise can affect the bit error rate. Furthermore, improving the overall energy efficiency of the wireless control system necessitates adjusting the allocation of sensing and communication resources according to the dynamic operation of industrial equipment. Therefore, the joint design of sensing and communication is crucial for improving the overall performance of IoT-based control systems. A key challenge in the co-design of sensing and communication is effectively integrating advanced sensing and communication technologies, and balancing limited sensing and network resources with flexibility and control performance requirements.B. Demirel, Z. Zou, P. Soldati, and M. Johansson, “Modular design of jointly optimal controllers and forwarding policies for wireless control,” IEEE Transactions on Automatic Control, vol. 59, no. 12, pp. 3252–3265, 2014. This paper proposes a co-design framework for cyber-physical systems, dynamically designing the number of data packets transmitted based on the physical state during system operation and reconstructing the network through a network manager. However, these existing co-design schemes do not consider, or only consider, a specific type of network or sensor interference, thus failing to comprehensively address wireless network and sensor interference.

[0005] Most existing technologies for addressing sensor or wireless network interference do not integrate the entire physical system with the wireless network for collaborative design of control systems and network communication. Existing co-design solutions for network and sensor interference do not consider the evolution of the physical state of industrial equipment, failing to achieve efficient utilization of sensing and network resources. Furthermore, most existing methods for detecting sensor interference do not consider how to differentiate between abnormal sensor readings caused by interference affecting the industrial equipment itself connected to the sensor.

[0006] Patent document CN114584429A discloses an industrial intelligent IoT gateway, relating to the field of IoT device technology. This invention includes an IoT gateway, a fieldbus communication module, and a serial communication module. The communication architecture of the IoT gateway includes a sensing layer, a network layer, and an application layer. The sensing layer includes a sensing module and a control module, used to complete the network control and physical access of sensing extension devices, and to match various sensor network technologies to achieve access to different communication protocols. The network layer is used to manage the IoT system and transmit remote data. The application layer includes an application module and a business management module, used to provide services for the application processes of end users. However, this invention cannot effectively solve the problem of abnormal sensor readings while ensuring the timely response of the entire system to physical disturbances of the controlled object. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for intelligent sensing and communication collaborative design of an Internet of Things (IoT) control system.

[0008] The present invention provides a method for intelligent sensing and communication collaborative design of an Internet of Things (IoT) control system, comprising:

[0009] Step S1: Deploy multiple sensors in each control loop and classify the sensors, and design the sensor mode table offline;

[0010] Step S2: Perform sensor transmission scheduling based on dual time scales;

[0011] Step S3: Intelligent sensing based on multiple sensors and Kalman filters;

[0012] Step S4: Perform sensor mode switching;

[0013] Step S5: Network reconstruction is performed by the wireless network coordinator.

[0014] Preferably, in step S1:

[0015] Step S1.1: Deploy multiple sensors according to the detectability conditions required by the state estimation;

[0016] Step S1.2: Divide the multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal.

[0017] Step S1.3: Solve the multi-objective optimization problem of minimizing the total control and communication costs under different physical states and network link quality, obtain the sensor and its corresponding network configuration, and generate a feasible solution set for subsequent online solution of the multi-objective optimization problem offline.

[0018] Preferably, in step S2:

[0019] All types of sensors transmit data within the system over a large timescale.

[0020] Here, a large timescale refers to a coordination period T. c The first sampling period T within s The beginning moment;

[0021] Only the transmission of the main sensor in each loop is scheduled on a small time scale;

[0022] Among them, the small time scale refers to T c Except for the first T s The start time of sampling periods other than those specified.

[0023] Preferably, in step S3:

[0024] Step S3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and the auxiliary filtering sub-module of the intelligent Kalman filter;

[0025] Step S3.2: Perform sensor anomaly detection based on the 3σ principle of process and measurement noise;

[0026] The intelligent sensing module generates a preset normal range for state estimation based on the 3σ principle of measurement noise and process noise. The module determines whether the state estimation is within the preset normal range of the state. If it exceeds the range, it is initially considered that there is sensor interference.

[0027] Step S3.3: Anomaly detection of sensors is performed on a large time scale through sensor collaboration;

[0028] On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors NS. n，t ;

[0029] Step S3.4: Calculate a reliable state estimate;

[0030] Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules.

[0031] Step S3.5: Online identification of model physical parameters;

[0032] The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module;

[0033] Step S3.6: Smart Kalman filter collaborative update;

[0034] Based on the state estimation results obtained in step S3.4 from Extract the measurement value corresponding to auxiliary sensor m The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with... As the initial state The iterative process used for its next time step.

[0035] Preferably, in step S4:

[0036] The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality PRR. t physical state And the collection of normal sensors NS n，t The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. * Routing set R * , the set of transmission quantities η * ;

[0037] In step S5:

[0038] Step S5.1: The IEEE 802.15.4 network reconfigures the wireless network through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling {s} via a beacon frame at the beginning of the next superframe. * R * η *};

[0039] Step S5.2: The node that receives the beacon frame determines the schedule result {s} * R * η * Update its own scheduling table.

[0040] According to the present invention, an intelligent sensing and communication collaborative design system for an Internet of Things (IoT) control system includes:

[0041] Module M1: Deploys multiple sensors in each control loop and classifies the sensors, and designs the sensor mode table offline;

[0042] Module M2: Performs sensor transmission scheduling based on dual time scales;

[0043] Module M3: Intelligent sensing based on multiple sensors and Kalman filters;

[0044] Module M4: Performs sensor mode switching;

[0045] Module M5: Enables network reconfiguration via the wireless network coordinator.

[0046] Preferably, in module M1:

[0047] Module M1.1: Deploy multiple sensors based on the detectability conditions required by the state estimation;

[0048] Module M1.2: Divides multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal.

[0049] Module M1.3: Solve a multi-objective optimization problem that minimizes the total control and communication costs under different physical states and network link quality, obtains the sensor and its corresponding network configuration, and generates a feasible solution set offline for subsequent online solving of the multi-objective optimization problem.

[0050] Preferably, in module M2:

[0051] All types of sensors transmit data within the system over a large timescale.

[0052] Here, a large timescale refers to a coordination period T. c The first sampling period T within s The beginning moment;

[0053] Only the transmission of the main sensor in each loop is scheduled on a small time scale;

[0054] Among them, the small time scale refers to T c Except for the first T s The start time of sampling periods other than those specified.

[0055] Preferably, in module M3:

[0056] Module M3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and auxiliary filtering sub-module of the intelligent Kalman filter;

[0057] Module M3.2: Sensor anomaly detection based on the 3σ principle of process and measurement noise;

[0058] The intelligent sensing module generates a preset normal range for state estimation based on the 3σ principle of measurement noise and process noise. The module determines whether the state estimation is within the preset normal range of the state. If it exceeds the range, it is initially considered that there is sensor interference.

[0059] Module M3.3: Large-scale anomaly detection of sensors through sensor collaboration;

[0060] On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors NS. n，t ;

[0061] Module M3.4: Calculates reliable state estimates;

[0062] Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules.

[0063] Module M3.5: Online identification of model physical parameters;

[0064] The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module;

[0065] Module M3.6: Intelligent Kalman Filter Co-updating;

[0066] Based on the state estimation results obtained in module M3.4 from Extract the measurement value corresponding to auxiliary sensor m The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with... As the initial state The iterative process used for its next time step.

[0067] Preferably, in module M4:

[0068] The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality PRR. t physical state And the collection of normal sensors NS n，t The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. * Routing set R * , the set of transmission quantities η * ;

[0069] In module M5:

[0070] Module M5.1: In IEEE 802.15.4 networks, wireless network reconfiguration is achieved through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling via a beacon frame at the beginning of the next superframe. * R * η *};

[0071] Module M5.2: The node that receives the beacon frame determines the scheduling result {s} * R * η * Update its own scheduling table.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] 1. This invention proposes a framework for intelligent sensing and communication collaborative design of control systems based on the Industrial Internet of Things. Based on the sensing and transmission modes of the sensor, controlled object and wireless network state switching control loop, it comprehensively handles wireless network interference and sensor interference. The sensing-transmission collaborative design framework solves the sensor and network interference problem in wireless control and ultimately optimizes the overall performance of the control system and wireless network.

[0074] 2. The perception-transmission scheduling not only uses the status of sensors and wireless networks as the basis for scheduling, but also considers the physical dynamics of industrial equipment during operation. When the equipment is in a transition process, the scheduling framework will provide it with more reliable network transmission.

[0075] 3. The intelligent sensing and transmission collaborative design algorithm proposed in this invention fully considers the dynamic evolution process of each industrial device, uses Lyapunov function to quantify the dynamic evolution process and uses it as the basis for scheduling;

[0076] 4. This invention utilizes the collaboration between multiple sensors and state observers to distinguish between sensor interference and physical disturbances experienced by industrial equipment. While effectively resolving abnormal sensor readings, it ensures the timely response of the entire system to physical disturbances to the controlled object. Attached Figure Description

[0077] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0078] Figure 1 A schematic diagram of an intelligent sensing-communication collaborative design framework for an IoT-based wireless control system;

[0079] Figure 2 A diagram illustrating the classification of sensor categories;

[0080] Figure 3 This is a schematic diagram of the intelligent sensing module architecture;

[0081] Figure 4 This is a schematic diagram of a smart Kalman filter structure;

[0082] Figure 5 This is a pseudocode diagram of a sensor anomaly detection algorithm;

[0083] Figure 6 This is a schematic diagram of the sensor mode switching module architecture;

[0084] Figure 7 This is a pseudocode diagram of the sensor mode switching algorithm;

[0085] Figure 8 This is a block diagram of a wireless control system based on the Internet of Things. Detailed Implementation

[0086] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0087] Example 1:

[0088] 1. IEEE 802.15.4 Network Modeling

[0089] This invention is based on the IEEE 802.15.4 standard for wireless network sensing and communication collaborative design. The IEEE 802.15.4 standard has been the most commonly used communication standard for wireless network control systems (IEEE 802.15.4 Standard. Accessed: May 20, 2020. [Online]. Available: https: / / standards.ieee.org / findstds / standard / 802.15.4-2015.html) and is widely used in industrial control. A superframe is a collection of time slots that repeat in time. In beacon-enabled mode, beacon frame transmission begins with the first time slot of each superframe. Beacon frames sent by the coordinator node are used for synchronization of interconnected devices, identification of the wireless network, and description of the superframe structure. During inactive periods, wireless nodes enter a low-power mode. The active period consists of a contention period (CAP) and a contention-free period (CFP). During the contention period, devices use a carrier sense multiple access / collision avoidance (CSMA / CA) media access control scheme. For applications with real-time performance requirements, the network manager provides guaranteed time slots for deterministic transmission during contention-free periods. This invention addresses the real-time performance requirements of IoT-based wireless control systems by placing all sensor superframe scheduling on guaranteed time slots (GTS). The network manager manages the entire wireless network and its nodes. The network manager, device controllers, and network coordinator are typically deployed in the same location or connected via a reliable wired network where packet loss and latency are negligible. The sensor mode switching module is directly deployed within the network manager. Upon receiving an updated sense-communication mode request from the network manager, the coordinator broadcasts the scheduling table via a beacon frame at the beginning of the next superframe. In this way, nodes receiving the beacon update their scheduling tables.

[0090] 2. IoT-based control system framework

[0091] Figure 8 The diagram shows an N-loop wireless control system based on the Internet of Things (IoT). Each loop includes an industrial device, sensors, a state observer, a controller, an actuator, and a wireless network shared with other loops. The sensors measure the physical state y of the industrial device. t The data is then transmitted to the state observer via a wireless network. The observer calculates the state estimate. Controller according to Calculate control command u t And send it to the actuator. The controlled object can be modeled as a discrete-time linear time-invariant system:

[0092] x n，t =A n x n，t-1 +Bn u n，t-1 +W n，t ,

[0093] y n，t =C n x n，t +v n，t ,

[0094] Where t is the time index and n is the cycle index. For state vectors, Let u be the input vector. For the controller, this invention employs a classic feedback control model, i.e., u... n，t =g n (x n，t ), where g n (·) is the control law. It is the output vector. and Both are Gaussian noise, with mean and covariance Q. n ≥0 and R n >0.

[0095] 3. Intermittent observation Kalman filter

[0096] The state estimation process of an intermittent observation Kalman filter consists of two steps: prediction and update. The update step is modified when packets are dropped by the wireless network (Sinopoli B, Schenato L, Franceschetti M, et al. Kalman filtering with intermittent observations[J]. IEEE transactions on Automatic Control, 2004, 49(9):1453-1464.). The end-to-end packet reception rate (PDR) has a critical constraint. To ensure P m，n，t Statistical convergence, This is derived by solving a quasi-convex optimization problem. Therefore, if the current wireless network configuration cannot guarantee the preconditions for filter convergence... The sensing mode switching module should switch to a sensing mode with a higher PDR to ensure the boundedness of state estimation.

[0097] This invention mainly relates to a method for the coordinated design of control and communication in an Internet of Things (IoT) control system, and particularly to a method for the coordinated design of state perception and network uplink transmission scheduling.

[0098] This invention proposes a framework for intelligent sensing and communication collaborative design of an Internet of Things (IoT) control system. It designs an intelligent sensing module based on multi-sensor sensing transmission and intermittent observation Kalman filter to detect sensor interference, and a sensing mode switching module based on sensor status, device physical status and wireless network link status to adjust the sensor and its network configuration online. This achieves the goal of overcoming interference between sensors and wireless networks while the device is running, and ultimately optimizes the overall performance of the wireless network and control system.

[0099] The invention's key features are: 1. The design concept of the intelligent sensing module and the sensing mode switching module, as well as the intelligent sensing communication collaborative design framework.

[0100] 2. Identify physical disturbances and sensor interference in the controlled process through the collaboration of multiple sensors.

[0101] According to the present invention, an intelligent sensing and communication collaborative design method for an Internet of Things (IoT) control system is provided, such as... Figures 1-8 As shown, it includes:

[0102] Step S1: Deploy multiple sensors in each control loop and classify the sensors, and design the sensor mode table offline;

[0103] Specifically, in step S1:

[0104] Step S1.1: Deploy multiple sensors according to the detectability conditions required by the state estimation;

[0105] Step S1.2: Divide the multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal.

[0106] Step S1.3: Solve the multi-objective optimization problem of minimizing the total control and communication costs under different physical states and network link quality, obtain the sensor and its corresponding network configuration, and generate a feasible solution set for subsequent online solution of the multi-objective optimization problem offline.

[0107] Step S2: Perform sensor transmission scheduling based on dual time scales;

[0108] Specifically, in step S2:

[0109] All types of sensors transmit data within the system over a large timescale.

[0110] Here, a large timescale refers to a coordination period T. c The first sampling period T within s The beginning moment;

[0111] Only the transmission of the main sensor in each loop is scheduled on a small time scale;

[0112] Among them, the small time scale refers to T c Except for the first T s The start time of sampling periods other than those specified.

[0113] Step S3: Intelligent sensing based on multiple sensors and Kalman filters;

[0114] Specifically, in step S3:

[0115] Step S3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and the auxiliary filtering sub-module of the intelligent Kalman filter;

[0116] Step S3.2: Perform sensor anomaly detection based on the 3σ principle of process and measurement noise;

[0117] The intelligent sensing module generates a preset normal range for state estimation based on the 3σ principle of measurement noise and process noise. The module determines whether the state estimation is within the preset normal range of the state. If it exceeds the range, it is initially considered that there is sensor interference.

[0118] Step S3.3: Anomaly detection of sensors is performed on a large time scale through sensor collaboration;

[0119] On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors NS. n，t ;

[0120] Step S3.4: Calculate a reliable state estimate;

[0121] Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules.

[0122] Step S3.5: Online identification of model physical parameters;

[0123] The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module;

[0124] Step S3.6: Smart Kalman filter collaborative update;

[0125] Based on the state estimation results obtained in step S3.4 from Extract the measurement value corresponding to auxiliary sensor m The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with... As the initial state The iterative process used for its next time step.

[0126] Step S4: Perform sensor mode switching;

[0127] Specifically, in step S4:

[0128] The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality PRR. t physical state And the collection of normal sensors NS n，t The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. * Routing set R * , the set of transmission quantities η * ;

[0129] Step S5: Network reconstruction is performed by the wireless network coordinator.

[0130] In step S5:

[0131] Step S5.1: The IEEE 802.15.4 network reconfigures the wireless network through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling {s} via a beacon frame at the beginning of the next superframe. * R * η *};

[0132] Step S5.2: The node that receives the beacon frame determines the schedule result {s} * R * η * Update its own scheduling table.

[0133] Example 2:

[0134] Example 2 is a preferred embodiment of Example 1, and is used to illustrate the present invention in more detail.

[0135] The present invention also provides an intelligent sensing and communication collaborative design system for an Internet of Things (IoT) control system. The intelligent sensing and communication collaborative design system for an IoT control system can be implemented by executing the process steps of the intelligent sensing and communication collaborative design method for an IoT control system. That is, those skilled in the art can understand the intelligent sensing and communication collaborative design method for an IoT control system as a preferred embodiment of the intelligent sensing and communication collaborative design system for an IoT control system.

[0136] According to the present invention, an intelligent sensing and communication collaborative design system for an Internet of Things (IoT) control system includes:

[0137] Module M1: Deploys multiple sensors in each control loop and classifies the sensors, and designs the sensor mode table offline;

[0138] Specifically, in module M1:

[0139] Module M1.1: Deploy multiple sensors based on the detectability conditions required by the state estimation;

[0140] Module M1.2: Divides multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal.

[0141] Module M1.3: Solve a multi-objective optimization problem that minimizes the total control and communication costs under different physical states and network link quality, obtains the sensor and its corresponding network configuration, and generates a feasible solution set offline for subsequent online solving of the multi-objective optimization problem.

[0142] Module M2: Performs sensor transmission scheduling based on dual time scales;

[0143] Specifically, in module M2:

[0144] All types of sensors transmit data within the system over a large timescale.

[0145] Here, a large timescale refers to a coordination period T. c The first sampling period T within s The beginning moment;

[0146] Only the transmission of the main sensor in each loop is scheduled on a small time scale;

[0147] Among them, the small time scale refers to T c Except for the first T s The start time of sampling periods other than those specified.

[0148] Module M3: Intelligent sensing based on multiple sensors and Kalman filters;

[0149] Specifically, in module M3:

[0150] Module M3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and auxiliary filtering sub-module of the intelligent Kalman filter;

[0151] Module M3.2: Sensor anomaly detection based on the 3σ principle of process and measurement noise;

[0152] The intelligent sensing module generates a preset normal range for state estimation based on the 3σ principle of measurement noise and process noise. The module determines whether the state estimation is within the preset normal range of the state. If it exceeds the range, it is initially considered that there is sensor interference.

[0153] Module M3.3: Large-scale anomaly detection of sensors through sensor collaboration;

[0154] On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors NS. n，t ;

[0155] Module M3.4: Calculates reliable state estimates;

[0156] Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules.

[0157] Module M3.5: Online identification of model physical parameters;

[0158] The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module;

[0159] Module M3.6: Intelligent Kalman Filter Co-updating;

[0160] Based on the state estimation results obtained in module M3.4 from Extract the measurement value corresponding to auxiliary sensor m The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with... As the initial state The iterative process used for its next time step.

[0161] Module M4: Performs sensor mode switching;

[0162] Specifically, in module M4:

[0163] The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality PRR. t physical state And the collection of normal sensors NS n，t The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. * Routing set R * The set of transmission quantities η;

[0164] Module M5: Enables network reconfiguration via the wireless network coordinator.

[0165] In module M5:

[0166] Module M5.1: In IEEE 802.15.4 networks, wireless network reconfiguration is achieved through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling via a beacon frame at the beginning of the next superframe. * R * η *};

[0167] Module M5.2: The node that receives the beacon frame determines the scheduling result {s} * R * η * Update its own scheduling table.

[0168] Example 3:

[0169] Example 3 is a preferred example of Example 1, and is used to illustrate the present invention in more detail.

[0170] The purpose of this invention is to provide a sensing-communication collaborative design framework that comprehensively addresses interference between sensors and wireless networks, and dynamically schedules sensors and corresponding wireless networks based on the physical state of industrial equipment. This design framework adopts the most advanced "cyber-physical collaborative design" concept in the field of wireless control today.

[0171] The intelligent sensing-communication collaborative design framework in this invention mainly consists of an intelligent sensing module, a sensing mode switching module, and their internal algorithms. For example... Figure 1As shown, the intelligent sensing module provides reliable state estimation, detects and distinguishes between sensor interference and process disturbances in the controlled process, and selects normal sensors based on the state estimates of different sensors and the cooperation of a series of corresponding state observers. The sensing mode switching module selects sensors and their wireless network configurations based on the set of normal sensors provided by the intelligent sensing module, the physical state of the device operation, and the wireless network link status. Since this invention focuses on framework and modular design, the implementation steps are described in detail below with reference to the module diagram and algorithm pseudocode.

[0172] Step 1: Deploy multiple sensors in each control loop and classify them, and design the sensor pattern table offline.

[0173] 1.1) Deploy multiple sensors based on the system detectability requirements specified in the state estimation;

[0174] Consider a discrete-time linear time-invariant state-space equation:

[0175] x n，t =A n x n，t-1 +B n u n，t-1 +w n，t ,

[0176] y n，t =C n x n，t +v n，t ,

[0177] Where t is the sampling time index and n is the control loop index. For state vectors, The input vector. Detectability refers to the ability of the system output y to detect as t→∞. n，t State estimation Converging to the true state x n，t ,Right now Design matrix C based on the PBH criterion of detectability n Make the matrix pair (A) n C n ) satisfies the detectability condition, and then according to matrix C n Select and deploy sensors that measure different physical states.

[0178] 1.2) Divide the multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type (sensing, collaboration) and status (normal, abnormal);

[0179] Based on their status and the tasks they perform, all sensors are divided into three categories, such as Figure 2As shown. The first type of task is perception, where the measurement results are sent to the intelligent sensing module via a wireless network for state estimation; the second type is cooperation, where the sensor measurements are over a large time scale (one coordination period T). c The first sampling period T within s At the start time, the signal is transmitted to the intelligent sensing module, which works with other sensors to detect sensor interference and process disturbances. The sensors are divided into three categories: (1) main sensors, which perform sensing tasks; (2) auxiliary sensors, which participate in sensing and collaborative tasks; and (3) abnormal sensors, which are a set of sensors that detect interference.

[0180] 1.3) Offline design of sensor mode table, that is, solving a multi-objective optimization problem to minimize the total control-communication cost under different physical states and network quality to obtain the sensor and its corresponding network configuration (routing, number of transmissions, etc.).

[0181] 1.3.1) Initialize the pattern table length L and candidate routes R n Number of transmissions η n Main sensor s n ;

[0182] 1.3.2) Determine the feasible physical states (using x) n，t The wireless link quality (represented by packet reception rate) is evenly divided into multiple intervals.

[0183] 1.3.3) By solving a multi-objective optimization problem that minimizes observation error, transmission cost, and network reconstruction cost, sensing mode entries {s, R, η} are generated under different physical and network conditions, where s = {s1, ..., s2}. n , ..., s N}, R={R1,...,R n , ..., R N}, and η={η1,...,η n , ..., η N},s n It is the main sensor of loop n, R n From s n The route to the network access point, η n It is assigned to s n The number of transmissions (quantized in terms of the number of slots within a superframe).

[0184] The multi-objective optimization problem is represented as The objective function is V n，t α is the Lyapunov function value of the nth control loop, representing the dynamic control performance of the controlled object during operation; α is V n，t Weighting coefficients; Represents route R n Total number of hops included, R n，h For multi-hop routing R n The h-th hop of the wireless link; For wireless link R n，h The corresponding PRR; utilizing w represents the transmission cost of loop n. T δ is the transmission cost coefficient; δ is the number of time slots updated within the superframe during each network reconfiguration, used to represent the cost of network scheduling reconfiguration; w G It is the switching cost coefficient, Q n Let be the process noise covariance matrix of loop n, and Tr(·) be the operator for calculating the trace of the matrix. L represents network resource constraints. frame This indicates the superframe length; the total number of transmissions across all control loops at the same time should not exceed L. frame ; The end-to-end packet reception rate (PDR) of a multi-hop network is expressed as p(s). n R n η n The minimum PDR requirement for convergence of the average state error covariance of the Kalman filter under all initial conditions should be met. Specific solution See Appendix reference [6].

[0185] Solving this optimization problem yields the routing set R, the transmission quantity set η, and the sensing mode switching result {s, R, η} of the main sensor s under the initial conditions.

[0186] 1.3.4) Repeat steps 1.3.2) to 1.3.3) until L sensing modes are generated, which will eventually form a sensing mode table and provide a set of feasible solutions for dynamically solving the sensing modes.

[0187] Step 2: Sensor transmission scheduling based on dual time scales, multiple sensors execute transmission strategies at two time scales:

[0188] Large timescales and coordination period T c Related, the coordination period is the sampling period T s A multiple of T c =H×T s ,in During the coordination period T c The first sampling period T within sAt the start, the scheduler schedules all sensors to transmit wirelessly. At the same time, the intelligent sensing module uses an anomaly detection algorithm to detect sensor anomalies and interference in all sensors. The sensing mode switching module updates the main sensor and the corresponding network configuration, such as routing and retransmission.

[0189] On a small timescale (coordination period T) c Except for the first sampling period T s At the start of a sampling period other than the previous one, the main sensor set s transmits the measurement results to the intelligent sensing module via a wireless network according to the network configuration (R and η) determined in the previous large time scale, while other sensors do not have their transmission schedules arranged.

[0190] Step 3: Intelligent Sensing Based on Multi-Sensor and Kalman Filter

[0191] like Figure 3 As shown, the intelligent sensing module integrates an anomaly detector and coordinator to detect sensor interference, a Kalman filter collaborative update module to trigger the update steps of the Kalman filter corresponding to the auxiliary sensor, and an online parameter identification module. Module input and T c Initial Receive State Set A smart Kalman filter is used for state estimation. The anomaly detector and coordinator module uses state estimation from multiple sensors. The system calculates the dynamic range of the normal state estimate, detects sensor interference through inter-sensor cooperation, and generates a set of normal sensors NS on a large time scale for the Kalman filter collaborative update module and the sensing mode switching module. n，t Module state estimate The parameters are sent to the parameter identification module, the intelligent Kalman filter, the sensing mode switching module, and the controller; the parameter estimation results are then sent. The data is sent to the intelligent Kalman filter, the sensing mode switching module, and the controller. The filter co-update module triggers an update of the Kalman filter in the auxiliary sensor.

[0192] 3.1) The intelligent sensing module adopts the intermittent observation Kalman filter algorithm. Reliable state estimation is obtained by designing the main filtering sub-module and the auxiliary filtering sub-module of the intelligent Kalman filter.

[0193] The Kalman filter algorithm is shown below:

[0194]

[0195]

[0196]

[0197]

[0198]

[0199] in C represents the predicted state value at time t. n The system only measures y m，n，t The corresponding output matrix. It is the real physical state x n，t and The error covariance matrix between them, K m，n，t It is Kalman gain. It is the state estimate, P m，n，t It is x n，t and The error covariance matrix between the two variables, binary variables The measurement value y transmitted via wireless network at time t m，n，t Successful reception Or lost if Equal to state prediction Add based on y m，n，t The correction item. If Equal to the state prediction value

[0200] The main filtering submodule and auxiliary filtering submodules of the intelligent Kalman filter are as follows: Figure 4 As shown, the two sub-modules calculate the state estimate respectively. and

[0201] 3.1.1) Prediction and update steps of the main Kalman filter submodule;

[0202] Input sensor measurement value Or in step 3.6.1) Perform state prediction and update to obtain state estimate. If the sensor corresponding to this Kalman filter has been identified as an abnormal sensor (I) m，n，t =1), within a small time scale, this submodule no longer uses the steps in step 3.6.1). For state estimation, the corresponding Kalman filter only performs a prediction step based on outliers.

[0203] 3.1.2) Prediction and update steps of the auxiliary Kalman filter submodule.

[0204] The Kalman filter algorithm (3.1) is also used for state estimation, but the update steps of the auxiliary Kalman filter submodule are modified as follows:

[0205] That is, first input Step 3.2) performs sensor anomaly detection, and the auxiliary Kalman filter submodule uses the obtained anomaly detection result I... m，n，t Determine the update steps for the auxiliary Kalman filter submodule (using...) or ).

[0206] 3.2) Sensor anomaly detection based on the 3σ principle of process and measurement noise;

[0207] 3.2.1) Calculate the normal state estimation interval based on the 3σ principle of noise distribution;

[0208] Will Defined as w respectively n，t and v m，n，t The normal interval is based on the 3σ criterion. A dynamic interval is obtained according to the Kalman filter formula. The normal state estimation range for sensor m, based on the above-mentioned normal range of process and measurement noise pair, is used to characterize the range. The calculation steps for the interval are as follows:

[0209]

[0210]

[0211]

[0212]

[0213] in Noise under normal conditions during maximum process The predicted state value is as follows. This is the corresponding measured value of sensor m at this moment. m It is a row vector used to... extract Indicates in Add the maximum measurement noise on the basis The corresponding measurement value of the rear sensor m.

[0214] KF m，n This represents the intermittent observation Kalman filter algorithm. It uses the infinite norm of the vector ||·||. ∞ The boundaries characterizing the normal state ultimately yield the interval.

[0215] 3.2.2) Determine the measured value from a normal sensor. Received Is it in Within: If exist Inside, set its abnormal sensor flag I m，n，t =0; otherwise, set its sensor flag bit I. m，n，t =1; if the abnormal sensor measurement value q consecutive times remain in the interval Internal (convergence), Set the initial state value of its smart Kalman filter (see step 3.6), and set its anomaly flag I. m，n，t =0.

[0216] 3.3) Sensor anomaly detection through sensor collaboration on a large time scale;

[0217] Once the anomaly detector and coordinator detect and flag one or more anomalous sensors, an inspection is performed on any large timescale. Is it at sensor m′(m′, m′∈IS)? n，t ∪NS n，t The normal state estimation range Inside;

[0218] When a new coordination cycle T c At time t1, if there is an abnormal sensor m∈IS n，t And m′≠m exist Scope, at the same time exist If the system is within a certain range, then a process disturbance has occurred, provided that the sensors m∈IS are within their respective normal ranges. n，t Classified as a normal sensor, I m，n，t =0; the judgment logic can ultimately be expressed by the following formula:

[0219]

[0220] If the abnormal sensor measurement value Still Otherwise, these sensors are all identified as being interfered with. m，n，t =1.

[0221] The pseudocode for steps 3.2) to 3.3) is as follows: Figure 5 As shown.

[0222] 3.4) Calculate a reliable state estimate;

[0223] On a small timescale, if the main sensor does not detect any anomalies, the final estimation result is: If I main，n，t=1, the intelligent observation module uses the state estimation results from the auxiliary Kalman filter submodule to calculate the state prediction based on the previous normal state until the next coordination period T. c At the start time, the main sensor s of loop n n Updated;

[0224] On a large time scale, if I main，n，t =0, the final estimate is If in the previous T c Or at the current moment I main，n，t =1, the sensing mode switching module will select a normal primary sensor from the normal candidate sensors that have successfully transmitted data packets at the current time. The state estimate of the new primary sensor is used as the final state estimate. However, when all sensors are classified as anomalous, the module continues to use the auxiliary Kalman filter submodule of the previous main sensor for calculations. As And set I main，n，t =0, until process interference or sensor interference is classified.

[0225] 3.5) Online identification of model physical parameters; The limited memory least squares algorithm is mainly used for online identification of model physical parameters.

[0226] 3.5.1) Establish a systematic source regression model (ARX):

[0227] A(z -1 )y n，t =B(z) -1 )u n，t +v n，t ,

[0228] in A(z -1 ) and B(z) -1 The coefficient of each term is about θ. n，t The function, for example a1=f1(θ n，t ). n a ,n b They are A(z) -1 ) and B(z) -1 The order of the model is ). The model parameter vector is

[0229] 3.5.2) Input to the parameter identification module and from Extract the required measurement value y m，n，t As input to the identification algorithm (y n，t );

[0230] 3.5.3) Perform parameter identification

[0231]

[0232] P n，t =[I+K n，h h n，t ]P n，t-1

[0233]

[0234] Where K n，t P is the gain matrix of the LMRLS algorithm for loop n at time t. n，t It is the covariance matrix at time t. It is time t The parameter estimation results. It is the LMRLS measurement data vector at time t. This represents the physical state measurement of the device at time t. At each sampling time, LMRLS calculates the model parameter vector. and from Derivation For example,

[0235] Module 3.5.4) will The data is transmitted to the intelligent Kalman filter, controller, and sensing mode switching module to update the model's physical parameters throughout system operation.

[0236] 3.6) Intelligent Kalman filter collaborative update;

[0237] 3.6.1) On a small timescale, the module uses NS n，t The extraction matrix c corresponding to each auxiliary sensor in the matrix m ,from extract And transmit it to the corresponding smart Kalman filter for updating;

[0238] 3.6.2) When the sensor interference ends, the Kalman filter co-update module provides the observer with... As the initial state The iterative process used for its next time step.

[0239] Step 4: Switching the sensing mode

[0240] like Figure 6 As shown, the sensor mode switching module runs the sensor mode switching algorithm at each large time scale. The module input NS... t ={NS 1，t ..., NS n，t , ..., NSN，t} and and the Radio Link Quality Set (PRR) t The main sensor s = {s1, ..., s2} for all loops is selected by solving the combinatorial optimization problem in step 1.3.3). n , ..., s N} and the corresponding network configuration: Route R = {R1, ..., R} n , ..., R N} and transmission quantity scheduling η={η1,...,η n , ..., η N}

[0241] 4.1) The sensing mode switching module is based on PRR. t PDRp for each control loop within L sensing modes i (s n R n η n Update;

[0242] 4.2) Search the sensor mode table and determine mode S i Does the constraint of this multi-objective optimization problem satisfy (see step 1.3.3 for constraints)? If satisfied, calculate the total control and communication cost. in If the conditions are not met, proceed to step 4.2) to determine and calculate the total control and communication cost corresponding to the next sensing mode in the sensing mode table.

[0243] 4.3) After searching for L sensing modes, determine the objective function to be minimized. Sensing mode

[0244] The pseudocode for the sensor mode switching algorithm is available. Figure 7 .

[0245] Step 5: Network reconfiguration is performed by the wireless network coordinator.

[0246] 5.1) In IEEE 802.15.4 networks, wireless network reconfiguration is achieved through a coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling via a beacon frame at the beginning of the next superframe. * R * η *};

[0247] 5.2) The node that receives the beacon frame determines the scheduling result {s}* R * η * Update its own scheduling table.

[0248] Current alternatives often involve independently designing sensor interference detection and filtering, or considering dynamic network scheduling algorithms that address network interference. Currently, there are no solutions, either domestically or internationally, that utilize sensor-communication co-design to address sensor and wireless network interference within IoT-based wireless network control systems.

[0249] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0250] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for intelligent sensing and communication collaborative design of an Internet of Things (IoT) control system, characterized in that, include: Step S1: Deploy multiple sensors in each control loop and classify the sensors, and design the sensor mode table offline; Step S2: Perform sensor transmission scheduling based on dual time scales; Step S3: Intelligent sensing based on multiple sensors and Kalman filters; Step S4: Perform sensor mode switching; Step S5: Network reconstruction is performed by the wireless network coordinator; In step S2: All types of sensors transmit data within the system over a large timescale. Here, a large timescale refers to a coordination cycle. The first sampling period within The beginning moment; Only the transmission of the main sensor in each loop is scheduled on a small time scale; Among them, small time scale refers to Except for the first one The start time of sampling periods other than those specified; In step S3: Step S3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and the auxiliary filtering sub-module of the intelligent Kalman filter; Step S3.2: Based on process and measurement noise The principle is to perform anomaly detection on the sensors; The intelligent sensing module is based on measurement noise and process noise. The principle is to generate a preset normal range for the state estimate. The module determines whether the state estimate is within the preset normal range for that state. If it exceeds the range, it is initially considered that there is sensor interference. Step S3.3: Anomaly detection of sensors is performed on a large time scale through sensor collaboration; On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors. ; Step S3.4: Calculate a reliable state estimate; Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules. ; Step S3.5: Online identification of model physical parameters; The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module; Step S3.6: Smart Kalman filter collaborative update; Based on the state estimation results obtained in step S3.4 ,from Extracting auxiliary sensors Corresponding measurement value The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with the data. As the initial state The iterative process used for its next moment; In step S4: The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality. physical state and a collection of normal sensors The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. Routing sets , number of transmissions ; In step S5: Step S5.1: The IEEE 802.15.4 network reconfigures the wireless network through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling via a beacon frame at the beginning of the next superframe. ; Step S5.2: The node that receives the beacon frame determines the schedule based on the result. Update its own scheduling table.

2. The intelligent sensing and communication collaborative design method for an IoT control system according to claim 1, characterized in that, In step S1: Step S1.1: Deploy multiple sensors according to the detectability conditions required by the state estimation; Step S1.2: Divide the multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal. Step S1.3: Solve the multi-objective optimization problem of minimizing the total control and communication costs under different physical states and network link quality, obtain the sensor and its corresponding network configuration, and generate a feasible solution set for subsequent online solution of the multi-objective optimization problem offline.

3. A collaborative design system for intelligent sensing and communication in an Internet of Things (IoT) control system, characterized in that, include: Module M1: Deploys multiple sensors in each control loop and classifies the sensors, and designs the sensor mode table offline; Module M2: Performs sensor transmission scheduling based on dual time scales; Module M3: Intelligent sensing based on multiple sensors and Kalman filters; Module M4: Performs sensor mode switching; Module M5: Enables network reconfiguration via the wireless network coordinator; In module M2: All types of sensors transmit data within the system over a large timescale. Here, a large timescale refers to a coordination cycle. The first sampling period within The beginning moment; Only the transmission of the main sensor in each loop is scheduled on a small time scale; Among them, small time scale refers to Except for the first one The start time of sampling periods other than those specified; In module M3: Module M3.1: The intelligent sensing module adopts the intermittent observation Kalman filter algorithm, and performs state estimation by designing the main filtering sub-module and auxiliary filtering sub-module of the intelligent Kalman filter; Module M3.2: Based on process and measurement noise The principle is to perform anomaly detection on the sensors; The intelligent sensing module is based on measurement noise and process noise. The principle is to generate a preset normal range for the state estimate. The module determines whether the state estimate is within the preset normal range for that state. If it exceeds the range, it is initially considered that there is sensor interference. Module M3.3: Large-scale anomaly detection of sensors through sensor collaboration; On a large timescale, the occurrence and termination of sensor interference are detected through collaboration among different sensors, distinguishing between sensor anomalies and process disturbances in industrial equipment, thus obtaining a set of normal sensors. ; Module M3.4: Calculates reliable state estimates; Based on the anomaly detection results, reliable state estimates are selected from the deployed Kalman filters and submodules. ; Module M3.5: Online identification of model physical parameters; The limited memory least squares algorithm is used to identify the physical parameters of the model online, and the parameter identification results are then processed. Feedback is sent to the controller, Kalman filter, and sensing mode switching module; Module M3.6: Intelligent Kalman Filter Co-updating; Based on the state estimation results obtained in module M3.4 ,from Extracting auxiliary sensors Corresponding measurement value The feed is sent to its Kalman filter for updating; when the sensor interference ends, the Kalman filter co-update module provides the observer with the data. As the initial state The iterative process used for its next moment; In module M4: The sensing mode switching module performs sensing mode switching at each large time scale, based on the wireless network link quality. physical state and a collection of normal sensors The set of all main sensors for the control loop in the next sampling period is obtained by searching the generated sensor pattern table. Routing sets , number of transmissions ; In module M5: Module M5.1: In IEEE 802.15.4 networks, wireless network reconfiguration is achieved through the coordinator node. After the coordinator node receives the updated sensing mode from the sensing mode switching module, it broadcasts the sensing and transmission scheduling via a beacon frame at the beginning of the next superframe. ; Module M5.2: The node that receives the beacon frame determines the schedule based on the results. Update its own scheduling table.

4. The intelligent sensing and communication collaborative design system for the Internet of Things control system according to claim 3, characterized in that, In module M1: Module M1.1: Deploy multiple sensors based on the detectability conditions required by the state estimation; Module M1.2: Divides multiple sensors in each loop into main sensors, auxiliary sensors, and abnormal sensors according to task type and status; task type includes sensing and collaboration, and status includes normal and abnormal. Module M1.3: Solve a multi-objective optimization problem that minimizes the total control and communication costs under different physical states and network link quality, obtains the sensor and its corresponding network configuration, and generates a feasible solution set offline for subsequent online solving of the multi-objective optimization problem.

Images