Ice mooring profile node adaptive task scheduling method and system

Abstract

The invention discloses an ice mooring section node adaptive task scheduling method and system, and belongs to the technical field of underwater observation, and the method comprises the steps: collecting the state information of a current ice mooring section node; obtaining observation coverage information of a cloud platform or an edge node to construct observation redundancy distribution; calculating observation priorities of the nodes in different depth intervals and time periods based on node state vectors, observation redundancy distribution and business requirements; performing optimization solution to obtain an observation task plan by taking maximization of the total observation value or the unit energy observation value in the residual life of the node as a target; executing the observation task plan, and returning the profile observation data and the updated state information to the cloud platform or the edge node; and dynamic feedback updating is carried out to realize continuous optimization. The method can maximize the observation value under the condition of limited energy, supports multi-node cooperation and cloud closed-loop control, and is especially suitable for long-term, unattended and adaptive underwater environment observation of ice-covered sea areas such as the north pole and the south pole.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of underwater observation, and particularly relates to an ice-tethered profile node adaptive task scheduling method and system. BACKGROUND

[0002] At present, long-term continuous observation of the polar ocean environment mainly relies on ice-tethered profilers (ITP), such as the ITP system developed by WHOI (Woods Hole Oceanographic Institution) and the commercial profiler produced by the Mc Lane company. This kind of equipment can perform vertical profile movement underwater to collect key hydrological parameters such as temperature, salinity and depth. With the development of sensor technology, some devices can integrate dissolved oxygen, optical or acoustic sensors to support the observation of biogeochemical processes. However, with the increasing complexity and refinement of observation requirements, the existing system still has obvious limitations in task scheduling, energy management and collaborative observation, which restricts the further improvement of its observation efficiency in dynamic polar ocean environment.

[0003] Firstly, in terms of task scheduling, most existing ice-tethered profilers rely on fixed or semi-fixed preset programs. The observation task related parameters are usually preset at the factory, or can only be configured through the local serial port in a limited way, and remain fixed throughout the life cycle, or can only be simply adjusted through extremely limited remote instructions. This static scheduling mode lacks the ability to adaptively adjust according to real-time environmental changes, node energy state and real-time scientific events.

[0004] Secondly, in terms of collaborative observation, the existing observation system mainly adopts single node independent operation, and lacks effective communication and collaboration mechanism between nodes, resulting in that when multiple profilers are deployed, the nodes cannot share position, observation results or environmental state information, leading to isolated observation plans, and easily causing problems such as spatial repeated observation or insufficient capture of some key ocean processes (such as frontal passage and ice edge biological activity).

[0005] Thirdly, in terms of energy management, since the polar observation faces severe energy constraints, the observation environment is harsh and energy supply is extremely difficult, and the energy management of the existing system mostly stays at the level of hardware low-power design, mostly without introducing task-level energy optimization strategy, which cannot maximize the observation value under limited battery capacity.

[0006] In addition, although some advanced devices have supported uploading observation data through satellite links such as Iridium, there is still a lack of closed-loop task optimization and feedback mechanism from the cloud to the node, which limits the improvement of the overall efficiency and intelligence level of the observation system.

[0007] Therefore, in response to the scientific challenges brought by the rapid changes in the polar ocean and to improve the sustainability and flexibility of long-term observation, it is urgent to break through the technical paradigm of the existing ITP system and develop a new intelligent ice-based ocean observation system with environmental perception and adaptive task scheduling capability, supporting efficient end-edge-cloud collaboration, to achieve more efficient, flexible and durable stereoscopic observation of the polar ocean environment and provide stronger data support for polar climate research and ecological protection. SUMMARY

[0008] In view of the above, the purpose of the present application is to provide an ice-based profiling node adaptive task scheduling method and system, which realizes node state perception, observation redundancy evaluation, task priority calculation and local optimization scheduling by integrating edge computing and Internet of Things communication, thereby maximizing observation value under limited energy conditions and supporting multi-node collaboration and cloud closed-loop control, especially suitable for long-term, unattended and adaptive underwater environmental observation in ice-covered waters such as the Arctic and Antarctic.

[0009] To achieve the above-mentioned purpose of the application, the technical solutions provided by the present application are as follows: In a first aspect, the ice-based profiling node adaptive task scheduling method provided by the embodiments of the present application comprises the following steps: Periodically collecting the state information of the current ice-based profiling node by using the state monitoring module to form a node state vector and transmitting it to the edge computing unit; Obtaining observation coverage information from the cloud platform or edge node by using the Internet of Things communication module, and constructing the observation redundancy distribution of the target sea area based on the observation coverage information by using the edge computing unit; Calculating the observation priority of the current ice-based profiling node in different depth intervals and time periods based on the node state vector, the observation redundancy distribution and the preset business requirements by using the edge computing unit; The edge computing unit optimizes and solves the observation task plan under constraints based on the observation priority, with the goal of maximizing the total observation value or the unit energy observation value within the remaining life of the node; The ice-based profiling node executes the observation task plan to complete the profile observation, and the profile observation data and the updated node state vector are transmitted back to the cloud platform or edge node by using the Internet of Things communication module; The edge computing unit dynamically updates the energy consumption model and observation priority inside the current ice-based profiling node according to the energy consumption and data quality feedback of this task execution, realizing continuous optimization.

[0010] Preferably, the ice-based profiling node comprises a surface layer cabin, a profiler and a tether, and a message format management module. The surface layer cabin is arranged on the surface or near the surface of sea ice and comprises an Internet of Things communication module, an edge computing unit and a state monitoring module. The Internet of Things communication module is used for connecting with a satellite communication network through a first communication submodule and performing short-range wireless communication with a nearby mother ship, an ice station or a relay node through a second communication submodule. The edge computing unit is in communication connection with the state monitoring module and the Internet of Things communication module. The state monitoring module is used for collecting state information of ice mooring profile nodes. The profiler ascends and descends along the tether and is used for hydrological and / or biogeochemical profile observation in a predetermined depth range. The profiler comprises a driving and attitude control module, a sensor module and a task execution control module. The message format management module is used for encoding and decoding command messages, state messages and data messages received or sent by the Internet of Things communication module. The command messages are used for issuing profile modes, depth intervals, time schedules and sensor configurations. The state messages are used for reporting node positions, power and task execution states. The data messages are used for reporting observation data and quality marks.

[0011] Preferably, the state information at least comprises residual power, power supply voltage, current, temperature, profile time consumption, position information, ice drift speed, historical profile energy consumption, historical standby energy consumption and estimated communication energy consumption.

[0012] Preferably, the observation redundancy distribution of the target sea area is constructed based on observation coverage information, comprising: receiving and analyzing observation coverage information including spatial position, observation time and depth interval information issued by a cloud platform or an edge node, counting the number of completed and planned observation tasks in a specific geographical position, depth and time period , and then generating an observation redundancy distribution reflecting the spatial and temporal coverage saturation of the observation : , wherein, is an indicator function, indicating whether the ice mooring profile node is observed at depth and time , represents a node set.

[0013] Preferably, the observation priority comprises: , wherein, represents the observation priority of the th ice mooring profile node at depth and time , , , Indicates the weighting coefficient; This represents a pre-defined business demand factor, reflecting a certain depth based on scientific objectives and environmental conditions. and time Its importance in business; This represents the observation redundancy distribution factor, reflecting the current region and depth. and time The degree to which something has been observed by other nodes is calculated using the following formula: , in, This indicates the critical value at which the observation redundancy set based on the node density within the region reaches saturation. This indicates the distribution of observation redundancy in the target sea area; This represents the node state factor, reflecting the impact of the node state vector on the ability to perform observation tasks.

[0014] Preferably, the edge computing unit, based on observation priorities, aims to maximize the total observation value or unit energy observation value within the remaining lifetime of a node, and optimizes the solution under constraints to obtain an observation task plan, including: The total observation value objective function is constructed based on the observation priority and the total observation value or unit energy observation value of the node corresponding to the observation action within the next scheduling cycle. The observation value includes at least data scarcity, objective priority, node state adaptability, and urgency of observation events. Construct energy constraints and communication constraints. The goal of energy constraints is to ensure that nodes do not fail due to energy depletion before the next resupply or throughout their entire lifespan. The goal of communication constraints is to ensure that observation data is reliably transmitted back and that the communication behavior itself does not deplete node energy or interfere with task execution. Under energy and communication constraints, the objective function of total observation value is solved, and the solution that maximizes the objective function of total observation value is output as the observation task plan for the node.

[0015] Preferably, the observation task plan includes: profile period, maximum profile depth, key layer encrypted sampling interval, and sensor switching strategy; wherein, the key layer encrypted sampling interval is a specific depth layer identified based on the observation priority and whose scientific value is higher than a set threshold, and data is collected in this depth layer using a smaller sampling interval compared to other water layers; the sensor switching strategy is to dynamically control the power supply and working mode of different sensors according to depth, time, and preset scientific objectives.

[0016] Preferably, the observation task plan also responds to an emergency event triggering command received via the IoT communication module from the cloud or edge node; the emergency event includes extreme weather, rapid sea ice changes, or biomass gathering events; upon receiving the command, the edge computing unit prioritizes executing observation tasks related to the emergency event.

[0017] Preferably, the step of dynamically updating the energy consumption model and observation priority within the current ice-retention profile node based on the energy consumption and data quality feedback of this task includes: The actual energy consumption of this task is compared with the predicted value of the energy consumption model inside the ice-stacked profile node, and the predicted parameters of the model are updated using a lightweight online learning algorithm. Based on the data quality feedback information, including at least the sensor's own diagnosis, data rationality analysis, comparison with external reference data, and task execution status, the weighting factors or business requirement mapping relationships in the observation priority calculation are adjusted.

[0018] Secondly, embodiments of the present invention also provide an adaptive task scheduling system for ice tethered profile nodes, implemented using the aforementioned adaptive task scheduling method for ice tethered profile nodes, comprising: a state perception unit, a collaborative perception unit, an observation and evaluation unit, a task optimization unit, a task execution unit, and a feedback update unit. The state perception unit is used to periodically collect the state information of the current ice tether profile node using the state monitoring module to form a node state vector and transmit it to the edge computing unit. The collaborative sensing unit is used to acquire observation coverage information from the cloud platform or edge nodes using the Internet of Things communication module, and to construct the observation redundancy distribution of the target sea area based on the observation coverage information using the edge computing unit. The observation and evaluation unit is used to calculate the observation priority of the current ice stagnation profile node in different depth ranges and time periods by using the edge computing unit, based on the node state vector, observation redundancy distribution and preset business requirements. The task optimization unit is used to utilize the edge computing unit to optimize and solve the observation task plan under constraints, based on the observation priority, with the goal of maximizing the total observation value or unit energy observation value within the remaining lifetime of the node. The task execution unit is used to execute the observation task plan through the ice-tethered profile node to complete the profile observation, and uses the Internet of Things communication module to transmit the profile observation data and the updated node state vector back to the cloud platform or edge node. The feedback update unit is used to utilize the edge computing unit to dynamically update the energy consumption model and observation priority inside the current ice stagnation profile node based on the energy consumption and data quality feedback of this task, thereby achieving continuous optimization.

[0019] Compared with the prior art, the beneficial effects of the present invention include at least the following: (1) This invention significantly improves energy utilization efficiency and task scheduling flexibility: Through the edge computing unit, the node status and observation redundancy are perceived in real time and tasks are optimized, maximizing the amount of effective observation information obtained per unit of energy under limited battery capacity, effectively extending the continuous observation life of the node in the harsh environment of the polar region; at the same time, it supports the remote configuration of task parameters and the rapid distribution of event-driven tasks through the Internet of Things communication module, and can dynamically adjust the observation strategy according to the sea ice status, scientific task priority or sudden environmental events, and has a high degree of responsiveness.

[0020] (2) This invention realizes network-level collaborative observation and enhances system robustness: Based on cloud-based global observation value assessment and information interaction between edge nodes, it realizes task division and collaborative scheduling of multi-ice tethered profile nodes, avoids spatial duplication of observation, and improves the ability to capture key ocean processes; In addition, relying on the end-edge-cloud three-level architecture, edge nodes can take over regional task optimization and scheduling functions when satellite communication is interrupted, ensuring the continuous operation of the system under extreme conditions, and supporting task redistribution when nodes fail or energy is insufficient, thereby significantly improving the overall system reliability.

[0021] (3) Support for model self-learning and adaptive optimization: The edge computing unit can dynamically update the parameters of the internal energy consumption model and observation priority function based on the energy consumption differences and data quality feedback of the actual task execution, so that the system can gradually improve itself and continuously improve the accuracy of task planning and adaptability to dynamic environment.

[0022] (4) It has good architectural versatility and engineering scalability: The node structure, message format and scheduling method proposed in this invention form a general technical framework that is not only applicable to the Ice-Stayed Profiling Instrument (ITP), but can also be extended to other polar observation platforms such as the Towed Profiling Instrument (TOP) and the Winch system, and has broad engineering promotion value. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an adaptive task scheduling method for ice tethering profile nodes provided in an embodiment of the present invention. Figure 2This is a schematic diagram of the framework of an adaptive task scheduling method for ice tethering profile nodes provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the polar ice-based observation Internet of Things system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an adaptive task scheduling system for ice tethered profile nodes provided in an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0026] The inventive concept of this invention is as follows: Addressing the problems of rigid task scheduling, lack of coordination, and insufficient energy management in existing ice-tethered profiling instruments, this invention provides an adaptive task scheduling method and system for ice-tethered profiling nodes. Its core lies in endowing each node with local intelligence based on edge computing: nodes dynamically calculate observation priorities by sensing their own status and observation redundancy in real time, and autonomously optimize and generate task plans under energy and communication constraints, achieving dynamic perception and optimized decision-making. Simultaneously, through IoT communication and a three-level collaborative architecture of "end-edge-cloud," task division and regional intelligent collaboration among multiple nodes are realized, and continuous self-optimization is achieved using an online learning mechanism, thereby improving the energy efficiency, coordination, flexibility, and environmental adaptability of polar observations.

[0027] like Figure 1 and Figure 2 As shown, the embodiment provides an adaptive task scheduling method for ice tethered profile nodes. By integrating an IoT communication module and an edge computing unit within the ice tethered profile instrument layer cabin, and combining cloud-based full-domain observation value assessment, adaptive observation task adjustment and network-level collaboration among multiple ice tethered profile nodes are achieved. The method includes the following steps: S1 uses the state monitoring module to periodically collect the state information of the current ice tether profile nodes to form node state vectors and transmit them to the edge computing unit.

[0028] In this embodiment, the status monitoring module periodically collects the status information of the current ice tethering profile node. The status information includes at least: remaining power. Power supply voltage Current ,temperature Sectioning time Location information Ice drift speed Historical profile energy consumption Historical standby power consumption Estimated communication energy consumption To form the node state vector : , Among them, subscript Both represent node indices. Remaining battery power. This determines the scale of the tasks performed by the node. Power supply voltage. and current Temperature reflects the health and stability of node operation. The system's node task strategy is determined as follows: when the temperature is below -40℃, the system decides to shut down sensors sensitive to low temperatures, activate the sensor preheating module, and perform only basic CTD observations (Conductive-Temperature-Depth) to avoid equipment damage or data anomalies. When the temperature is between 0℃ and -10℃, the system decides to activate all sensors and increase the sampling frequency to fully utilize environmental conditions and acquire high-quality data. (Profile time consumption) The total time it takes for a node to complete one profile observation directly reflects task execution efficiency. The system will optimize the profile cycle and node operation mode based on this parameter. Location information With ice drift speed Used to adjust the location where nodes execute tasks. Historical profile energy consumption. Compared with historical standby power consumption This reflects energy consumption during historical missions and between mission intervals, used to predict future node energy consumption and calculate remaining node lifetime. Estimated communication energy consumption. It represents the cost of sending existing data and is used for decisions regarding communication methods.

[0029] S2 uses the Internet of Things (IoT) communication module to obtain observation coverage information from the cloud platform or edge nodes, and uses the edge computing unit to construct the observation redundancy distribution of the target sea area based on the observation coverage information.

[0030] In this embodiment, the edge computing unit receives observation coverage information from other ice-tethered profile nodes within a certain area, such as spatial location, most recent profile time, and observed depth range, from the cloud platform or nearby edge nodes via an IoT communication module. Based on this information, an observation redundancy distribution function for the target sea area is constructed. This function is used to statistically analyze data at a specific geographical location and depth. and time period The number of observation profiles that have been completed or are planned to be completed is used to determine the redundancy, which is positively correlated with the number of profiles. The specific calculation method is as follows: , in, The indicator function represents the ice tethered profile node. In depth and time Whether observations were conducted Represents a set of nodes.

[0031] A sort of The rules for determining the value are as follows: , Higher redundancy indicates more comprehensive observation coverage in that region and time, thus its observation value is relatively low. Edge computing units dynamically update by parsing the received message format. .

[0032] S3 utilizes edge computing units to calculate the observation priority of the current ice stagnation profile node in different depth ranges and time periods based on node state vectors, observation redundancy distribution, and preset business requirements.

[0033] In this embodiment, the edge computing unit uses the node state vector Observation redundancy distribution In conjunction with pre-defined operational requirements, such as seasonality rules, sea ice conditions, and one or more pre-defined priority observation depth ranges, the observation priority function for this node is calculated. Used to quantitatively describe nodes At a specific depth in the target sea area , specific time The priority of observation needs is determined. Seasonal rules are preset either manually before system deployment or remotely via a cloud platform, enabling the system to dynamically adjust its observation focus and strategies based on typical oceanographic phenomena and scientific objectives in different seasons. Observation priority function. Represented as: , Among them, the weighting coefficient It can be adjusted according to specific application scenarios.

[0034] This represents a pre-defined business demand factor, reflecting a certain depth based on scientific objectives and environmental conditions. and time In practical applications, the specific values ​​of business requirement factors can be determined by a mapping table to assess their importance in the business context. A feasible mapping table is shown in Table 1. Table 1. A feasible business demand factor mapping table

[0035] This represents the observation redundancy distribution factor, reflecting the current region and depth. and time The degree to which observations have already been observed by other nodes; the greater the observation redundancy, the smaller the observation priority value. The calculation formula is as follows: , in, This indicates the critical value at which the observation redundancy set based on the node density within the region reaches saturation. It needs to be set according to the node density within the region; the more nodes there are, the higher the redundancy level. The larger the value, the more likely it is to be, typically ranging from 3 to 5. The redundancy at a certain spatiotemporal depth exceeds... It is judged as excessive redundancy.

[0036] This represents the node state factor, reflecting the impact of the node state vector on the observation task's execution capability. The better the node state matches the observation task, the higher the observation priority value. For example, in the remaining battery power... Increase priority when sufficient, or reduce profile time. Priority is increased for shorter timeframes, and the quantitative relationship between the status parameters of each node and the node status factor can be set manually in advance according to task requirements.

[0037] S4, the edge computing unit optimizes the observation task plan under constraints based on observation priority, with the goal of maximizing the total observation value or unit energy observation value within the remaining lifetime of the node.

[0038] In this embodiment, the edge computing unit will observe the priority function. This serves as a direct basis for measuring the value of a single observation action. Assume there are a total of [number missing] actions within a future scheduling cycle. K Each candidate observation action, Corresponding depth and time Its observational value is Energy consumption is Introducing decision variables Indicates whether to perform the action. Specifically, it is expressed as follows: , If the objective is to maximize the total observation value over the remaining lifetime of a node, then this periodic optimization problem can be formulated as follows: , in, The maximum energy consumption allowed for this cycle is determined based on the ratio of the node's remaining energy to its remaining lifetime.

[0039] If the objective is to maximize the observation value per unit energy within the remaining lifetime of a node, then the optimization problem is: , By solving the above optimization problem, we can obtain the set of observation actions that maximize the objective, i.e., the observation task plan.

[0040] In this embodiment, the edge computing unit constructs a total observation value objective function based on the total observation value or unit energy observation value of the node within its remaining lifetime corresponding to the node's observation action, determined according to the observation priority, within a future scheduling cycle. Simultaneously, energy constraints and communication constraints are constructed. The energy constraint aims to ensure that the node does not fail due to energy depletion before the next resupply or throughout its entire lifetime. The communication constraint aims to ensure reliable transmission of observation data and that communication activities themselves do not deplete node energy or interfere with task execution.

[0041] Under energy and communication constraints, the objective function of the total observation value is solved, and the solution that maximizes the objective function is output as the observation task plan for each node. The observation task plan specifically includes: profile period... Maximum profile depth The study includes: intensified sampling intervals for key layers (such as the thermocline and the chlorophyll-maximum layer); and sensor switching strategies (such as activating high-power sensors at specific depths). Intensified sampling of key layers involves not using a uniform sampling interval across the entire water column during vertical profile observations, but rather conducting dense data collection at smaller depth intervals in specific depth layers with the highest scientific value and most dramatic changes. The sensor switching strategy involves intelligently controlling the power supply and operating status of different sensors based on depth, time, and scientific objectives, rather than keeping all sensors constantly on throughout the profile observation process.

[0042] The optimization solution can be obtained using a priority-based greedy algorithm or a reinforcement learning model based on a Markov decision process (MDP). In the reinforcement learning model, the state is a node state vector. The action is the observation task plan, and the reward function is the observation value minus the energy penalty. The edge computing unit selects an appropriate algorithm based on computing power limitations and outputs the task plan after solving the problem.

[0043] Furthermore, the observation mission plan also responds to emergency event triggering commands received via the IoT communication module from the cloud or edge nodes, such as extreme weather, rapid sea ice changes, or biomass gathering events. Upon receiving the command, the edge computing unit prioritizes the observation tasks related to the emergency event and adjusts and optimizes the objectives.

[0044] S5, the ice-system profile node executes the observation task plan to complete the profile observation, and uses the Internet of Things communication module to transmit the profile observation data and the updated node state vector back to the cloud platform or edge node.

[0045] In this embodiment, the observation task plan is sent to the task execution control module of the profiler body for execution. The task execution control module, based on the plan control drive, attitude control module, and sensor module, completes the profile observation. The obtained observation data consists of raw environmental parameters collected by the sensor modules on the profiler body during underwater profile movement, and is related to the types of sensors integrated on the profiler. After the profile observation is completed, the observation data and updated node state vectors are transmitted back to the cloud platform via the IoT communication module. The data message includes the observation data and quality markers, while the status message includes the latest battery level, location, etc., for the cloud to perform the next round of optimization.

[0046] S6 utilizes edge computing units to dynamically update the energy consumption model and observation priority within the current ice-bound profile node based on the energy consumption and data quality feedback of this task, thereby achieving continuous optimization.

[0047] In this embodiment, the edge computing unit dynamically updates its internal energy consumption model and observation priority function based on the difference between the actual and estimated energy consumption of the current task execution, as well as the quality feedback of the sensor data. The parameters are as follows. Quality feedback includes: sensor self-diagnosis, data rationality analysis, comparison with external reference data, and mission execution status. External reference data includes: synchronous observation data from other nearby nodes, historical average data, and satellite remote sensing data.

[0048] An energy consumption model is a mathematical function or a set of functions used to predict the energy required to execute a planned observation task. A feasible energy consumption model function is as follows: , in, The total energy consumption for a node to perform a single complete observation task, in relation to ambient temperature. related, It is the energy consumption of profile observation, and the maximum profile depth. The set of sensors that have been activated related, It is the data communication consumption, and the amount of communication data. Communication frequency Related, It is standby power consumption, which is equal to the product of the node's standby power and the standby duration. It is a temperature correction factor, which is determined by humans based on the actual situation.

[0049] Profile observation energy consumption is the most significant energy source for nodes, encompassing two core energy-consuming components: drive lifting and sensor operation. The calculation logic is as follows: , , , in, It refers to the energy consumption of the drive module in raising and lowering the profiler, and the maximum profile depth. Directly related The rated power of the driver module is determined by the selected hardware itself. It is the total time for the profiler to rise and fall, and A linear relationship exists. It is the total energy consumption of the sensors activated during the observation process, which is related to the sensor type, the number activated, and the operating time. It is a sensor The rated operating power is determined by the sensor selection. This is the sensor's operating time.

[0050] The data communication consumption is calculated as follows: , in, It is the energy consumption for short-range communication, which is the energy consumption for communication between a node and its neighboring edge nodes. It is equal to the product of the rated power of the short-range communication module, the duration of a single communication, and the number of communication times within a task cycle. It is the energy consumption of satellite communication and the amount of data. The relationship is linear, and the calculation method is as follows: , in, This is the rated power of the satellite communication module. It refers to the amount of data transmitted in a single communication. It refers to the satellite communication rate.

[0051] Energy consumption model and observation priority function The dynamic updating of parameters employs lightweight online learning algorithms, including: recursive least squares, exponentially weighted recursive least squares with a forgetting factor, and online stochastic gradient descent.

[0052] For example, recursive least squares can be used to update the parameters of the energy consumption model: , in, and These are the model parameters before and after the update, respectively. The learning rate. The observation priority function. The weights of different factors are adjusted based on data quality feedback (such as sensor calibration error and data anomaly rate) to gradually improve the accuracy of task planning and environmental adaptability.

[0053] like Figure 3As shown, the embodiment specifically provides a polar ice-based observation IoT system that coordinates "end, edge, and cloud" to implement the above-mentioned adaptive task scheduling method for ice tethered profile nodes, specifically including: ice tethered profile nodes, cloud platform, and edge nodes.

[0054] The ice-tethered profiling node includes: a surface module, a profiler and mooring rope, and a message format management module. The surface module is deployed on or near the sea ice surface and includes: an IoT communication module, an edge computing unit, and a status monitoring module. The IoT communication module connects to a satellite communication network via a first communication submodule and conducts short-range wireless communication with nearby mother ships, ice stations, or relay nodes via a second communication submodule. The first communication submodule of the IoT communication module supports polar-orbiting satellite communication protocols, including Iridium and BeiDou. The second communication submodule supports UHF (Ultra High Frequency), WiFi, or relay links with the mother ship or ice station.

[0055] The edge computing unit communicates with the status monitoring module and the IoT communication module to generate a local observation task plan for the next cycle based on the node status information provided by the status monitoring module and the network-level observation redundancy information obtained through the IoT communication module.

[0056] The status monitoring module is used to collect status information of ice tether profile nodes.

[0057] The profiler ascends and descends along a tether for hydrological and / or biogeochemical profiling within a predetermined depth range. Internally, the profiler includes a drive and attitude control module, a sensor module, and a mission execution control module. The drive and attitude control module drives the profiler body to ascend and descend along the tether. The sensor module collects environmental parameter data during profiling observations and includes CTD sensors, dissolved oxygen sensors, optical sensors, or acoustic sensors. The mission execution control module, connected to the drive and attitude control module and the sensor module, controls the profiling observation operations according to the received observation mission plan.

[0058] The message format management module encodes and decodes command messages, status messages, and data messages received or sent by the IoT communication module. Command messages are used to issue profile modes, depth ranges, time schedules, and sensor configurations. Status messages are used to report node locations, battery levels, and task execution status. Data messages are used to report observation data and quality tags. The command messages defined by the message format management module include: profile mode ID, depth range, time schedule, and sensor configuration parameters. Status messages include node locations, battery levels, and task execution status. Data messages include observation data and data quality tags. Messages support multicast or regional broadcast via region ID or group ID.

[0059] Edge nodes, deployed on ice stations or ships, act as regional hubs, connecting ice-tethered profiling nodes within the region, aggregating their status and observation data, and taking over the cloud platform's optimization functions when satellite links are interrupted or latency is too high. Based on observation redundancy and node status within the region, they perform local task re-optimization and dynamic allocation. When a node in the region is detected to be faulty or severely underpowered, the edge node can reassign some of its core observation tasks to other functional nodes within the region, ensuring the robustness of the observation network.

[0060] The cloud platform communicates with the ice-tethered profile nodes and the edge nodes, enabling it to estimate the observation uncertainty field of the target sea area based on a global numerical model or data assimilation system. , The coordinates of the target sea area are indicated; areas with higher uncertainty have greater observational value. The cloud platform integrates the status information of all nodes with the global uncertainty field, calculates task weights for each node, and generates node-level task constraints, such as the suggested number of profiles to be completed and the target depth zone within the next 24 hours. These constraints are then distributed to the corresponding ice-tethered profile nodes or edge nodes via the IoT communication module. The cloud platform supports sending multicast or regional broadcast commands to multiple ice-tethered profile nodes via region ID or group ID to achieve rapid deployment of collaborative observation strategies. When specific events occur, including extreme weather events, rapid sea ice change events, or other pre-set scientific observation opportunities, the cloud platform can issue event-driven observation tasks to one or more edge nodes or ice-tethered profile nodes. The cloud platform can also predict the remaining lifetime of ice-tethered profile nodes based on historical observation data and node status data, and dynamically adjust the energy allocation strategy in the task constraints based on the prediction results.

[0061] Based on the same inventive concept, such as Figure 4 As shown, this embodiment of the invention also provides an adaptive task scheduling system 400 for ice tethered profile nodes, including: a state perception unit 410, a collaborative perception unit 420, an observation and evaluation unit 430, a task optimization unit 440, a task execution unit 450, and a feedback update unit 460.

[0062] The state awareness unit 410 is used to periodically collect the state information of the current ice tether profile node using the state monitoring module to form a node state vector and transmit it to the edge computing unit.

[0063] The collaborative sensing unit 420 is used to acquire observation coverage information from the cloud platform or edge nodes using the Internet of Things communication module, and to construct the observation redundancy distribution of the target sea area based on the observation coverage information using the edge computing unit.

[0064] The observation and evaluation unit 430 is used to calculate the observation priority of the current ice stagnation profile node in different depth ranges and time periods by utilizing the edge computing unit, based on the node state vector, observation redundancy distribution and preset business requirements.

[0065] The task optimization unit 440 is used to utilize the edge computing unit to optimize and solve the observation task plan under constraints, based on the observation priority, with the goal of maximizing the total observation value or the observation value per unit energy within the remaining lifetime of the node.

[0066] The task execution unit 450 is used to execute the observation task plan through the ice-fixed profile node to complete the profile observation, and uses the Internet of Things communication module to transmit the profile observation data and the updated node state vector back to the cloud platform or edge node.

[0067] The feedback update unit 460 is used to utilize the edge computing unit to dynamically update the energy consumption model and observation priority inside the current ice stagnation profile node based on the energy consumption and data quality feedback of this task, thereby achieving continuous optimization.

[0068] It should be noted that the adaptive task scheduling system for ice mooring profile nodes provided in the above embodiments belongs to the same inventive concept as the adaptive task scheduling method for ice mooring profile nodes. For details of its implementation process, please refer to the embodiment of the adaptive task scheduling method for ice mooring profile nodes, which will not be repeated here.

[0069] The specific embodiments described above illustrate the technical solution and beneficial effects of the present invention in detail. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An ice series profiling node self-adaptive task scheduling method, characterized in that, The method comprises the following steps: Periodically collecting state information of the current ice-stay profile node by the state monitoring module to form a node state vector and transmit to the edge computing unit; Obtaining observation coverage information from the cloud platform or edge node by the Internet of Things communication module, and constructing an observation redundancy distribution of the target sea area based on the observation coverage information by the edge computing unit; Calculating the observation priority of the current ice-stay profile node in different depth intervals and time periods based on the node state vector, the observation redundancy distribution and the preset business demand by the edge computing unit; Optimizing and solving the observation task plan under the constraints based on the observation priority by the edge computing unit, with the goal of maximizing the total observation value or the unit energy observation value within the remaining life of the node; Executing the observation task plan by the ice-stay profile node to complete the profile observation, and returning the profile observation data and the updated node state vector to the cloud platform or edge node by the Internet of Things communication module; Dynamically updating the energy consumption model and observation priority in the current ice-stay profile node based on the energy consumption and data quality feedback of this task execution by the edge computing unit, to realize continuous optimization.

2. The ice containment profile node adaptive task scheduling method of claim 1, wherein, The ice-stay profile node comprises a surface cabin, a profile instrument and a tether, and a message format management module; The surface cabin is deployed on the sea ice surface or near the surface, and comprises an Internet of Things communication module, an edge computing unit and a state monitoring module; the Internet of Things communication module is used to connect with the satellite communication network through a first communication submodule, and to perform short-range wireless communication with nearby mother ships, ice stations or relay nodes through a second communication submodule; the edge computing unit is in communication connection with the state monitoring module and the Internet of Things communication module; the state monitoring module is used to collect state information of the ice-stay profile node; The profile instrument is used to perform hydrological and / or biogeochemical profile observation in a predetermined depth range, and comprises a driving and attitude control module, a sensor module and a task execution control module inside the profile instrument; The message format management module is used to encode and decode command messages, state messages and data messages received or sent by the Internet of Things communication module, wherein the command messages are used to issue profile modes, depth intervals, time schedules and sensor configurations, the state messages are used to report node positions, power and task execution states, and the data messages are used to report observation data and quality marks.

3. The ice containment profile node adaptive task scheduling method of claim 1, wherein, The state information at least includes residual power, power supply voltage, current, temperature, profile time consumption, position information, ice drift speed, historical profile energy consumption, historical standby energy consumption and estimated communication energy consumption.

4. The ice containment profile node adaptive task scheduling method of claim 1, wherein, The observation redundancy distribution of the target sea area is constructed based on the observation coverage information, which comprises: Receiving and analyzing the observation coverage information including spatial position, observation time and depth interval information issued by the cloud platform or edge node, counting the number of completed and planned observation tasks in a specific geographic location, depth and time period , and then generating an observation redundancy distribution reflecting the observation spatial and temporal coverage saturation : ， wherein, is an indicator function, indicating ice retention profile nodes at depth and time whether an observation was made, denotes a set of nodes.

5. The ice containment profile node adaptive task scheduling method according to claim 1 or 4, characterized in that, The observation priority comprises: ， wherein, represents the observation priority of the i-th ice-related profile node at depth and time , , , , represents the weight coefficient; represents the preset service demand factor, reflecting the importance of a certain depth and time in the service according to the scientific target and environmental conditions; represents the observation redundancy distribution factor, reflecting the degree of the current region and depth and time having been observed by other nodes, and the calculation formula is as follows: ， wherein, represents a critical value at which the observation redundancy set according to the node density within the region reaches saturation, represents an observation redundancy distribution of the target sea area; represents a node state factor reflecting the influence of the node state vector on the observed task execution capability.

6. The ice containment profile node adaptive task scheduling method of claim 1, wherein, The edge computing unit optimizes and solves the observation task plan under the constraints based on the observation priority, with the goal of maximizing the total observation value or the unit energy observation value within the remaining life of the node, which comprises: A total observation value objective function is constructed based on the total observation value or the unit energy observation value of the node remaining life corresponding to the node observation action determined based on the observation priority in the future one scheduling period, and the observation value at least includes data scarcity, target priority, node state adaptation degree, and urgency of observation event; An energy constraint and a communication constraint are constructed, where the target of the energy constraint is to ensure that the node does not fail due to energy depletion before the next replenishment or throughout the life cycle, and the target of the communication constraint is to ensure that the observation data is reliably returned and the communication behavior itself does not deplete the node energy or interfere with task execution; The total observation value objective function is solved under the energy constraint and the communication constraint, and the solution that maximizes the total observation value objective function is output as the observation task plan of the node.

7. The ice containment profile node adaptive task scheduling method according to claim 1 or 6, characterized in that, The observation task plan includes a profile period, a maximum profile depth, a key layer encryption sampling interval, and a sensor switching strategy; the key layer encryption sampling interval is a specific depth layer identified based on the observation priority, where the scientific value is higher than a set threshold, and a smaller sampling interval is used in the depth layer than in other water layers for data collection; and the sensor switching strategy is to dynamically control the power supply and working mode of different sensors according to depth, time, and a preset scientific goal.

8. The ice containment profile node adaptive task scheduling method according to claim 1 or 6, characterized in that, The observation task plan also responds to an emergency event trigger instruction received from the cloud or the edge node through the Internet of Things communication module; the emergency event includes extreme weather, rapid change of sea ice, or biological aggregation event; after receiving the instruction, the edge computing unit preferentially executes the observation task related to the emergency event.

9. The ice containment profile node adaptive task scheduling method of claim 1, wherein, The energy consumption model and the observation priority in the current ice-staying profile node are dynamically updated according to the energy consumption and data quality feedback of the current task execution, including: The actual energy consumption of the current task execution is compared with the predicted value of the energy consumption model in the ice-staying profile node, and a lightweight online learning algorithm is used to update the prediction parameters of the model; According to the data quality feedback information, including at least sensor self-diagnosis, data rationality analysis, comparison with external reference data, and task execution status, the weight factor or the business demand mapping relationship in the observation priority calculation is adjusted.

10. An ice-retaining profile node adaptive task scheduling system, implemented by using the ice-retaining profile node adaptive task scheduling method in any one of claims 1-9. It includes: a state perception unit, a collaborative perception unit, an observation evaluation unit, a task optimization unit, a task execution unit, and a feedback updating unit; The state perception unit is used to periodically collect the state information of the current ice-staying profile node by using the state monitoring module to form a node state vector and transmit it to the edge computing unit; The collaborative perception unit is used to obtain observation coverage information from the cloud platform or the edge node by using the Internet of Things communication module, and to construct the observation redundancy distribution of the target sea area based on the observation coverage information by using the edge computing unit; The observation evaluation unit is used to calculate the observation priority of the current ice-staying profile node in different depth intervals and time periods based on the node state vector, the observation redundancy distribution, and the preset business demand by using the edge computing unit; The task optimization unit is configured to utilize the edge computing unit to solve the observation task plan by optimization under constraints based on the observation priority, with the goal of maximizing the total observation value or the unit energy observation value within the remaining life of the node; The task execution unit is configured to execute the observation task plan by the ice system retention profile node to complete the profile observation, and utilize the Internet of Things communication module to return the profile observation data and the updated node state vector to the cloud platform or the edge node; The feedback updating unit is configured to utilize the edge computing unit to dynamically update the energy consumption model and the observation priority in the current ice system retention profile node according to the energy consumption and the data quality feedback of the current task execution, so as to realize continuous optimization.

Images