Cloud edge collaboration-based water conservancy flood control monitoring management system
The cloud-edge collaborative water conservancy flood control monitoring and management system collects and analyzes the health information of edge nodes in real time, determines load adjustment and early warning strategies, solves the problem of insufficient identification of the causes of edge node overload in the existing technology, improves the system's response rate, load adjustment accuracy and early warning accuracy, and ensures the reliability and stability of flood control early warning.
Patent Information
- Application Number
- CN202611065728.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-17
- Publication Date
- 2026-11-10
AI Technical Summary
The existing water conservancy flood control monitoring system lacks an identification mechanism when the edge node is overloaded, which leads to a mismatch between the data processing measures taken and the actual cause of the overload, resulting in ineffective degradation, continuous obstruction of the flood control early warning function and inability to recover autonomously, resulting in low system reliability.
The water conservancy flood control monitoring and management system based on cloud-edge collaboration is adopted. The load perception module collects the health information of edge nodes in real time, the edge computing module analyzes the network health, the cloud management module determines whether to start load adjustment, the load adjustment module decides frequency adjustment or cloud takeover based on the load index, the flood control early warning module generates early warning index and starts the scheduling decision module, and the scheduling decision module determines mode switching or threshold adjustment.
This improved the response rate and load adjustment accuracy of the water conservancy flood control monitoring and management system to the data acquisition quality of edge nodes, enhanced the accuracy of early warning results and the stability of the system, and ensured the reliability and autonomous recovery capability of flood control early warning.
Smart Images

Figure CN123072166A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of flood control monitoring and management, and in particular to a water conservancy flood control monitoring and management system based on cloud-edge collaboration. Background Technology
[0002] With the development of IoT, edge computing, and AI technologies, intelligent sensing and prediction technologies are increasingly being applied to flood control monitoring in water conservancy. In scenarios such as reservoirs and rivers, existing flood control monitoring systems primarily rely on centralized cloud-based processing architectures or simple edge-based direct data acquisition models. However, existing technologies suffer from the following problems in practical applications: First, existing early warning systems typically only use basic resource usage such as CPU and memory as the basis for determining node status, failing to incorporate data quality-related parameters such as image quality and node synchronization rate into the detection mechanism. This leads to a disconnect between the identification of edge node load status and data acquisition quality. Second, when nodes experience overload, existing technologies lack mechanisms to identify the causes of overload, making it impossible to distinguish the root cause of the fault. They often adopt a uniform frequency reduction or offloading strategy, resulting in a mismatch between the implemented strategy and the underlying cause, failing to effectively alleviate node congestion. Third, the lack of a closed-loop verification mechanism after implementing optimization measures makes it impossible to confirm whether the strategy has solved the fundamental problem, potentially causing the system to operate inefficiently for extended periods.
[0003] Chinese Patent Publication No. CN104092750A discloses a wireless monitoring station for flood control and rainfall conditions and its monitoring method, including the following steps: First, a solar charging control module controls a solar panel to charge a battery, and sets a clamping voltage threshold for the solar charging control module; Second, a measuring device collects one or more of the following data: water level, rainfall, flow rate, temperature, and humidity in the rainfall area, and uploads the collected data to a monitoring and control terminal; Third, a video device collects video images and on-site photos, and uploads them to the monitoring and control terminal; Fourth, the monitoring and control terminal analyzes and determines whether the uploaded collected data, video images, and on-site photos have been received, whether the data has been received completely, and whether the network is available, and uploads the complete received data to the superior dispatch center and cloud computing terminal; Fifth, the superior dispatch center and cloud computing terminal determine whether the received data exceeds the warning threshold. If it exceeds the warning threshold, an alarm is triggered in time; otherwise, the monitoring is completed.
[0004] However, existing technologies do not incorporate the load of edge nodes themselves into the system evaluation framework. They only detect whether the data is received completely and whether the network is available. They do not consider indicators that affect data quality, such as the signal-to-noise ratio of video images and the time synchronization rate of sensors. When the signal-to-noise ratio of edge nodes decreases or the time synchronization deviates, the system cannot accurately detect it and still performs prediction and scheduling according to the normal process, resulting in a decrease in the reliability of the early warning results. Summary of the Invention
[0005] To address this, the present invention provides a cloud-edge collaborative water conservancy flood control monitoring and management system to overcome the problem in the prior art that, in scenarios of real-time water conservancy flood control early warning, the lack of an identification mechanism for the cause of overload of edge nodes leads to a mismatch between the data processing measures taken and the actual cause of overload, resulting in ineffective degradation, continuous obstruction of flood control early warning functions and inability to recover autonomously, thus causing low reliability of the water conservancy flood control early warning system.
[0006] To achieve the above objectives, the present invention provides a cloud-edge collaborative water conservancy flood control monitoring and management system, comprising: The load sensing module is used to collect health information of several edge nodes in real time; wherein, the health information includes the signal-to-noise ratio of video images and the time synchronization rate between distributed sensors. An edge computing module is used to analyze network health based on the health information; A cloud management module is used to determine whether to activate the load adjustment module based on the network health status; The load regulation module is used to determine the start-up frequency regulation unit or the cloud takeover unit based on the load index. The flood warning module is used to determine the flood warning index based on the water level prediction curve generated by hydrological information, and to determine whether to activate the dispatch decision module based on the flood warning index; wherein, the hydrological information includes the water level, inflow and rainfall of the target water area in historical periods; The scheduling decision module is used to determine the activation mode switching unit or the activation threshold adjustment unit based on the flood warning index.
[0007] Furthermore, the edge computing module also includes: The health determination unit is used to determine the network health based on the signal-to-noise ratio and time synchronization rate.
[0008] Furthermore, the cloud management module also includes: A health detection unit is used to determine whether to activate the load adjustment module based on the network health status; wherein the condition for activating the load adjustment module is that the network health status is less than or equal to the network health status threshold.
[0009] Furthermore, the load adjustment module also includes: The load index calculation unit is used to determine the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient. An edge node detection unit is used to determine whether to start a frequency adjustment unit or a cloud takeover unit based on the load index; wherein, the condition for determining to start the frequency adjustment unit is that the load index is less than or equal to a load index threshold; and the condition for determining to start the cloud takeover unit is that the load index is greater than the load index threshold.
[0010] Furthermore, the load adjustment module also includes: The frequency adjustment unit is used to determine the adjustment range of the data acquisition frequency based on the Lipschitz constant; The re-detection unit is used to re-detect the network health after the adjustment range of the determined data acquisition frequency has been completed. If the network health is still less than or equal to the network health threshold, the cloud takeover unit is activated.
[0011] Furthermore, the load adjustment module also includes: Takeover index calculation unit, which is used to determine the takeover index based on CPU utilization and the interaction frequency of adjacent edge nodes; The cloud takeover unit is used to determine the adjustment range of the proportion of edge node tasks taken over by the cloud based on the takeover index.
[0012] Furthermore, the load adjustment module also includes: The detection unit is used to re-detect the network health under the condition that the adjustment range of the proportion of the task of taking over the edge node by the cloud has been completed. If the network health is still less than or equal to the network health threshold, the flood prevention and early warning module is activated.
[0013] Furthermore, the flood warning module also includes: The early warning index determination unit is used to obtain the expected arrival time of the flood peak and the flood peak water level value based on the water level prediction curve generated based on the hydrological information, and to determine the flood control early warning index based on the expected arrival time of the flood peak and the flood peak water level value. A flood warning detection unit is used to determine whether to control the scheduling decision module to activate the mode switching unit or the threshold adjustment unit based on the flood warning index. The condition for determining whether to control the scheduling decision module to activate the mode switching unit is that the flood warning index is greater than the flood warning index threshold. The condition for determining whether to control the scheduling decision module to activate the threshold adjustment unit is that the flood warning index is less than or equal to the flood warning index threshold.
[0014] Furthermore, the scheduling decision module also includes: A threshold adjustment unit is used to determine the adjustment range of the network health threshold based on the flood warning index; The mode switching unit is used to change the processing mode of the edge node for video data from real-time analysis of video data to timed analysis of screenshot data.
[0015] Furthermore, the scheduling decision module also includes: A threshold recovery unit is used to adjust the network health threshold to the initial network health threshold when the detected water level is less than or equal to the water level threshold and the predicted water level within a preset time period is less than or equal to the water level threshold. The data backhaul unit is used to read the temporarily stored video data during the flood warning period from the cloud and backhaul it to the edge node in the order of timestamps, provided that the flood warning has been lifted.
[0016] Compared with existing technologies, the beneficial effects of this invention are that it provides a cloud-edge collaborative water conservancy flood control monitoring and management system, including a load perception module, an edge computing module, a cloud management module, a load adjustment module, a flood control early warning module, and a scheduling decision module. This invention uses the load perception module to collect real-time health information of several edge nodes; the edge computing module to analyze network health based on the health information; the cloud management module to determine whether to activate the load adjustment module based on network health; the load adjustment module to determine whether to activate the frequency adjustment unit or the cloud takeover unit based on the load index; the flood control early warning module to determine the flood control early warning index based on the water level prediction curve generated from hydrological information, and to determine whether to activate the scheduling decision module based on the flood control early warning index; and the scheduling decision module to determine whether to activate the mode switching unit or the threshold adjustment unit based on the flood control early warning index. This invention overcomes the problem in existing technologies where, in real-time flood control early warning scenarios, the lack of a mechanism to identify the causes of edge node overload leads to a mismatch between the data processing measures taken and the actual overload causes, resulting in ineffective degradation, continuous obstruction of the flood control early warning function, and inability to recover autonomously, leading to low reliability of the water conservancy flood control early warning system.
[0017] In particular, this invention uses the load sensing module to collect the signal-to-noise ratio of video images from edge nodes and the time synchronization rate between distributed sensors in real time as health information. The edge computing module determines the network health based on the health information, and the cloud management module determines whether to activate the load adjustment module based on the network health. This improves the response rate of the water conservancy flood control monitoring and management system to the decline in the quality of data collection from edge nodes.
[0018] In particular, the present invention determines the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient through a load adjustment module, and determines whether to activate the frequency adjustment unit or the cloud takeover unit based on the load index. This enables the detection of the load status of edge nodes, thereby improving the accuracy of load adjustment in the water conservancy flood control monitoring and management system.
[0019] In particular, this invention collects water level, inflow, and rainfall data of the target water area within a historical period as hydrological information through a flood warning module. Based on the hydrological information, it generates a water level prediction curve, obtains the expected arrival time of the flood peak and the flood peak water level, and calculates a flood warning index. At the same time, it determines whether to activate the scheduling decision module based on the flood warning index. This allows the edge node to complete the water conservancy flood control prediction task while informing the cloud of the reliability of a single prediction task, thereby improving the accuracy of the verification of the early warning results of the edge node of the water conservancy flood control monitoring and management system.
[0020] In particular, the present invention, through a scheduling decision module, determines whether to activate the mode switching unit or the activation threshold adjustment unit based on the flood warning index. This enables the implementation of a computing resource allocation strategy to address the surge in data volume triggered by flood warnings, thereby relaxing detection standards and reducing the data processing volume of video analysis. This improves the stability of flood warnings in the water conservancy flood control monitoring and management system. Attached Figure Description
[0021] Figure 1 This is a structural block diagram of the cloud-edge collaborative water conservancy flood control monitoring and management system according to an embodiment of the present invention; Figure 2 This is a logic diagram illustrating how the load adjustment module is activated based on network health in an embodiment of the present invention. Figure 3 This is a logic diagram illustrating how the frequency adjustment unit is activated based on the load index in an embodiment of the present invention. Figure 4 This is a logic diagram illustrating how the flood warning index determines whether to activate the dispatch decision module in an embodiment of the present invention. Detailed Implementation
[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.
[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.
[0024] It should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the term "connected" should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral connection; it can refer to a mechanical connection or an electrical connection; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0025] Please see Figure 1 The diagram shown is a structural block diagram of a cloud-edge collaborative water conservancy flood control monitoring and management system according to an embodiment of the present invention. The present invention provides a cloud-edge collaborative water conservancy flood control monitoring and management system, comprising: The load sensing module is used to collect health information of several edge nodes in real time; wherein, the health information includes the signal-to-noise ratio of video images and the time synchronization rate between distributed sensors. An edge computing module, which is connected to the load awareness module, is used to analyze network health based on the health information; A cloud management module, which is connected to the edge computing module, is used to determine whether to activate the load adjustment module based on the network health status; A load adjustment module, which is connected to the load sensing module, the edge computing module and the cloud management module respectively, is used to determine whether to start the frequency adjustment unit or the cloud takeover unit based on the load index; wherein, the load index is determined based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes and the channel multiplexing coefficient; A flood warning module, connected to the load regulation module, is used to determine a flood warning index based on a water level prediction curve generated from hydrological information, and to determine whether to activate the dispatch decision module based on the flood warning index; wherein, the hydrological information includes the water level, inflow, and rainfall of the target water area within a historical period; The scheduling decision module, which is connected to the flood warning module, is used to determine the activation mode switching unit or threshold adjustment unit based on the flood warning index.
[0026] Specifically, the load sensing module further includes: The signal-to-noise ratio (SNR) acquisition unit includes several cameras deployed at monitoring points in the target water area. It determines the SNR of the video image based on the ratio of the intensity of image information to noise signal in the acquired video image signal. This determines the prominence of effective information in the video image. A higher SNR indicates that the effective information in the video image is more prominent relative to the noise signal, and the more accurately the system performs video analysis tasks such as water level gauge recognition and floating object detection. Conversely, a lower SNR indicates more severe noise interference in the image, and lower reliability of the video analysis results. The synchronization rate calculation unit, which is a network time protocol server deployed at the edge node, determines the time synchronization rate between distributed sensors based on the average clock deviation between the network time protocol server and each distributed sensor. This determines the consistency of the time reference of the data collected by the radar water level gauges, tipping bucket rain gauges, and acoustic Doppler current profilers under the same edge node. The higher the time synchronization rate, the higher the consistency of the time reference of the data collected by each instrument, the higher the time alignment accuracy of the data points for the water level prediction curve fitted by the system, and the better the spatiotemporal consistency of the prediction results. Conversely, the lower the time synchronization rate, the lower the consistency of the time reference of the data collected by each instrument, the lower the time alignment accuracy of the data points for the water level prediction curve fitted by the system, and the worse the spatiotemporal consistency of the prediction results. The clock deviation is obtained based on the timestamp messages periodically exchanged between each distributed sensor and the master clock.
[0027] Specifically, the flood warning module also includes: The water level acquisition unit includes several radar water level gauges deployed at the center of the reservoir area, the inlet section, and the outlet section of the target water area. These gauges are used to determine the water level of the target water area based on the vertical elevation of the free water surface relative to the flood season limit water level, thereby detecting the water storage status of the target water area. The higher the water level, the greater the water storage in the target water area, and the closer the system is to the flood control warning state. Conversely, the lower the water level, the smaller the water storage in the target water area, and the relatively lighter the flood control pressure on the system. The flow calculation unit, which is an acoustic Doppler current profiler deployed at the inlet of the target water area, is used to determine the inflow rate of the target water area based on the volume of water passing through the inlet section of the target water area per unit time, so as to detect the intensity of the upstream water flow to the target water area. The larger the inflow rate, the more abundant the upstream water flow, the greater the reservoir capacity pressure, and the larger the amount of flood control data processed by the system. The smaller the inflow rate, the gentler the upstream water flow, the relatively smaller the reservoir regulation pressure, and the smaller the amount of flood control data processed by the system. The rainfall acquisition unit includes several tipping bucket rain gauges deployed at watershed surface rainfall monitoring stations in the target water area. These rain gauges determine the rainfall in the target water area based on the product of the number of tipping buckets and the tipping bucket volume, thereby detecting the precipitation intensity. The greater the rainfall, the faster the surface runoff flows into the target water area, the stronger the driving effect on water level rise, and the higher the frequency of hydrological data reporting by the system. Conversely, the smaller the rainfall, the more dispersed the precipitation within the watershed, the weaker the driving effect on water level in the target water area, and the lower the frequency of hydrological data reporting by the system.
[0028] This invention provides a cloud-edge collaborative water conservancy flood control monitoring and management system. Through a load sensing module, it collects real-time health information from several edge nodes; through an edge computing module, it analyzes network health based on the health information; through a cloud management module, it determines whether to activate a load adjustment module based on network health; through the load adjustment module, it determines whether to activate a frequency adjustment unit or a cloud takeover unit based on a load index; through a flood warning module, it determines a flood warning index based on a water level prediction curve generated from hydrological information, and determines whether to activate a scheduling decision module based on the flood warning index; through a scheduling decision module, it determines whether to activate a mode switching unit or a threshold adjustment unit based on the flood warning index, thereby improving the reliability of the water conservancy flood control monitoring and management system in monitoring and warning of flood conditions at water conservancy facilities.
[0029] Specifically, the edge computing module further includes: The health determination unit is used to determine the network health based on the signal-to-noise ratio and time synchronization rate, in order to determine the data acquisition quality of edge nodes and the reliability of LSTM model prediction results. The lower the network health, the worse the integrity of the hydrological data collected by the edge nodes for the target water area and the lower the accuracy of the LSTM model prediction results. The lower the reliability of the flood control early warning decision generated by the system based on the data stored by the edge nodes. The higher the network health, the better the edge nodes' ability to collect hydrological data for the target water area, generate hydrological prediction curves by the LSTM model, and execute scheduling instructions. The higher the reliability of the flood control early warning decision generated by the system based on the data stored by the edge nodes. Specifically, the health determination unit further includes: The signal-to-noise ratio (SNR) determination subunit is used to determine the SNR based on the ratio of the SNR to the SNR threshold, in order to detect the clarity of the effective information in the video image. The larger the SNR, the more obvious the effective information in the video image, and the higher the accuracy of the system in performing water level gauge recognition and floating object detection. The smaller the SNR, the less obvious the effective information in the video image, and the lower the accuracy of water level gauge recognition and floating object detection. The synchronization rate deviation determination subunit is used to determine the synchronization rate deviation ratio based on the ratio of the time synchronization rate to the time synchronization rate threshold, in order to detect the time consistency of the data collected by each distributed sensor. The larger the synchronization rate deviation ratio, the higher the time reference alignment accuracy of the data collected by each distributed sensor, and the smaller the fitting error of the water level prediction curve constructed by the system based on the data collected by the distributed sensors. Conversely, the smaller the synchronization rate deviation ratio, the lower the time reference alignment accuracy of the data collected by each distributed sensor, the lower the accuracy of the water level-flow relationship curve fitting, and the larger the fitting error of the water level prediction curve constructed by the system based on the data collected by the distributed sensors.
[0030] The health determination subunit is used to determine the network health based on the sum of the signal-to-noise ratio and the synchronization rate deviation.
[0031] In this embodiment, the signal-to-noise ratio threshold is determined based on the statistical distribution of the signal-to-noise ratio of video images during the flood season and non-flood season in the historical database, combined with the minimum image clarity requirement required for OCR recognition of water level gauges; the time synchronization rate threshold is determined based on the maximum value of the calculation error of the water level-flow relationship caused by the time synchronization deviation of distributed sensors.
[0032] This invention uses an edge computing module to analyze the signal-to-noise ratio (SNR) of video images and the time synchronization rate between distributed sensors into signal-to-noise deviation ratio (SNR) and synchronization rate deviation ratio (SNR), respectively, and sums them to determine network health. In rainy weather, the camera's field of view is blurred and the data collection volume of each distributed sensor surges, causing both the SNR of video images and the time synchronization rate between distributed sensors to decrease simultaneously. Both factors contribute to a decline in network health. This invention can utilize the quality parameters of the data collected by the edge nodes themselves to detect whether the edge nodes have the conditions for stable operation, thereby improving the accuracy of the water conservancy flood control monitoring and management system in detecting the load status of edge nodes.
[0033] Please see Figure 2 As shown, this is a logic diagram for determining whether to activate the load balancing module based on network health in an embodiment of the present invention. The cloud management module in this embodiment of the present invention also includes: A health detection unit is used to determine whether to activate the load adjustment module based on the network health status; wherein the condition for activating the load adjustment module is that the network health status is less than or equal to the network health status threshold.
[0034] In this embodiment, the network health threshold is determined based on the statistical distribution of network health in the normal startup state of edge nodes in the historical database, combined with the critical point where the probability of data loss or prediction failure of nodes increases significantly when the network health is below this value.
[0035] In this embodiment of the invention, the cloud management module determines to start the load adjustment module only when the network health status is less than or equal to the network health status threshold. This avoids erroneous triggering of adjustment operations when the load status of edge nodes is normal, thereby improving the accuracy of the water conservancy flood control monitoring and management system in adjusting the load of edge nodes.
[0036] Please see Figure 3 As shown, this is a logic diagram for determining whether to activate the frequency adjustment unit based on the load index in an embodiment of the present invention. The load adjustment module in this embodiment of the present invention further includes: The load index calculation unit is used to determine the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient, so as to determine the congestion level of the wireless communication environment in which the edge nodes are located and the degree of local computing resource occupation. The smaller the load index, the more edge nodes there are, the more frequent the interaction, or the higher the degree of channel reuse, which leads to the congestion or even overload of the system's communication link. The larger the load index, the more CPU resources the edge nodes occupy due to performing tasks such as LSTM model prediction and video image analysis, resulting in insufficient local computing resources of the system. An edge node detection unit, connected to the load index calculation unit, is used to determine whether to activate the frequency adjustment unit or the cloud takeover unit based on the load index. The condition for activating the frequency adjustment unit is that the load index is less than or equal to a load index threshold; the condition for activating the cloud takeover unit is that the load index is greater than the load index threshold.
[0037] Specifically, the load index calculation unit further includes: The density deviation ratio calculation subunit is used to determine the distribution density of edge nodes based on the ratio of the number of adjacent edge nodes that can exchange data with the same edge node within the communication range to the communication coverage area, and to determine the density deviation ratio based on the ratio of the distribution density of edge nodes to the distribution density threshold, so as to detect the congestion level of the wireless communication environment in which the system is located. The larger the density deviation ratio, the more nodes there are per unit area, the more congested the wireless communication environment in which the system is located, and the fewer communication time slots that each edge node can obtain per unit time. The smaller the density deviation ratio, the fewer nodes there are per unit area, the more sufficient the communication bandwidth that a single edge node can obtain in the wireless channel, and the more idle the wireless communication environment in which the system is located. The frequency deviation ratio determination subunit is used to determine the interaction frequency of adjacent edge nodes based on the number of successful data exchanges between an edge node and its neighboring edge nodes per unit time, and to determine the frequency deviation ratio based on the ratio of the interaction frequency of adjacent edge nodes to the interaction frequency threshold, so as to detect the level of data communication activity between adjacent edge nodes in the system. The larger the frequency deviation ratio, the more active the data exchange between adjacent edge nodes, the higher the probability of data packet collision on the system's wireless channel, and the greater the risk of data congestion on the communication link. The smaller the frequency deviation ratio, the less active the data exchange between adjacent edge nodes, the higher the idle level of the communication link, and the lower the probability of data packet collision on the system's wireless channel. The channel multiplexing deviation ratio determination subunit is used to determine the channel multiplexing coefficient based on the ratio of the channel busy time to the total detection time on the same wireless channel, and to determine the channel multiplexing deviation ratio based on the ratio of the channel multiplexing coefficient to the channel multiplexing coefficient threshold, so as to detect the degree to which the same wireless channel is used by multiple edge nodes in an overlapping manner. The larger the channel multiplexing deviation ratio, the higher the degree of channel overlap, the greater the risk of multiple edge nodes in the system simultaneously occupying the same wireless channel for data transmission, and the lower the transmission stability of the communication link. The smaller the channel multiplexing deviation ratio, the lower the degree of channel overlap, the lower the risk of multiple edge nodes simultaneously occupying the same wireless channel for data transmission, and the higher the transmission stability of the communication link. The load index determination subunit is used to determine the load index based on the sum of the density offset ratio, frequency offset ratio, and channel multiplexing offset ratio.
[0038] In this embodiment, the distribution density threshold is determined based on the upper limit of the number of adjacent nodes that can stably maintain data exchange within the communication coverage area of the edge node in the water conservancy flood control monitoring scenario; the interaction frequency threshold is determined based on the upper limit of the number of data packet exchanges that the edge node communication interface can normally handle under high concurrency conditions during the flood season; the channel multiplexing coefficient threshold is determined based on the upper limit of the proportion of busy time of the wireless channel that can maintain stable transmission under the carrier sense multi-point access mechanism; and the load index threshold is determined based on the statistical distribution of the load index of edge nodes in the historical database under the two states of communication link overload and insufficient local computing resources, combined with the critical value that has the highest accuracy in distinguishing between the two states.
[0039] In this embodiment of the invention, the load adjustment module determines the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient. When a large number of edge nodes simultaneously report hydrological monitoring data due to torrential rains during the flood season, the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient will all increase simultaneously, leading to an increase in the load index. At the same time, the load index determines whether to activate the frequency adjustment unit or the cloud takeover unit, which can quantify the actual load of edge nodes into a load index, thereby improving the accuracy of load adjustment of edge nodes in the water conservancy flood control monitoring and management system.
[0040] Specifically, the load adjustment module further includes: A frequency adjustment unit, connected to the edge node detection unit, is used to determine the adjustment range of the data acquisition frequency based on the Lipschitz constant, so as to reduce the number of data packets sent from each distributed sensor to the edge node in a single cycle. The re-detection unit, which is connected to the frequency adjustment unit, is used to re-detect the network health after the adjustment range of the determined data acquisition frequency has been completed. If the network health is still less than or equal to the network health threshold, the cloud takeover unit is activated. The takeover index calculation unit is connected to the edge node detection unit and the re-detection unit respectively. It is used to determine the takeover index based on the CPU utilization and the interaction frequency of adjacent edge nodes to reduce the load of the local node. The larger the takeover index, the higher the CPU utilization of the edge node and the more active the interaction between adjacent edge nodes, and the greater the load of the edge node. The smaller the takeover index, the lower the CPU utilization of the edge node and the less active the interaction between adjacent edge nodes, and the smaller the load of the edge node. The cloud takeover unit is connected to the edge node detection unit and the takeover index calculation unit respectively, and is used to determine the adjustment range of the proportion of edge node tasks taken over by the cloud based on the takeover index, so as to migrate the edge node tasks to the cloud server. The detection unit, which is connected to the cloud takeover unit, is used to re-detect the network health under the condition that the proportion of tasks to be determined for cloud takeover edge nodes has been adjusted. If the network health is still less than or equal to the network health threshold, it indicates that the flood warning has caused a surge in data volume. In this case, the flood warning module is activated and the flood warning is triggered.
[0041] Specifically, the frequency adjustment unit also includes: The data change determination subunit is used to determine the Lipschitz constant based on the maximum ratio of the data change to the interval duration between several adjacent collection times within the collection period, and to determine the Lipschitz deviation ratio based on the ratio of the Lipschitz constant to the Lipschitz constant threshold, in order to detect the data processing load of the system. The larger the Lipschitz deviation ratio, the faster the rate of water level change between adjacent collection times in the hydrological information time series, and the more violent the water situation fluctuations, requiring a higher data collection frequency to fully record the flood peak rising process; the smaller the Lipschitz deviation ratio, the slower the rate of water level change between adjacent collection times in the hydrological information time series, and the more gradual the water situation changes, allowing for an appropriate reduction in the data collection frequency. The frequency adjustment subunit is used to determine the reduction range of the data acquisition frequency based on the product of the Lipschitz bias ratio and the original data acquisition frequency.
[0042] In this embodiment, the Lipschitz constant threshold is determined based on the statistical distribution of the rate of water level change of conventional rivers and reservoirs during the flood season and non-flood season in the historical hydrological database, combined with the critical value at which the information on the flood peak rise process is lost when the rate of water level change exceeds this value; the original data acquisition frequency is determined based on the data acquisition frequency requirements of the target water area during the flood season and non-flood season.
[0043] Specifically, the takeover of the index calculation unit also includes: The utilization rate deviation ratio determination subunit, which is a task manager deployed on the edge node, is used to determine the CPU utilization rate based on the proportion of time the edge node's central processing unit spends processing tasks per unit time, and to determine the utilization rate deviation ratio based on the ratio of the CPU utilization rate to the CPU utilization rate threshold, so as to detect the CPU resource utilization level of the edge node. The larger the utilization rate deviation ratio, the longer the edge node's CPU is in a high utilization state, and the less the system has remaining available computing resources; the smaller the utilization rate deviation ratio, the shorter the edge node's CPU is in a high utilization state, and the more the system has remaining available computing resources. The takeover index determination sub-unit is used to determine the takeover index based on the product of the occupancy deviation ratio and the frequency deviation ratio.
[0044] In this embodiment, the CPU utilization threshold is determined based on the maximum CPU utilization of the edge node embedded processor that can maintain real-time response to interrupts and emergency tasks under high concurrency during the flood season; the interaction frequency threshold is determined based on the upper limit of the number of data packet exchanges that the edge node communication interface can normally handle under high concurrency during the flood season.
[0045] In this embodiment, the cloud takeover unit determines the proportion of edge node tasks to be taken over by the cloud based on the product of the maximum takeover ratio and the takeover index, so as to adjust the workload of edge nodes migrating to the cloud. Preferably, the maximum takeover ratio is determined based on the minimum proportion of computing resources reserved by the edge nodes for performing the two core functions of hydrological information collection and real-time triggering of flood warnings.
[0046] The embodiments of the present invention improve the accuracy of load regulation in the water conservancy flood control monitoring and management system by determining the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes and the channel reuse coefficient through a load adjustment module, and determining whether to activate the frequency adjustment unit or the cloud takeover unit based on the load index.
[0047] Please see Figure 4 As shown, this is a logic diagram illustrating the determination of whether to activate the dispatch decision module based on the flood warning index in an embodiment of the present invention. The flood warning module in this embodiment of the present invention further includes: The early warning index determination unit is used to obtain the expected arrival time of the flood peak and the flood peak water level value based on the water level prediction curve generated by the hydrological information, and to determine the flood control early warning index based on the expected arrival time of the flood peak and the flood peak water level value, so as to detect the urgency of the flood situation. The larger the flood control early warning index, the higher the flood peak water level and the shorter the arrival time of the flood peak. The greater the increase in the amount of hydrological information data processed by the system, it indicates that the flood peak water level has reached the water level corresponding to the design flood control standard and the arrival time of the flood peak has entered the emergency response window. It is necessary to switch the real-time video analysis to timed screenshots to release computing resources to ensure the core hydrological data processing. The smaller the flood control early warning index, the smaller the increase in the amount of hydrological information data processed by the system, indicating that the current water situation has not reached the level of emergency dispatch and only the network health detection standard needs to be relaxed to cope with the slight increase in the amount of data. A flood warning detection unit, connected to the warning index determination unit, is used to determine whether to control the dispatch decision module to activate the mode switching unit or the threshold adjustment unit based on the flood warning index. The condition for controlling the dispatch decision module to activate the mode switching unit is that the flood warning index is greater than the flood warning index threshold. The condition for controlling the dispatch decision module to activate the threshold adjustment unit is that the flood warning index is less than or equal to the flood warning index threshold.
[0048] Specifically, the early warning index determination unit further includes: The water level index determination subunit is used to determine the water level index based on the ratio of the water level threshold to the flood peak water level, in order to predict the degree of exceedance of the future flood peak water level. The arrival time index determination subunit is used to determine the arrival time index based on the ratio of the flood peak arrival time threshold to the expected flood peak arrival time, in order to predict the urgency of the future flood peak arrival. The early warning index determination sub-unit is used to determine the flood control early warning index based on the sum of the water level index and the arrival time index.
[0049] In this embodiment, the water level threshold is determined based on the design flood control standard of the target water area. When the target water area is a reservoir, the water level threshold is selected as the flood limit water level; when the target water area is a river, the water level threshold is selected as the warning water level. The flood peak arrival time threshold is determined based on the shortest propagation time of the historical flood peak from the upstream monitoring section to the downstream protection section of the target water area, combined with the effective forecast period of short-term hydrological forecasts. The flood control warning index threshold is determined based on the statistical distribution of events corresponding to different flood control warning levels of the target water area in the historical database, combined with the critical value with the highest accuracy in classifying warning levels.
[0050] In this embodiment of the invention, the flood warning module obtains the expected arrival time of the flood peak and the flood peak water level value by generating a water level prediction curve based on hydrological information, and determines the flood warning index. At the same time, based on the flood warning index, it determines whether to control the scheduling decision module to start the mode switching unit or start the threshold adjustment unit, thereby improving the reliability of the water conservancy flood control monitoring and management system for monitoring hydrological information of the target water area.
[0051] Specifically, the scheduling decision module further includes: A threshold adjustment unit is used to determine the adjustment range of the network health threshold based on the ratio of the flood warning index to the flood warning index threshold, so as to temporarily relax the detection standard of the water conservancy flood control monitoring and management system for the network health of edge nodes. A mode switching unit, which is connected to the threshold adjustment unit, is used to change the processing mode of the edge node for video data from real-time analysis of video data to timed analysis of screenshot data, so as to reduce the amount of data processed by the edge node for video data. A threshold recovery unit, connected to the threshold adjustment unit, is used to adjust the network health threshold to the initial network health threshold under the condition that the detected water level is less than or equal to the water level threshold and the predicted water level within a preset time period is less than or equal to the water level threshold, so as to lift the flood warning. The data feedback unit, which is connected to the mode switching unit, is used to read the video data temporarily stored during the flood warning period from the cloud when it is determined that the flood warning has been lifted, and to send it back to the edge node in the order of timestamps so as to continue to monitor the load status of the edge node.
[0052] In this embodiment of the invention, the scheduling decision module continuously monitors the water level and the predicted water level within a preset time period after completing the scheduling decision. When both are less than or equal to the water level threshold, the flood warning is lifted. At the same time, the load detection threshold of the edge node is adjusted to the initial benchmark, and the video data temporarily stored during the flood warning period is read from the cloud and transmitted back to the local edge node in batches according to the timestamp order. The automatic recovery mechanism after the flood warning is lifted solves the problem of the water conservancy flood control monitoring and management system still relaxing the detection standards after the flood warning is lifted, thereby improving the adaptability of the water conservancy flood control monitoring and management system at different water conditions.
[0053] It is understood that the acquisition devices, communication protocols, calculation methods, and threshold setting methods involved in the above modules are all preferred embodiments of the present invention in specific application scenarios. In actual deployment, the signal-to-noise ratio of video images can be calculated through edge node post-processing algorithms, sensor time synchronization can be achieved using a precise time protocol, hydrological information acquisition can be replaced by pressure level gauges, ultrasonic flow meters, or other equivalent devices, and the calculation method of load index-related parameters can be adjusted according to the actual network environment; all of these are equivalent embodiments of the present invention.
[0054] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.
Claims
1. A cloud-edge collaborative water conservancy flood control monitoring and management system, characterized in that, include: The load sensing module is used to collect health information of several edge nodes in real time; wherein, the health information includes the signal-to-noise ratio of video images and the time synchronization rate between distributed sensors. An edge computing module is used to analyze network health based on the health information; A cloud management module is used to determine whether to activate the load adjustment module based on the network health status; The load regulation module is used to determine the start-up frequency regulation unit or cloud takeover unit based on the load index. The flood warning module is used to determine the flood warning index based on the water level prediction curve generated by hydrological information, and to determine whether to activate the dispatch decision module based on the flood warning index; wherein, the hydrological information includes the water level, inflow and rainfall of the target water area in historical periods; The scheduling decision module is used to determine the activation mode switching unit or the activation threshold adjustment unit based on the flood warning index.
2. The water conservancy flood control monitoring and management system based on cloud-edge collaboration according to claim 1, characterized in that, The edge computing module also includes: The health determination unit is used to determine the network health based on the signal-to-noise ratio and time synchronization rate.
3. The water conservancy flood control monitoring and management system based on cloud-edge collaboration according to claim 2, characterized in that, The cloud management module also includes: A health detection unit is used to determine whether to activate the load adjustment module based on the network health status; wherein the condition for activating the load adjustment module is that the network health status is less than or equal to the network health status threshold.
4. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 3, characterized in that, The load adjustment module also includes: The load index calculation unit is used to determine the load index based on the distribution density of edge nodes, the interaction frequency of adjacent edge nodes, and the channel reuse coefficient. An edge node detection unit is used to determine whether to start a frequency adjustment unit or a cloud takeover unit based on the load index; wherein, the condition for determining to start the frequency adjustment unit is that the load index is less than or equal to a load index threshold; and the condition for determining to start the cloud takeover unit is that the load index is greater than the load index threshold.
5. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 4, characterized in that, The load adjustment module also includes: The frequency adjustment unit is used to determine the adjustment range of the data acquisition frequency based on the Lipschitz constant; The re-detection unit is used to re-detect the network health after the adjustment range of the determined data acquisition frequency has been completed. If the network health is still less than or equal to the network health threshold, the cloud takeover unit is activated.
6. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 5, characterized in that, The load adjustment module also includes: Takeover index calculation unit, which is used to determine the takeover index based on CPU utilization and the interaction frequency of adjacent edge nodes; The cloud takeover unit is used to determine the adjustment range of the proportion of edge node tasks taken over by the cloud based on the takeover index.
7. The water conservancy flood control monitoring and management system based on cloud-edge collaboration according to claim 6, characterized in that, The load adjustment module also includes: The detection unit is used to re-detect the network health under the condition that the adjustment range of the proportion of the task of taking over the edge node by the cloud has been completed. If the network health is still less than or equal to the network health threshold, the flood prevention and early warning module is activated.
8. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 7, characterized in that, The flood warning module also includes: The early warning index determination unit is used to obtain the expected arrival time of the flood peak and the flood peak water level value based on the water level prediction curve generated based on the hydrological information, and to determine the flood control early warning index based on the expected arrival time of the flood peak and the flood peak water level value. A flood warning detection unit is used to determine whether to control the scheduling decision module to activate the mode switching unit or the threshold adjustment unit based on the flood warning index. The condition for determining whether to control the scheduling decision module to activate the mode switching unit is that the flood warning index is greater than the flood warning index threshold. The condition for determining whether to control the scheduling decision module to activate the threshold adjustment unit is that the flood warning index is less than or equal to the flood warning index threshold.
9. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 8, characterized in that, The scheduling decision module also includes: A threshold adjustment unit is used to determine the adjustment range of the network health threshold based on the flood warning index; The mode switching unit is used to change the processing mode of the edge node for video data from real-time analysis of video data to timed analysis of screenshot data.
10. The cloud-edge collaborative water conservancy flood control monitoring and management system according to claim 9, characterized in that, The scheduling decision module also includes: A threshold recovery unit is used to adjust the network health threshold to the initial network health threshold when the detected water level is less than or equal to the water level threshold and the predicted water level within a preset time period is less than or equal to the water level threshold. The data backhaul unit is used to read the temporarily stored video data during the flood warning period from the cloud and backhaul it to the edge node in the order of timestamps, provided that the flood warning has been lifted.
Citation Information
Patent Citations
Flood defense rainwater condition wireless monitoring station and monitoring method thereof
CN104092750A
