Electromagnetic spectrum energy harvesting and optimization method for low-power multi-sensor cooperation

Abstract

The application discloses a kind of electromagnetic spectrum energy collection and optimization method for low-power multi-sensor cooperation, to improve the energy efficiency and network stability of Internet of Things equipment.The method is deployed by multiple self-powered sensor nodes in the target area, collects multi-band radio frequency energy, captures electromagnetic spectrum energy in the environment and converts it into electrical energy storage, realizes self-sufficient energy.The three-dimensional electromagnetic energy density distribution map is constructed based on the half variance clustering algorithm, accurately identifies high-energy areas, optimizes node deployment, avoids frequent failure caused by low-energy areas, and reduces the blindness of network deployment.The energy-spectrum joint optimization model is established, and the reinforcement learning algorithm is combined to dynamically adjust the node working mode, balance energy collection, data transmission and sleep state, and improve resource utilization.Adaptive sleep scheduling strategy is designed, and according to the node energy state and task priority, the sleep cycle and wake-up timing are dynamically adjusted to avoid network paralysis caused by single-point energy depletion.Through the cooperative wake-up mechanism, the energy sharing and data transmission of low-energy nodes are realized, and the robustness and adaptability of the network are enhanced.The application is suitable for complex and variable Internet of Things environment, can effectively prolong the network life, improve the reliability of data transmission, and provide strong support for the sustainable development of low-power Internet of Things.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication and Internet of Things technology, specifically relating to a method and system for electromagnetic spectrum energy harvesting and optimization for low-power multi-sensor collaboration. Background Technology

[0002] With the rapid development of modern military informatization, low-power IoT equipment plays a crucial role in military fields such as battlefield situational awareness, equipment status monitoring, and unmanned swarm collaborative operations. However, traditional military IoT nodes are mostly powered by special batteries, which have three major drawbacks: First, battery performance degrades rapidly in extreme battlefield environments such as high altitude, high temperature, humidity, and strong vibrations, potentially causing critical nodes to fail and directly affecting the execution of combat missions; second, replacing batteries for nodes deployed behind enemy lines poses a risk of tactical exposure, potentially triggering enemy radio reconnaissance and positioning; and third, existing equipment lacks energy resilience mechanisms, and the depletion of energy in some nodes will lead to monitoring blind spots in the battlefield sensor network. To address the aforementioned issues, self-powered technology based on electromagnetic spectrum energy has become a new direction for the development of military IoT. However, it faces unique challenges in battlefield environments: First, in complex electromagnetic warfare environments, high-intensity interference signals coexist with low signal-to-noise ratio target frequency bands, requiring solutions to the anti-interference problem of spectrum energy acquisition; second, the distribution of radio frequency energy in dynamic battlefields exhibits strong spatiotemporal heterogeneity, such as near-field energy field distortion caused by UAV swarm maneuvers; third, military equipment needs to support multi-modal adaptive switching of "silent reconnaissance-sudden transmission-emergency wake-up," and traditional fixed-threshold energy management strategies are insufficient to meet tactical response requirements; finally, tactical-level energy coordination needs to address the issue of secure energy sharing across classified nodes to prevent electromagnetic feature exposure during energy transmission. Therefore, this invention proposes an electromagnetic spectrum energy sabotage-resistant acquisition and tactical optimization method for military IoT. By constructing an anti-interference multi-band acquisition array, dynamic battlefield energy field modeling, task-driven energy scheduling strategies, and a covert collaborative power supply mechanism, it achieves full-domain energy security and sabotage-resistant continuous operation of military IoT equipment in complex combat environments. Summary of the Invention

[0003] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0004] The purpose of this invention is to propose a method and system for electromagnetic spectrum energy harvesting and optimization for low-power multi-sensor collaboration. Addressing the numerous challenges faced by low-power IoT devices in energy harvesting and management, this invention proposes a comprehensive solution based on multi-band energy harvesting, node layout optimization, energy management strategies, and a collaborative wake-up mechanism. This solution aims to achieve efficient energy utilization and long-life operation of IoT devices, while reducing system energy consumption and hardware deployment costs.

[0005] Specifically, this disclosure presents a method for electromagnetic spectrum energy harvesting and optimization for low-power multi-sensor collaboration, characterized by the following steps: Multiple self-powered sensor nodes are deployed in the target area. The self-powered sensor nodes collect multi-band radio frequency energy, capture electromagnetic spectrum energy in the environment, and convert it into electrical energy for storage. A three-dimensional electromagnetic energy density distribution map is constructed based on the semi-variance clustering algorithm, and the deployment of self-powered sensor nodes is optimized based on the energy density distribution map. An energy-spectrum joint optimization model is established, and the working mode of the self-powered sensor node is dynamically adjusted by combining reinforcement learning algorithms. An adaptive sleep scheduling strategy is designed to dynamically adjust the sleep cycle and wake-up sequence of the self-powered sensor node based on its energy status and task priority. Energy sharing and data transmission between low-energy self-powered sensor nodes are achieved through a collaborative wake-up mechanism.

[0006] Furthermore, the self-powered sensor node described in this disclosure includes a multi-band microstrip patch antenna array, a π-type impedance matching network, and a multi-stage Dickson charge pump rectifier circuit; The self-powered sensor node collects data in the 2.4GHz, 5GHz, and 900MHz frequency bands. The conversion efficiency formula is as follows:

[0007] in, The output DC power is the actual power supplied to the load after rectification. Input radio frequency power, i.e., the radio frequency energy power captured by the antenna from the environment; The voltage of the kth RF input signal represents the RF signal voltages of multiple frequency bands received by the antenna; N represents the number of frequency bands. For output voltage, For load resistance, This is the input impedance.

[0008] Furthermore, the self-powered sensor node described in this disclosure has a supercapacitor and a lithium battery, wherein the supercapacitor is used for transient energy buffering, and the lithium battery serves as a long-term energy storage unit. Charge and discharge logic control is implemented through a PMIC, and the charge and discharge strategy is as follows:

[0009] in, Indicates the charging current. Indicates the maximum charging current. This indicates the current voltage of the supercapacitor. This indicates the voltage threshold of the supercapacitor. This represents a coefficient that adjusts the charging current.

[0010] Furthermore, the method described in this disclosure for constructing a three-dimensional electromagnetic energy density distribution map based on a semi-variance clustering algorithm, and for optimizing the deployment of self-powered sensor nodes based on this energy density distribution map, specifically includes: A three-dimensional electromagnetic energy density distribution map was constructed based on a semi-variance clustering algorithm; The K-means++ algorithm is used to cluster energy hotspots in the three-dimensional electromagnetic energy density distribution map; The simulated annealing algorithm is used to find the global optimum.

[0011] Furthermore, the specific formula for constructing a three-dimensional electromagnetic energy density distribution map based on the semi-variance clustering algorithm described in this disclosure is as follows:

[0012] in, For position The radio frequency energy density at that location, where d is the spatial distance. Let be the number of sample point pairs at a distance d; The K-means++ algorithm is used to cluster energy hotspots, with the objective function being to minimize the within-cluster variance:

[0013] Where K is the number of clusters, For the i-th cluster, Let be the center energy value of the i-th cluster. Location point The energy density value.

[0014] Furthermore, the self-powered sensor node layout optimization described in this disclosure is achieved through the following formula:

[0015] Where S is the set of node deployment locations, For position energy density, Limited by; Maximum number of nodes allowed to be deployed; This is a location weighting factor used to prevent nodes from being deployed too densely.

[0016] The candidate locations currently being evaluated, Other selected locations in set S; For the set of selected locations, for and The Euclidean distance between them.

[0017] 7. The method according to claim 4, wherein the objective function of the energy-spectrum joint optimization model is:

[0018] in, The baseline lifespan; This refers to the actual lifespan of the network. As a weighting factor; The root mean square error of the spectrum coverage; The constraints on the objective function include energy balance constraints and spectral coverage constraints.

[0019] Furthermore, the state space of the reinforcement learning strategy disclosed herein is defined as follows:

[0020] in, The current energy level stored in the node; The availability of radio frequency energy in the current environment; The priority of the current task; Action space is defined as:

[0021] Among them, Active is for nodes to perform energy harvesting or data processing; Sleep is for nodes to enter a low-power sleep state; Transmit is for nodes to perform data communication. The reward function is defined as:

[0022] in, The current energy level stored in the node; Energy threshold; RMSE (root mean square error) for spectral map reconstruction; This indicates that the node energy is within a reasonable range. This indicates that the accuracy of the spectrum map has improved compared to before.

[0023] Furthermore, the adaptive sleep scheduling strategy described in this disclosure, which dynamically adjusts the sleep cycle and wake-up sequence of the self-powered sensor node based on its energy state and task priority, specifically includes: The sleep scheduling strategy optimizes the node state transition probability through a Markov decision process; A formula for calculating the adaptive sleep cycle is established, and the formula is as follows:

[0024] in, This refers to the duration of the dormancy cycle; This represents the energy storage capacity of the current node. To wake up energy consumption, This refers to the power consumption during sleep mode. The energy status of self-powered sensor nodes is classified and the sampling frequency is adjusted. The wake-up sequence is adjusted according to the task priority. Energy sharing and data transmission between low-energy self-powered sensor nodes are achieved through a collaborative wake-up mechanism.

[0025] Furthermore, the collaborative wake-up mechanism described in this disclosure is based on TDMA beacon frames. High-energy self-powered sensor nodes transmit energy to low-energy self-powered sensor nodes through near-field coupling, and its transmission efficiency model is as follows:

[0026] in, The coupling coefficient is... , This is the coil quality factor.

[0027] This disclosure presents a proposed method for low-power multi-sensor collaborative electromagnetic spectrum energy harvesting and optimization. Through the design of a multi-band energy harvesting module, it can efficiently capture electromagnetic energy in different frequency bands within the environment. By constructing a three-dimensional electromagnetic energy density distribution map and optimizing node layout, it achieves rational node deployment and improves energy harvesting efficiency. Through an energy-spectrum joint optimization model and reinforcement learning algorithm, it dynamically adjusts node operating modes to maximize network lifetime. Furthermore, through an adaptive sleep scheduling strategy and a collaborative wake-up mechanism, it further optimizes energy utilization efficiency and system reliability. This invention is applicable to low-power IoT devices, enabling efficient harvesting and utilization of electromagnetic spectrum energy in complex environments, extending device lifespan, reducing maintenance costs, and possessing broad application prospects.

[0028] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a flowchart illustrating a low-power, multi-sensor collaborative electromagnetic spectrum energy harvesting and optimization method provided in an embodiment of this disclosure.

[0031] Figure 2 The flowchart illustrates a method for constructing a three-dimensional electromagnetic energy density distribution map based on a semi-variance clustering algorithm and optimizing the deployment of self-powered sensor nodes, as provided in this embodiment of the disclosure.

[0032] Figure 3 A flowchart illustrating the method for establishing an energy-spectrum joint optimization model to maximize network lifetime provided in this embodiment of the disclosure.

[0033] Figure 4 The flowchart illustrates a method for designing an adaptive sleep scheduling strategy and dynamically adjusting the sleep cycle and wake-up timing, as provided in this embodiment of the disclosure.

[0034] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. Detailed Implementation

[0035] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0036] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0037] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0038] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0039] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0040] For detailed implementation procedures, please refer to the appendix. Figure 1 , attached Figure 1 A method for electromagnetic spectrum energy harvesting and optimization for low-power multi-sensor collaboration is disclosed, comprising the following steps: Step S100: Deploy self-powered sensor nodes, integrate multi-band radio frequency energy harvesting modules, capture electromagnetic spectrum energy in the environment and convert it into electrical energy for storage.

[0041] The RF energy harvesting module employs a multi-band microstrip patch antenna array, capable of simultaneously covering multiple frequency bands including 2.4GHz (Wi-Fi / Bluetooth), 5GHz (5G NR n78 band), and 900MHz (LoRa band), with an antenna gain ≥5dBi to ensure efficient capture of RF signals in the environment. To achieve efficient energy transfer between the antenna and subsequent circuitry, a π-type impedance matching network is used for impedance matching to minimize signal energy loss. Subsequently, a multi-stage Dickson charge pump rectifier circuit converts the AC RF signal received by the antenna into a DC voltage output. The circuit input impedance is optimized to 50Ω, and low-threshold Schottky diodes are selected to improve rectification efficiency. Furthermore, considering the instability of RF energy intensity in the environment, a dynamic voltage regulator (LDO or DC-DC converter) is designed to stabilize the output voltage at 3.3V±5%, ensuring stable power supply to the sensor nodes.

[0042] Specifically, the radio frequency energy harvesting module uses a multi-band microstrip patch antenna array. This type of antenna has the advantages of small size, light weight, easy integration, and low cost, making it suitable for use in IoT sensor nodes.

[0043] Design an antenna array that simultaneously covers the following frequency bands: 2.4GHz band: Wi-Fi (802.11b / g / n) and Bluetooth signals; 5GHz band: 5G NR n78 band and Wi-Fi (802.11ac / ax) signals; 900MHz band: LoRa (Long Range Low Energy Communication) signals.

[0044] Antenna gain ≥ 5dBi (the higher the gain, the stronger the signal capture capability), VSWR ≤ 2 (the smaller the VSWR, the less signal reflection and the higher the energy transmission efficiency), linear polarization or circular polarization, selected according to actual environmental requirements.

[0045] A π-type impedance matching network (Pi-network), consisting of two capacitors and one inductor, is employed. This network is simple in structure and easy to implement. It achieves impedance matching between the antenna and the subsequent rectifier circuit, ensuring efficient signal energy transmission and reducing reflection loss. RF simulation software (such as ADS and HFSS) is used for simulation optimization to determine the optimal capacitor and inductor parameters, achieving a standard 50Ω impedance matching.

[0046] A multi-stage Dickson charge pump rectifier circuit is employed. This circuit has a simple structure, is easy to integrate, and is suitable for low-power applications. Through a multi-stage capacitor and diode network, the AC RF signal received by the antenna is progressively multiplied and rectified, converting it into a DC voltage output.

[0047] By adjusting the circuit parameters to make the input impedance of the rectifier circuit close to 50Ω, matching the antenna impedance, the energy transmission efficiency is improved; by selecting diodes with low threshold voltage and fast recovery (such as Schottky diodes), the conduction loss is reduced and the rectification efficiency is improved.

[0048] Conversion efficiency calculation formula: Conversion efficiency formula:

[0049] in, The output DC power is the actual power supplied to the load after rectification. Input radio frequency power, i.e., the radio frequency energy power captured by the antenna from the environment; The voltage of the kth RF input signal represents the RF signal voltages of multiple frequency bands received by the antenna, and the unit is volts (V); N represents the number of frequency bands. For output voltage, For load resistance, The input impedance is [value missing]. Additionally, a dynamic voltage regulator is used to stabilize the output voltage at 3.3V ± 5%. Vout is the power conversion efficiency; RL is the output DC voltage; and Pin is the load resistance.

[0050] A low-dropout linear regulator (LDO) or a switching regulator (DC-DC converter) is selected to stabilize the output voltage at 3.3V ± 5%. Due to the unstable radio frequency energy intensity in the environment, the rectified output voltage may fluctuate significantly. The regulator can ensure the stability of the output voltage and meet the power supply requirements of the sensor node.

[0051] A collaborative energy storage scheme involving supercapacitors and lithium batteries is proposed. The supercapacitor, acting as a transient energy buffer, features high charge / discharge efficiency (≥95%) and fast response, enabling rapid energy storage and release, and adapting to fluctuations in radio frequency (RF) energy. The lithium battery, serving as a long-term energy storage unit, stores energy overflowing from the supercapacitor, ensuring stable operation of the node even during prolonged periods without RF energy input. To achieve intelligent charge / discharge management, a dedicated power management chip (PMIC) is used to monitor parameters such as voltage, current, and temperature of the supercapacitor and lithium battery in real time, and automatically control the charging and discharging process according to a preset strategy. Specifically, when the supercapacitor voltage is below a set threshold, it is charged rapidly with the maximum current; when the voltage is above the threshold, the charging current is reduced to avoid overcharging and extend the lifespan of the energy storage components.

[0052] For supercapacitor parameter selection, the capacitance range is 1F to 10F (determined based on node power consumption requirements), the rated voltage is 2.7V to 5.5V, and the equivalent series resistance (ESR) should be as low as possible (generally below tens of milliohms) to improve charge and discharge efficiency. Supercapacitors act as transient energy buffer units, rapidly storing and releasing energy to adapt to fluctuations in radio frequency energy in the environment. With a charge and discharge efficiency ≥95%, far exceeding that of lithium batteries, they are suitable for fast charge and discharge scenarios.

[0053] The lithium battery is selected from lithium-ion or lithium polymer batteries, depending on the node power consumption requirements (e.g., 500mAh to 2000mAh), with a rated voltage of 3.7V (typical value) and an output of 3.3V through a voltage regulator circuit.

[0054] The power management chip (PMIC) selected is TI's BQ series or Maxim's MAX series to achieve intelligent charge and discharge management of the supercapacitor and lithium battery. The PMIC is used to monitor parameters such as voltage, current, and temperature of the supercapacitor and lithium battery in real time; according to preset strategies, it automatically controls the charging and discharging process to ensure the safe and efficient operation of the energy storage system. The specific strategies are as follows: The charging and discharging strategy is as follows:

[0055] in, Indicates the charging current. Indicates the maximum charging current. This indicates the current voltage of the supercapacitor. This indicates the voltage threshold of the supercapacitor. When the voltage is below this value, the supercapacitor will be charged with the maximum current. This represents the coefficient for adjusting the charging current; it is set here. That is, when the voltage of the supercapacitor is higher than the threshold, the charging current is halved. When the voltage of the supercapacitor is lower than the threshold, it is charged quickly with the maximum current; when the voltage is higher than the threshold, the charging current is reduced to avoid overcharging.

[0056] The supercapacitor voltage is monitored in real time by the ADC (analog-to-digital converter) built into the PMIC chip; based on the above strategy, the charging current is automatically adjusted to achieve intelligent charging and discharging management.

[0057] Finally, during the system integration and testing phase, professional PCB design tools (such as Altium Designer or Cadence) were used for circuit board layout, with particular attention paid to the shielding and interference suppression of the RF circuits. Supercapacitors, lithium batteries, and PMIC chips were strategically placed to ensure good heat dissipation and system safety. After system completion, a vector network analyzer (VNA) was used to test antenna performance (such as gain and VSWR), an oscilloscope and multimeter were used to test the output voltage and current of the rectifier circuit, and a power analyzer was used to test the charging and discharging efficiency and stability of the energy storage system.

[0058] Step S200: Construct a three-dimensional electromagnetic energy density distribution map based on the semi-variance clustering algorithm, and optimize the deployment of self-powered sensor nodes based on the energy density distribution map. See [link to implementation details] for more information. Figure 2 . Step S2001: Construct a three-dimensional electromagnetic energy density distribution map based on the semi-variance clustering algorithm.

[0059] In a real-world environment, portable radio frequency (RF) energy measurement devices (such as spectrum analyzers and RF power meters) are used to perform gridded measurements of the target area. During measurement, the spatial coordinates (latitude, longitude, and altitude) of each measurement point and the corresponding RF energy density value (unit: μW / cm² or dBm) are recorded. The measurement frequency bands cover key bands such as 2.4GHz (Wi-Fi / Bluetooth), 5GHz (5G / Wi-Fi), and 900MHz (LoRa). The collected data is cleaned to remove outliers and noise. Data normalization is performed to unify the units of measurement for easier subsequent analysis.

[0060] Calculate the energy correlation between different locations in space using the semivariance function formula:

[0061] in, For position The radio frequency energy density at that location, where d is the spatial distance. Let be the number of sample point pairs at a distance d.

[0062] Plot the semivariance function to analyze the variation of spatial correlation with distance. Determine the effective range, sill, and nugget of spatial correlation to provide parameter basis for subsequent spatial interpolation.

[0063] Using the semivariance function analysis results, Kriging interpolation is employed to predict the radio frequency energy density at unmeasured locations. Kriging interpolation considers spatial correlation, providing an optimal unbiased estimate and specifying the estimation error. Essentially, Kriging interpolation is a weighted average: each known point is assigned a weight based on its observations, and the estimated value for unknown points is the sum of the products of the known point observations and their corresponding weights. Unlike simple interpolation methods such as inverse distance weighting (IDW), Kriging's weights are not solely determined by distance but are calculated based on the spatial structure of the data (e.g., correlation, variability). The ultimate goal is to minimize the error (variance) of the estimate while maintaining unbiasedness (the expected value of the estimate equals the true value).

[0064] The interpolation results are stored in the form of a 3D grid, with each grid point containing spatial coordinates and the corresponding radio frequency energy density value. A 3D electromagnetic energy density distribution map is then created using professional visualization software (such as MATLAB, Python's Mayavi, or Plotly) to visually represent the energy distribution.

[0065] Step S2002: Cluster energy hotspots on the three-dimensional electromagnetic energy density distribution map using the K-means++ algorithm.

[0066] The data of each grid point in the three-dimensional electromagnetic energy density distribution map is converted into the input format required by the clustering algorithm, that is, each sample point contains spatial coordinates (x, y, z) and energy density value E.

[0067] The K-means++ algorithm is an improved version of the K-means algorithm, capable of selecting initial cluster centers more rationally and avoiding getting trapped in local optima. The algorithm steps can be summarized as follows: (1) Initialize cluster centers: The first cluster center is randomly selected.

[0068] For each subsequent cluster center, select the sample point farthest from the previously selected cluster center as the new cluster center, until K cluster centers are selected.

[0069] (2) Sample point allocation: Calculate the distance from each sample point to each cluster center (usually using Euclidean distance), and assign the sample point to the nearest cluster center.

[0070] (3) Update cluster centers: Recalculate the centroid of each cluster (i.e., the mean of the sample points within each cluster).

[0071] (4) Iterative optimization: Repeat steps (2) and (3) until the cluster centers no longer change or the maximum number of iterations is reached.

[0072] The objective function used for cluster analysis of energy hotspot regions aims to group locations with similar energy densities into the same class, thereby identifying energy hotspot regions and minimizing the intra-class variance.

[0073] Where K is the number of clusters, For the i-th cluster, Let be the center energy value of the i-th cluster. The energy density value at location ν is the radio frequency energy intensity at that location.

[0074] Suppose we measured the energy density at 100 locations in a region, and now we want to divide these locations into 3 clusters (K=3) to identify 3 energy hotspot regions. Each cluster, C1, C2, and C3, contains some locations; each location ν has an energy value E(ν); each cluster has a central energy value μi, for example, μ1=10, μ2=20, μ3=30. The formula calculates the sum of the squares of the differences between the energy value E(ν) of each location and the central energy value μi of its cluster; the smaller the sum, the better. In other words, the closer the energy values ​​of the locations within each cluster are, the better the clustering effect and the more accurate the energy hotspot region identification.

[0075] Analyze the spatial distribution, mean energy density, variance, and other statistical characteristics of each cluster. Identify regions with high energy density and concentrated spatial distribution as energy hotspots. Display different clusters on a 3D map using different colors or markers to visually represent the spatial location and extent of energy hotspots.

[0076] The goal is to maximize the energy density of node deployment locations while considering the sparsity of node deployment to avoid excessive node concentration.

[0077] Where S is the set of node deployment locations, For position energy density, It is an abbreviation of "subject to," meaning "subject to." The maximum number of nodes allowed to be deployed (e.g., budget or resource limitations); This is a location weighting factor used to prevent nodes from being deployed too densely.

[0078] The candidate locations currently being evaluated, Other selected locations in set S (i.e., locations where the deployment nodes have been decided); This refers to the set of selected locations (i.e., the set of locations where nodes have already been decided to be deployed). Location and The Euclidean distance between two points (i.e., the straight-line distance between two points).

[0079] Node deployment locations must be within feasible spatial limits (such as inside buildings, along roads, etc.). The number of nodes may be limited by factors such as cost and network scale.

[0080] Step S2003: Use the simulated annealing algorithm to solve for the global optimal solution.

[0081] Since node layout optimization is a combinatorial optimization problem, it may have multiple local optima. Simulated annealing, with its global search capability, effectively avoids getting trapped in local optima. The specific algorithm steps are as follows: (1) Initial solution generation: Randomly generate an initial node layout scheme as the current solution.

[0082] (2) Neighborhood solution generation: make a small perturbation to the current solution (such as randomly moving the position of a node) to generate a neighborhood solution.

[0083] (3) Calculate the objective function value of the neighborhood solution (i.e., the sum of the products of energy density and weighting factor).

[0084] (4) Criteria for accepting solutions: If a neighboring solution is better than the current solution, then the neighboring solution is accepted as the new current solution.

[0085] If the neighboring solution is worse than the current solution, the worse solution is accepted with a certain probability (the probability is related to the temperature parameter).

[0086] (5) Temperature update: reduce the temperature according to the temperature decay coefficient α=0.95.

[0087] (6) The iteration ends, and steps (2) to (5) are repeated until the maximum number of iterations (≥1000 times) is reached or the temperature drops to the preset threshold.

[0088] The simulated annealing algorithm parameters are set as follows: temperature decay coefficient: α=0.95, maximum number of iterations ≥1000.

[0089] Analyze the optimized node layout scheme and calculate the total energy density at each node deployment location. Check the sparsity of the node deployment to ensure a reasonable node distribution. In a real-world environment, deploy nodes according to the optimized scheme and monitor the actual energy harvesting of the nodes. Compare the optimized scheme with the original random deployment scheme to verify the optimization effect.

[0090] Construct a three-dimensional electromagnetic energy density distribution map to optimize the deployment location of sensor nodes and ensure that nodes are preferentially deployed in energy-rich areas, thereby improving the energy utilization efficiency and lifespan of the entire Internet of Things network.

[0091] By constructing a three-dimensional electromagnetic energy density distribution map, we can identify which areas of the environment are energy-rich and which are energy-scarce. Based on this energy distribution map, we can rationally plan node locations, prioritizing node deployment in energy-rich areas to prevent premature node failure in energy-scarce regions. Optimizing node layout improves the overall energy utilization efficiency of the network and extends its lifespan.

[0092] Step S300: Establish an energy-spectrum joint optimization model, and dynamically adjust the working mode of self-powered sensor nodes using reinforcement learning algorithms to maximize network lifetime. See Appendix for specific implementation details. Figure 3 .

[0093] Step S3001: Establish an energy-spectrum joint optimization model.

[0094] In IoT environments, sensor nodes typically rely on radio frequency (RF) energy collected from the environment to maintain their operation. However, the distribution of RF energy in the environment is often uneven and dynamically changes over time, making the switching between energy collection, data transmission, and sleep states complex and difficult to predict. Therefore, establishing an optimization model that comprehensively considers energy utilization efficiency and spectrum coverage quality is particularly important.

[0095] The core idea of ​​the energy-spectrum joint optimization model is to extend the overall lifespan of the network as much as possible while maintaining high spectrum map reconstruction accuracy, under the premise of ensuring node energy balance. Specifically, the model defines a clear objective function:

[0096] in, The lifespan is based on hours; The actual lifespan (or lifetime) of the network. As a trade-off factor, it is used to balance network lifetime and spectrum coverage error; The root mean square error (RMSE) of the spectrum coverage is used to measure the accuracy of the spectrum coverage. Constraints include energy balance constraints. and spectrum coverage constraints: .in, The energy stored at time t. The energy consumed at time t. This is the energy margin at time t, used to ensure that the system can still operate normally when energy is insufficient.

[0097] This constraint ensures that the storage energy of a node is sufficient to meet its current consumption and certain marginal energy requirements at any given time, preventing the node from failing due to insufficient energy. This constraint ensures that the reconstruction error of the spectrum map does not exceed 2dB, guaranteeing the map's accuracy and reliability. By establishing the aforementioned energy-spectrum joint optimization model, the optimization objective can be clearly defined, providing a clear optimization direction and decision-making basis for subsequent reinforcement learning algorithms.

[0098] Step S3002: Design reinforcement learning strategies.

[0099] To achieve intelligent decision-making at the nodes, step S300 introduces a reinforcement learning strategy. The core idea of ​​reinforcement learning is to allow nodes to autonomously learn the optimal decision-making strategy through trial and error and feedback mechanisms during their interaction with the environment.

[0100] In designing reinforcement learning strategies, it is first necessary to clearly define three key elements: state space, action space, and reward function.

[0101] The state space defines the environmental information perceived by a node at a given moment. In this scheme, the state space is defined as follows:

[0102] in, The current energy level stored in the node; The availability of radio frequency energy in the current environment; This sets the priority for the current task. Through the above state space definition, nodes can perceive their own energy status, environmental energy availability, and task requirements in real time, providing a basis for subsequent decision-making.

[0103] The action space defines the actions a node can take in a given state. In this scheme, the action space is defined as follows:

[0104] In this context, Active is for nodes to perform energy harvesting or data processing; Sleep is for nodes to enter a low-power sleep state; and Transmit is for nodes to perform data communication. Based on its current state, a node selects an action from its action space to execute, achieving a balance between energy utilization and task requirements.

[0105] The reward function defines the feedback signal a node receives after performing an action, used to guide the node in learning the optimal policy. In this scheme, the reward function is defined as:

[0106] in, The current energy level stored in a node reflects whether the node's energy state is healthy; The energy threshold is a preset critical value used to determine whether a node is in a low-energy state; RMSE is the root mean square error of the spectrum map reconstruction, which measures the accuracy of spectrum coverage, and the smaller the better. indicates that the node energy is within a reasonable range (e.g., neither too high nor too low), that is, the energy state is stable, indicates that the accuracy of the spectrum map is improved compared to before (i.e., the RMSE is reduced).

[0107] When the node energy state is good (balanced) and the spectrum map accuracy is improved (RMSE improved): it indicates that the node has taken a "good" action (such as harvesting energy in due time and transmitting data reasonably); a positive reward of +1 is given to encourage this behavior.

[0108] When the node energy is lower than the safety threshold (Estore<Ethreshold): it indicates that the node may have taken a "poor" action (such as excessive transmission and failing to sleep in time); a negative reward of -0.5 is given to punish this behavior, prompting the node to learn to avoid low energy states.

[0109] Through the above reward function design, the node can gradually learn the optimal strategy to strike a balance between energy utilization and spectrum coverage.

[0110] Step S3003, designing a reinforcement learning (DQN) algorithm, so that the node can independently decide the optimal working mode according to the real-time energy state, radio frequency energy availability and task priority.

[0111] To effectively implement the reinforcement learning strategy, the DQN (Deep Q-Network) algorithm in deep reinforcement learning is adopted. The DQN algorithm combines a deep neural network and the Q-learning method, which can process high-dimensional state spaces and realize intelligent decision-making in complex environments.

[0112] The core idea of the DQN algorithm is to use a deep neural network to approximate the Q function (action value function), that is, Q(s,a;θ)≈Q (s,a).

[0113] Wherein, Q(s,a;θ) represents the value of taking action a in state s, and θ is the parameter of the neural network. By continuously training the neural network, the Q value output by it is gradually made to approach the optimal action value function Q (s,a), so as to realize optimal decision-making.

[0114] In this scheme, the DQN algorithm employs a simple three-layer fully connected neural network. The input layer represents the node's current state *st* (energy storage, RF energy availability, task priority); the hidden layers consist of two fully connected layers with 64 and 32 neurons respectively; and the output layer represents the Q-values ​​of the action space *at* (Q-values ​​of the three actions: Active, Sleep, and Transmit). This network structure is simple and efficient, making it suitable for implementation on resource-constrained sensor nodes.

[0115] The training process of the DQN algorithm includes the following key steps: (1) Experience Replay: Store the historical experience (st, at, rt, st+1) of the node's interaction with the environment in the experience pool; during each training session, randomly select a batch of samples from the experience pool for training to break the data correlation and improve training stability.

[0116] (2) Define the loss function as the mean square error between the current Q value and the target Q value.

[0117] (3) Network parameter update: use gradient descent algorithm to minimize the loss function and update the network parameter θ; every certain number of steps, copy the current network parameter to the target network to stabilize the training process.

[0118] (4) Exploration vs. Exploitation, using Strategy, with a certain probability Randomly select an action (exploration) with a probability of 1. Select the current optimal action (utilize). Gradually decrease this as training progresses. Value, reduce exploration, increase utilization.

[0119] Through the training and implementation of the DQN algorithm, nodes can gradually learn the optimal action strategy under different states, achieving a dynamic balance between energy utilization and task requirements, and ultimately extending network lifetime and improving spectrum coverage quality.

[0120] Step S300 establishes an energy-spectrum joint optimization model and combines it with reinforcement learning algorithms to dynamically adjust the working modes (acquisition, transmission, and hibernation) of nodes. This ensures that nodes hibernate promptly when energy is insufficient and operate efficiently when energy is sufficient, thereby maximizing the overall network lifetime. The objective function of step S300 not only focuses on network lifetime (Tnet) but also introduces spectrum coverage error (RMSEmap) as a penalty term to ensure that while extending network lifetime, high spectrum map reconstruction accuracy is maintained. The importance of both is adjusted by a tradeoff factor λ (=0.1) to achieve the optimal balance between energy efficiency and spectrum coverage quality. Traditional static strategies cannot adapt to complex dynamic environments. The reinforcement learning (DQN) algorithm is introduced, enabling nodes to autonomously decide on the optimal working mode based on real-time energy status, RF energy availability, and task priorities, thereby improving the network's intelligence and adaptability.

[0121] Step S400: Design an adaptive sleep scheduling strategy to dynamically adjust the sleep cycle and wake-up sequence based on the energy status and task priority of the self-powered sensor nodes. See Appendix for specific implementation details. Figure 4 .

[0122] Step S4001: Use Markov Decision Process (MDP) to model and optimize the state transition probabilities of nodes.

[0123] Markov Decision Processes (MDPs) are a mathematical framework for modeling sequential decision-making problems under uncertain environments, widely used in reinforcement learning, operations research, and cybernetics. At its core, MDPs describe how an agent (or decision-maker) maximizes long-term cumulative rewards by choosing actions in a stochastic environment through a "state-action-reward-transition" logic. The foundation of MDPs is the Markov property: future states depend only on the current state and the currently chosen action, and are independent of past states.

[0124] In conjunction with this disclosure, an MDP can be composed of quintuples. S,A,T,R, The definitions and meanings of each element are as follows: State Space: The node's current energy state (high, medium, low energy); current task priority (high, medium, low priority tasks); and the node's current working state (active, sleeping, transmitting).

[0125] Action Space: The actions that a node can perform include: remain active, enter sleep mode, and wake up to transmit data.

[0126] Transition Probability: Defines the probability that a node will transition from one state to another, taking into account the uncertainties of energy consumption, task requirements, and energy harvesting.

[0127] Reward Function: Based on the node's energy state, task completion status, and spectrum map accuracy, a reward function is designed to guide the node to learn the optimal state transition strategy.

[0128] Through MDP modeling, nodes can select the optimal action based on their current state and expected benefits, achieving a balance between energy utilization and task requirements.

[0129] Step S4002: Establish the calculation of adaptive sleep cycle.

[0130] Construct an adaptive sleep cycle calculation formula to dynamically adjust the sleep duration based on the node's current energy state:

[0131] in, This refers to the sleep cycle duration, which is the length of time a node remains in a sleep state. This represents the energy storage capacity of the current node. To wake up energy consumption, This refers to sleep power consumption. The energy state classification is as follows: In... That is, when the energy is high, high-frequency sampling (10Hz) is performed. The energy currently stored in the node; At medium energy, the sampling frequency is 5Hz; When energy is low, it enters deep sleep (only keeping the clock awake).

[0132] When a node has sufficient energy, the sleep cycle is shorter and the node is woken up more frequently; when a node has insufficient energy (lower Estore), the sleep cycle is extended to reduce energy consumption; the sleep cycle is smoothly adjusted through a nonlinear function (logarithmic function) to avoid drastic fluctuations in the sleep cycle caused by small changes in energy state.

[0133] Step S4003: Energy state classification and sampling frequency adjustment.

[0134] The table below categorizes node energy states into three levels, and the sampling frequency is adjusted according to each level:

[0135] Through this hierarchical management, nodes can flexibly adjust their workload according to their energy status, thus preventing rapid energy depletion.

[0136] Step S4004: Adjust the wake-up sequence according to task priority.

[0137] Each node also takes into account the differences in task priorities. For example, high-priority tasks (such as emergency alarms and critical data transmission) will be immediately woken up and executed even if the node is in a dormant state, while ensuring the real-time performance and reliability of critical tasks.

[0138] For medium-priority tasks (such as regular data uploads), nodes are woken up and executed as scheduled, provided that energy allows; balancing energy consumption and task requirements.

[0139] Low-priority tasks (such as periodic status reports) are executed by nodes when energy is sufficient, and delayed or canceled when energy is insufficient; this is to avoid low-priority tasks affecting the energy security of nodes.

[0140] By adjusting the wake-up timing based on task priority, we can ensure that critical tasks are executed first and improve the quality of network services.

[0141] Step S400 involves adjusting the wake-up sequence dynamically based on task priority, according to the varying real-time and energy consumption requirements of different tasks. This ensures that high-priority tasks are executed first, improving network service quality. Through hierarchical energy state management (high, medium, and low energy states), fine-grained energy control at the node level is achieved, enhancing the system's adaptability to environmental changes.

[0142] Step S500: Through the collaborative wake-up mechanism, energy sharing and data retransmission of low-energy nodes are realized, and high-precision reconstruction of the spectrum map is completed.

[0143] Time Division Multiple Access (TDMA) beacon frames are used to achieve synchronous wake-up between nodes. Each node sends a beacon frame within a preset time slot, broadcasting its own energy status and task requirements. The beacon frame contains information such as node ID, energy level, and data priority.

[0144] Wake-up Time Window =10ms, during which the node remains awake, listening for beacon frames from other nodes. Time synchronization avoids communication conflicts between nodes, ensuring the efficiency of coordinated wake-up.

[0145] Low-energy nodes declare their insufficient energy in the beacon frame and request energy sharing or data retransmission. Upon receiving the request, high-energy nodes respond in the next time slot, establishing an energy transfer or data communication link. Energy or data transfer between nodes is completed via near-field coupling (NFC) or wireless communication.

[0146] Near-field coupling (NFC) technology is used to achieve energy transfer from a high-energy node to a low-energy node. Utilizing the principle of electromagnetic induction, energy transfer is achieved through magnetic field coupling between coils. The high-energy node acts as the energy transmitter, and the low-energy node acts as the energy receiver.

[0147] Energy transfer efficiency model:

[0148] in, , The quality factor of the transmitting and receiving coils reflects the energy storage and transmission capabilities of the coils. Energy transmission efficiency represents the proportion of energy effectively received by the receiver from the transmitting end. The coupling coefficient reflects the degree of coupling between two coils (0 ≤ k ≤ 1). By optimizing coil design (such as coil size, number of turns, and materials), the quality factors Q1 and Q2 can be improved; by adjusting the coil spacing and angle, the coupling coefficient k can be increased, thereby improving energy transfer efficiency.

[0149] This disclosure not only enables energy sharing but also supports data retransmission, ensuring the integrity of the spectrum map. When a low-energy node is unable to upload data due to insufficient energy, it caches the data in its local storage. When a high-energy node approaches, the low-energy node requests data retransmission through a cooperative wake-up mechanism. After receiving the data, the high-energy node uploads it to the base station or aggregation node on its behalf, completing the data retransmission.

[0150] Base stations or aggregation nodes collect data from all nodes (including retransmitted data) to reconstruct the spectrum map; due to improved data integrity, the accuracy of the spectrum map (such as RMSE index) is significantly improved.

[0151] The collaborative wake-up mechanism disclosed herein also possesses intelligent scheduling capabilities. Nodes dynamically adjust the frequency of wake-up request transmission based on task priority and energy status. High-priority tasks (such as emergency alarms) receive priority access to energy sharing and data retransmission.

[0152] In a network, multiple high-energy nodes can work together to provide energy to multiple low-energy nodes, preventing excessive energy consumption by a single node. By optimizing the scheduling algorithm, a balanced distribution of energy can be achieved, extending the overall lifespan of the network.

[0153] Step S500 enables energy sharing and balancing, allowing high-energy nodes to transfer energy to low-energy nodes to prevent node failure; it also improves data retransmission and map accuracy, enabling low-energy nodes to retransmit their data and significantly improving the accuracy of spectrum map reconstruction; it enhances network robustness, allowing nodes to work collaboratively and greatly improving the network's adaptability to dynamic environments; and it enhances intelligence, enabling the network to autonomously repair and optimize itself through TDMA beacon frames and intelligent scheduling.

[0154] To verify that the method proposed in this embodiment is significantly superior to traditional methods in terms of energy capture, node deployment, network lifetime, and stealth, the following experiment was conducted.

[0155] In this example, the test environment simulates a high-intensity combat battlefield environment (mixed mountainous and urban terrain, area 2 km²), deploying 30 survivable, self-powered tactical nodes. The energy distribution characteristics of the target frequency bands are as follows: the 2.4 GHz (Wi-Fi / Bluetooth) band has an average power density of 0.1 μW / cm², with short-range communication jamming signals present; the 5 GHz (5G NR n78 band) band has an average power density of 0.25 μW / cm², accompanied by dynamic spectrum interference (such as frequency-hopping radar signals); and the 900 MHz (tactical LoRa band) band has an average power density of 0.05 μW / cm², requiring countermeasures against low-probability-of-interception reconnaissance signals. These frequency bands cover tactical communication, data backhaul, and covert reconnaissance needs, and their energy harvesting must be compatible with high-intensity jamming suppression and concealment requirements. A traditional military single-band (900 MHz) energy harvesting system with a fixed node layout is then used as a comparison scheme. The evaluation period is 15 days of continuous high-intensity combat exercises, and tactical performance indicators are recorded. The results are shown in the following two tables. Table 1.1 is a comparison table of multi-band energy harvesting efficiency and anti-interference performance, and Table 1.2 is a table of tactical node deployment optimization and network survivability assessment. Table 1.1 Comparison of Multi-Band Energy Harvesting Efficiency and Anti-interference Performance

[0156] Table 1.2 Tactical Node Deployment Optimization and Network Survivability Assessment Table

[0157] Tables 1.1 and 1.2 reveal the core shortcomings of traditional military IoT solutions in multi-band coordination and survivability: uneven utilization of frequency band resources (2.4 GHz has strong anti-interference capabilities but low power density, 5 GHz has high energy but insufficient interference suppression, and 900 MHz has weak overall performance), resulting in limited energy harvesting efficiency (44%-52%); rigid deployment strategies lead to only 55% coverage in high-threat areas and a 50-second dynamic reconfiguration time, making it difficult to cope with sudden situations; and poor network survivability (survival period of only 6 days under electronic interference). This invention, through multi-band adaptive switching, semi-variance clustering optimization deployment, and reinforcement learning dynamic scheduling, increases coverage to 90%, reduces reconfiguration time to 10 seconds, and extends survival to 14 days, systematically solving the problems of low energy efficiency, slow response, and insufficient survivability, significantly enhancing the tactical adaptability and covert combat capabilities of battlefield IoT.

[0158] This disclosure constructs a three-dimensional electromagnetic energy density distribution map, which can accurately identify high-energy areas, guide node optimization, avoid frequent failures caused by low-energy areas, and reduce the randomness of network deployment. This solution proposes an energy-spectrum joint optimization model, which uses reinforcement learning to dynamically balance energy harvesting, data transmission, and sleep states, improving resource utilization. Adaptive sleep scheduling and collaborative wake-up mechanisms can dynamically adjust behavior based on the real-time energy status and task requirements of nodes, preventing network paralysis caused by single-point energy depletion and adapting to complex and changing environments.

[0159] A computer device according to embodiments of the present disclosure includes a memory and a processor. The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0160] The processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of this disclosure, the processor is used to execute computer-readable instructions stored in the memory, causing the computer device to perform all or part of the steps of the learning outcome prediction method based on learning behavior data mining of the foregoing embodiments of this disclosure.

[0161] Those skilled in the art will understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this disclosure.

[0162] like Figure 5 This is a schematic diagram of a computer device provided for an embodiment of the present disclosure. It illustrates a structural schematic diagram suitable for implementing the computer device in the embodiments of the present disclosure. Figure 5 The computer device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0163] like Figure 5As shown, a computer device may include a processor (such as a central processing unit, graphics processing unit, etc.), which performs various appropriate actions and processes based on programs stored in read-only memory (ROM) or programs loaded from storage devices into random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer device. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0164] Typically, the following devices can be connected to the I / O interface: input devices, such as sensors or visual information acquisition devices; output devices, such as displays; storage devices, such as magnetic tapes or hard drives; and communication devices. Communication devices allow the computer device to communicate wirelessly or wiredly with other devices (such as edge computing devices) to exchange data. Although a computer device with various devices is illustrated, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0165] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the learning outcome prediction method based on learning behavior data mining according to embodiments of this disclosure are performed.

[0166] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0167] A computer-readable storage medium according to embodiments of the present disclosure stores non-transitory computer-readable instructions. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the learning outcome prediction method based on learning behavior data mining according to the foregoing embodiments of the present disclosure are performed.

[0168] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0169] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0170] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0171] In this disclosure, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The block diagrams of devices, apparatuses, devices, and systems involved in this disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as "comprising," "including," "having," etc., are open-ended terms meaning "including but not limited to," and are used interchangeably with them. The terms "or" and "and" as used herein refer to the terms "and / or," and are used interchangeably with them unless the context clearly indicates otherwise. The term "such as" as used herein refers to the phrase "such as but not limited to," and is used interchangeably with it.

[0172] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0173] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0174] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0175] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0176] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for electromagnetic spectrum energy harvesting and optimization for low-power multi-sensor collaboration, characterized in that, Includes the following steps: Multiple self-powered sensor nodes are deployed in the target area. The self-powered sensor nodes collect multi-band radio frequency energy, capture electromagnetic spectrum energy in the environment, and convert it into electrical energy for storage. A three-dimensional electromagnetic energy density distribution map is constructed based on the semi-variance clustering algorithm, and the deployment of self-powered sensor nodes is optimized based on the energy density distribution map. An energy-spectrum joint optimization model is established, and the working mode of the self-powered sensor node is dynamically adjusted by combining reinforcement learning algorithms. An adaptive sleep scheduling strategy is designed to dynamically adjust the sleep cycle and wake-up sequence of the self-powered sensor node based on its energy status and task priority. Energy sharing and data transmission between low-energy self-powered sensor nodes are achieved through a collaborative wake-up mechanism.

2. The method according to claim 1, characterized in that, The self-powered sensor node includes a multi-band microstrip patch antenna array, a π-type impedance matching network, and a multi-stage Dickson charge pump rectifier circuit. The self-powered sensor node collects data in the 2.4GHz, 5GHz, and 900MHz frequency bands. The conversion efficiency formula is as follows: in, The output DC power is the actual power supplied to the load after rectification. Input radio frequency power, i.e., the radio frequency energy power captured by the antenna from the environment; The voltage of the kth RF input signal represents the RF signal voltages of multiple frequency bands received by the antenna; N represents the number of frequency bands. For output voltage, For load resistance, This is the input impedance.

3. The method according to claim 2, characterized in that, The self-powered sensor node has a supercapacitor and a lithium battery. The supercapacitor is used for transient energy buffering, and the lithium battery serves as a long-term energy storage unit. Charge and discharge logic control is implemented via a PMIC, and the charge and discharge strategy is as follows: in, Indicates the charging current. Indicates the maximum charging current. This indicates the current voltage of the supercapacitor. This indicates the voltage threshold of the supercapacitor. This represents a coefficient that adjusts the charging current.

4. The method according to claim 1, characterized in that, The construction of a three-dimensional electromagnetic energy density distribution map based on the semi-variance clustering algorithm, and the optimization of the deployment of self-powered sensor nodes based on the energy density distribution map, specifically include: A three-dimensional electromagnetic energy density distribution map was constructed based on a semi-variance clustering algorithm; The K-means++ algorithm is used to cluster energy hotspots in the three-dimensional electromagnetic energy density distribution map; The simulated annealing algorithm is used to find the global optimum.

5. The method according to claim 4, characterized in that, The specific formula for constructing a three-dimensional electromagnetic energy density distribution map based on the semi-variance clustering algorithm is as follows: in, For position The radio frequency energy density at that location, where d is the spatial distance. Let be the number of sample point pairs at a distance d; The K-means++ algorithm is used to cluster energy hotspots, with the objective function being to minimize the within-cluster variance: Where K is the number of clusters, For the i-th cluster, Let be the center energy value of the i-th cluster. Location point The energy density value.

6. The method according to claim 5, characterized in that, The self-powered sensor node layout optimization is achieved through the following formula: Where S is the set of node deployment locations, For position energy density, Limited by; Maximum number of nodes allowed to be deployed; This is a location weighting factor used to prevent nodes from being deployed too densely. The candidate locations currently being evaluated, Other selected locations in set S; For the set of selected locations, for and The Euclidean distance between them.

7. The method according to claim 4, characterized in that, The objective function of the energy-spectrum joint optimization model is: in, The baseline lifespan; This refers to the actual lifespan of the network. As a weighting factor; The root mean square error of the spectrum coverage; The constraints on the objective function include energy balance constraints and spectral coverage constraints.

8. The method according to claim 7, characterized in that, The state space of a reinforcement learning policy is defined as follows: in, The current energy level stored in the node; The availability of radio frequency energy in the current environment; The priority of the current task; Action space is defined as: Among them, Active is for nodes to perform energy harvesting or data processing; Sleep is for nodes to enter a low-power sleep state; Transmit is for nodes to perform data communication. The reward function is defined as: in, The current energy level stored in the node; Energy threshold; RMSE (root mean square error) for spectral map reconstruction; This indicates that the node energy is within a reasonable range. This indicates that the accuracy of the spectrum map has improved compared to before.

9. The method according to claim 8, characterized in that, The adaptive sleep scheduling strategy designed dynamically adjusts the sleep cycle and wake-up sequence of the self-powered sensor node based on its energy status and task priority. Specifically, this includes: The sleep scheduling strategy optimizes the node state transition probability through a Markov decision process; A formula for calculating the adaptive sleep cycle is established, and the formula is as follows: in, This refers to the duration of the dormancy cycle; This represents the energy storage capacity of the current node. To wake up energy consumption, This refers to the power consumption during sleep mode. The energy status of self-powered sensor nodes is classified and the sampling frequency is adjusted. The wake-up sequence is adjusted according to the task priority. Energy sharing and data transmission between low-energy self-powered sensor nodes are achieved through a collaborative wake-up mechanism.

10. The method according to claim 9, characterized in that, The cooperative wake-up mechanism is based on TDMA beacon frames. High-energy self-powered sensor nodes transmit energy to low-energy self-powered sensor nodes through near-field coupling. The transmission efficiency model is as follows: in, The coupling coefficient is... , This is the coil quality factor.

Images