Energy-saving routing method and system for wireless sensor network based on intelligent algorithm

Abstract

Disclosed is a wireless sensor network energy-saving routing method and system based on an intelligent algorithm. The method collects the residual energy values of each wireless sensor in the wireless sensor network, introduces a data processing and analysis algorithm at the back end to perform time sequence analysis on the residual energy values of each wireless sensor, and simultaneously performs residual energy time sequence feature mapping correction of each wireless sensor in combination with the network space topology of the wireless sensor network and the distance information between the nodes and the base station, so as to realize prediction of the probability of a node becoming a cluster head and selection of the optimal cluster head node. In this way, the energy utilization efficiency can be improved, and the network life cycle can be prolonged.

Description

Technical Field

[0001] This application relates to the field of wireless sensor networks, and more specifically, to an energy-saving routing method and system for wireless sensor networks based on intelligent algorithms. Background Technology

[0002] Wireless sensor networks (WSNs) are self-organizing networks composed of a large number of sensor nodes distributed across a sensing area. They can monitor, collect, process, and transmit data about targets or the environment within that area. WSNs have broad application prospects, such as military reconnaissance, environmental monitoring, intelligent transportation, and smart homes. However, since nodes in WSNs are typically powered by batteries or micro-energy harvesting technologies, their energy is limited and difficult to replace or recharge. Therefore, how to effectively utilize limited energy resources to achieve energy conservation and extend the network's lifespan is an important issue in WSN research.

[0003] Clustering routing protocols are a commonly used energy-saving routing method. Clustering-based routing protocols divide the network into several clusters, and each cluster selects a cluster head node to be responsible for collecting and forwarding data, thereby reducing the communication overhead between nodes.

[0004] The low-energy adaptive clustering hierarchy (LEACH) protocol is one of the most representative clustering routing protocols. Each node randomly generates a random number and compares it with a threshold. If the number is less than the threshold, the node becomes the cluster head. Other nodes join clusters according to proximity. The cluster head merges the collected data and sends it directly to the base station. In this scheme, each node has an equal probability of becoming a cluster head, the number of clusters is determined experimentally, and the remaining energy of the nodes is not considered when selecting cluster heads.

[0005] The LEACH-C (LEACH-centralized) protocol is an improved version of the LEACH protocol. Unlike the distributed cluster head election used by LEACH, LEACH-C uses a centralized cluster head election method. Each node sends its location and remaining energy information to the base station (BS). The base station uses simulated annealing to select a cluster head from nodes with remaining energy above the average, based on the principle of minimizing the total transmission distance. However, this method requires estimation of the optimal number of clusters through experiments or formulas before the algorithm runs. These formulas often make certain assumptions and are only applicable to specific networks. Some formulas can only calculate the range of optimal cluster numbers; the accurate value still needs to be determined experimentally.

[0006] Therefore, an energy-saving routing scheme for wireless sensor networks based on intelligent algorithms is desired. Summary of the Invention

[0007] In view of this, this application proposes an energy-saving routing method and system for wireless sensor networks based on intelligent algorithms. When selecting cluster head nodes, it can take into account the remaining energy of the nodes, the network spatial topology, and the distance information between the nodes and the base station. Moreover, it does not require estimating the optimal number of clusters before the algorithm runs. It can adapt to different data fusion rates for cluster routing, thereby improving energy utilization efficiency and extending the network lifespan.

[0008] According to one aspect of this application, an energy-saving routing method for wireless sensor networks based on intelligent algorithms is provided, comprising:

[0009] Model the wireless sensor network to obtain the network space topology matrix;

[0010] Construct the degree matrix of the wireless sensor network relative to the base station;

[0011] Obtain the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period;

[0012] The remaining energy values ​​of each wireless sensor at multiple predetermined time points within a predetermined time period are arranged into input vectors according to the time dimension to obtain the time sequence input vector of the remaining energy of multiple wireless sensors.

[0013] The time-series feature vectors of the remaining energy of the multiple wireless sensors are obtained by extracting features from the time-series input vectors of the multiple wireless sensors using a time-series feature extractor based on a deep neural network model.

[0014] Each of the remaining energy time-series feature vectors of the multiple wireless sensors is first mapped to the feature space of the network space topology matrix, and then mapped to the feature space of the degree matrix to obtain multiple corrected remaining energy time-series feature vectors of the wireless sensors, which serve as the remaining energy time-series features of the multiple wireless sensors; and

[0015] The cluster head is determined based on the time-series characteristics of the remaining energy of the multiple wireless sensors.

[0016] According to another aspect of this application, an energy-saving routing system for wireless sensor networks based on intelligent algorithms is provided, comprising:

[0017] The modeling module is used to model wireless sensor networks to obtain the network space topology matrix;

[0018] A degree matrix construction module is used to construct the degree matrix of the wireless sensor network relative to the base station;

[0019] The remaining energy value acquisition module is used to acquire the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period.

[0020] The vectorization module is used to arrange the remaining energy values ​​of each wireless sensor at multiple predetermined time points within a predetermined time period into input vectors according to the time dimension to obtain the time sequence input vector of the remaining energy of multiple wireless sensors.

[0021] The time-series feature extraction module is used to extract features from the time-series input vectors of the remaining energy of the multiple wireless sensors using a time-series feature extractor based on a deep neural network model to obtain the time-series feature vectors of the remaining energy of the multiple wireless sensors.

[0022] The mapping module is used to map each of the remaining energy time-series feature vectors of the multiple wireless sensors to the feature space of the network space topology matrix and then to the feature space of the degree matrix to obtain multiple corrected remaining energy time-series feature vectors of the wireless sensors as multiple remaining energy time-series features of the wireless sensors; and

[0023] The cluster head determination module is used to determine the cluster head based on the time-series characteristics of the remaining energy of the multiple wireless sensors.

[0024] According to embodiments of this application, the remaining energy values ​​of each wireless sensor in a wireless sensor network are collected. Data processing and analysis algorithms are then introduced at the backend to perform time-series analysis of the remaining energy values ​​of each wireless sensor. Simultaneously, the network spatial topology of the wireless sensor network and the distance information between nodes and base stations are combined to perform time-series feature mapping correction of the remaining energy values ​​of each wireless sensor. This allows for the prediction of the probability of a node becoming a cluster head and the selection of the optimal cluster head node. This improves energy efficiency and extends the network's lifespan.

[0025] Other features and aspects of this application will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0026] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this application together with the specification and serve to explain the principles of this application.

[0027] Figure 1 A flowchart illustrating an energy-saving routing method for wireless sensor networks based on intelligent algorithms according to an embodiment of this application is shown.

[0028] Figure 2A flowchart illustrating sub-step S170 of a smart algorithm-based energy-saving routing method for wireless sensor networks according to an embodiment of this application is shown.

[0029] Figure 3 A flowchart illustrating sub-step S171 of an energy-saving routing method for wireless sensor networks based on intelligent algorithms according to an embodiment of this application is shown.

[0030] Figure 4 A block diagram of an energy-efficient routing system for wireless sensor networks based on intelligent algorithms, according to an embodiment of this application, is shown.

[0031] Figure 5 The diagram illustrates an application scenario of an energy-saving routing method for wireless sensor networks based on intelligent algorithms, according to an embodiment of this application. Detailed Implementation

[0032] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are also within the scope of protection of this application.

[0033] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0034] Various exemplary embodiments, features, and aspects of this application will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0035] Furthermore, to better illustrate this application, numerous specific details are provided in the following detailed embodiments. Those skilled in the art should understand that this application can be implemented without certain specific details. In some instances, methods, means, components, and circuits well-known to those skilled in the art have not been described in detail in order to highlight the main points of this application.

[0036] To address the aforementioned technical problems, the technical concept of this application is to collect the remaining energy values ​​of each wireless sensor in a wireless sensor network, and then introduce data processing and analysis algorithms in the backend to perform time-series analysis of the remaining energy values ​​of each wireless sensor. Simultaneously, by combining the network spatial topology of the wireless sensor network and the distance information between nodes and base stations, the time-series characteristic mapping correction of the remaining energy of each wireless sensor is performed. This allows for the prediction of the probability of a node becoming a cluster head and the selection of the optimal cluster head node. In this way, when selecting cluster head nodes, the remaining energy of the nodes, the network spatial topology, and the distance information between the nodes and base stations are considered. Furthermore, it eliminates the need to estimate the optimal number of clusters before the algorithm runs, and can adaptively perform cluster routing for different data fusion rates, thereby improving energy utilization efficiency and extending the network's lifetime.

[0037] Figure 1 A flowchart illustrating an energy-saving routing method for wireless sensor networks based on intelligent algorithms according to an embodiment of this application is shown. Figure 1 As shown, the energy-saving routing method for wireless sensor networks based on intelligent algorithms according to an embodiment of this application includes the following steps: S110, modeling the wireless sensor network to obtain a network space topology matrix; S120, constructing a degree matrix of the wireless sensor network relative to a base station; S130, obtaining the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period; S140, arranging the remaining energy values ​​of each wireless sensor at multiple predetermined time points within the predetermined time period into input vectors according to the time dimension to obtain a time-series input vector of the remaining energy of multiple wireless sensors; S150, through deep... The temporal feature extractor of the degree neural network model extracts features from the temporal input vectors of the remaining energy of the multiple wireless sensors to obtain temporal feature vectors of the remaining energy of the multiple wireless sensors; S160, each temporal feature vector of the remaining energy of the multiple wireless sensors is first mapped to the feature space where the network space topology matrix is ​​located, and then mapped to the feature space where the degree matrix is ​​located to obtain multiple corrected temporal feature vectors of the remaining energy of the multiple wireless sensors as temporal features of the remaining energy of the multiple wireless sensors; and S170, the cluster head is determined based on the temporal features of the remaining energy of the multiple wireless sensors.

[0038] Specifically, in the technical solution of this application, firstly, the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period are obtained. Next, considering that the remaining energy values ​​of each wireless sensor exhibit a dynamic temporal variation pattern, meaning that the remaining energy values ​​of each wireless sensor at multiple predetermined time points within the predetermined time period are temporally correlated, the technical solution of this application requires arranging the remaining energy values ​​of each wireless sensor at multiple predetermined time points within the predetermined time period into input vectors according to the time dimension to obtain multiple temporal input vectors of the remaining energy of the wireless sensors. This is used to integrate the temporal distribution information of the remaining energy values ​​of each wireless sensor in the wireless sensor network.

[0039] Then, the time-series input vectors of the remaining energy of the multiple wireless sensors are respectively processed by a time-series feature extractor based on a one-dimensional convolutional layer to extract the time-series dynamic feature information of the remaining energy value of each wireless sensor in the time dimension, thereby obtaining the time-series feature vectors of the remaining energy of multiple wireless sensors.

[0040] Accordingly, in step S150, the temporal feature extractor based on the deep neural network model is a temporal feature extractor based on a one-dimensional convolutional layer. It is worth noting that a one-dimensional convolutional layer is a commonly used layer type in deep neural networks, used to extract features from temporal data. It is mainly applied to processing data with a time-series structure. In this application, a one-dimensional convolutional layer is used as a temporal feature extractor to extract features from the temporal input vector of the remaining energy of multiple wireless sensors. The one-dimensional convolutional layer performs convolution operations on the input sequence by sliding a fixed-size window (convolutional kernel), thereby obtaining local features. The weight parameters of the convolutional kernel are automatically learned based on training data to capture important patterns and features in the input sequence. The main function of the one-dimensional convolutional layer is to extract local features from the input sequence through convolution operations, thereby capturing temporal patterns and related features in the input sequence. These features can be further used for subsequent classification, regression, or other tasks. One-dimensional convolutional layers have good feature extraction capabilities and parameter efficiency in deep learning, helping the model learn more meaningful representations from temporal data. In summary, a one-dimensional convolutional layer is a deep neural network layer used to extract features from temporal data. It is used in energy-efficient routing methods for wireless sensor networks to extract temporal features of the remaining energy of multiple wireless sensors to help identify cluster head nodes.

[0041] Furthermore, considering the temporal correlation between the remaining energy values ​​of the various wireless sensors in the wireless sensor network, this correlation exists in the spatial topology between the various wireless sensors. Therefore, in the technical solution of this application, the wireless sensor network is further modeled to obtain a network spatial topology matrix, wherein the values ​​at each off-diagonal position in the network spatial topology matrix are the spatial distance values ​​between the corresponding two wireless sensors.

[0042] Furthermore, when selecting cluster head nodes in practice, it is necessary to consider not only the spatial topology information of the wireless sensor network, but also the distance information between each node in the wireless sensor network and the base station, in order to improve energy utilization efficiency. Therefore, in the technical solution of this application, a degree matrix of the wireless sensor network relative to the base station is further constructed, wherein the value of each position on the diagonal of the degree matrix is ​​the spatial distance value between the corresponding wireless sensor and the base station.

[0043] Next, each of the remaining energy time-series feature vectors of the multiple wireless sensors is first mapped to the feature space of the network space topology matrix, and then mapped to the feature space of the degree matrix to obtain multiple corrected remaining energy time-series feature vectors of the wireless sensors. It should be understood that by mapping each remaining energy time-series feature vector of the wireless sensors to the feature space of the network space topology matrix, the spatial relationships between nodes can be considered. That is, since the network space topology matrix describes the spatial distance between nodes, by mapping the remaining energy time-series feature vectors of the wireless sensors to this feature space, the spatial information between nodes can be captured. In this way, the remaining energy time-series feature vectors of the wireless sensors can better reflect the spatial relationships between nodes, thereby better guiding the selection of cluster heads. Further, after mapping the remaining energy time-series feature vectors of the wireless sensors to the feature space of the network space topology matrix, they are further mapped to the feature space of the degree matrix, where the degree matrix describes the spatial distance between nodes and the base station. By mapping the feature vectors to this feature space, the spatial relationship between nodes and the base station can be considered. In this way, the spatial relationships between nodes and base stations, as well as the spatial relationships between nodes, can be comprehensively considered to obtain a more comprehensive and accurate feature representation.

[0044] Next, the time-series feature vectors of the remaining energy of the multiple corrected wireless sensors are processed by a classifier to obtain multiple probability values. That is, the mapping feature information obtained by mapping the time-series features of the remaining energy of each wireless sensor to the high-dimensional feature space of the spatial topology of the wireless sensor network and the spatial distance between nodes and the base station is used for classification processing. This is used to calculate the probability of each wireless sensor, and the wireless sensor corresponding to the highest probability value is selected as the cluster head. In this way, when selecting cluster head nodes, the remaining energy of the nodes, the network spatial topology, and the distance information between the nodes and the base station are considered. Furthermore, it eliminates the need to estimate the optimal cluster number before the algorithm runs, and can adaptively perform cluster routing for different data fusion rates, thereby improving energy utilization efficiency.

[0045] Accordingly, in step S170, as Figure 2 As shown, determining the cluster head based on the time-series characteristics of the remaining energy of the multiple wireless sensors includes: S171, passing the time-series characteristic vectors of the remaining energy of the multiple corrected wireless sensors through a classifier to obtain multiple probability values; and S172, taking the wireless sensor corresponding to the one with the largest probability value as the cluster head.

[0046] It should be understood that in step S171, the classifier determines the probability of each wireless sensor becoming a cluster head based on the input feature vector and outputs the corresponding probability value. In step S172, the wireless sensor with the highest probability value is selected as the cluster head from multiple probability values. This means that after evaluation by the classifier, the wireless sensor with the highest probability value is selected as the cluster head. Selecting the wireless sensor with the highest probability as the cluster head ensures that the sensor has the highest energy level and lower energy consumption, thereby improving the energy efficiency of the entire wireless sensor network. In summary, step S171 calculates the probability value of each wireless sensor becoming a cluster head through the classifier, while step S172 selects the wireless sensor with the highest probability value as the cluster head to achieve energy-saving routing in the wireless sensor network.

[0047] In step S171, as follows Figure 3 As shown, the multiple corrected wireless sensor remaining energy time-series feature vectors are passed through a classifier to obtain multiple probability values, including: S1711, optimizing the corresponding corrected wireless sensor remaining energy time-series feature vectors based on the remaining energy time-series feature vectors of each wireless sensor to obtain multiple optimized wireless sensor remaining energy time-series feature vectors; and S1712, passing the multiple optimized wireless sensor remaining energy time-series feature vectors through the classifier to obtain the multiple probability values.

[0048] It should be understood that in step S171, the process of obtaining multiple probability values ​​from multiple corrected wireless sensor residual energy time-series feature vectors through a classifier includes two sub-steps: S1711 and S1712. Step S1711 aims to further process the feature vectors to extract more informative features, thereby improving the performance and accuracy of the classifier.

[0049] Specifically, in the technical solution of this application, each of the multiple wireless sensor remaining energy time-series feature vectors expresses the local time-series correlation feature of the remaining energy value of the corresponding wireless sensor. Thus, after mapping the wireless sensor remaining energy time-series feature vector to the feature space where the network space topology matrix is ​​located and then to the feature space where the degree matrix is ​​located, the topological correlation mapping of the remaining energy value time-series correlation feature of the corresponding wireless sensor under the wireless sensor space topology can be further performed. That is, each set of corresponding corrected wireless sensor remaining energy time-series feature vectors is equivalent to the interpolated spatial topological correlation mapping hybrid of the corresponding wireless sensor remaining energy time-series feature vectors.

[0050] Thus, in order to enhance the spatial topological association mapping effect based on the consistency of the expression of the corrected wireless sensor remaining energy time-series feature vector with the local temporal correlation features of its corresponding wireless sensor remaining energy time-series feature vector, based on the wireless sensor remaining energy time-series feature vector, for example denoted as... The corrected time-series feature vector of the remaining energy of the wireless sensor, for example, is denoted as... The optimization process includes the following steps:

[0051] The remaining energy time-series feature vectors of the wireless sensor are respectively... and the corrected wireless sensor remaining energy time-series feature vector Through common mapping weight matrix Perform a common multidimensional mapping to obtain the time-series mapping feature vector of the remaining energy of the wireless sensor. and the corrected wireless sensor remaining energy time-series mapping feature vector :

[0052] ;

[0053] ;

[0054] in, Represents matrix multiplication;

[0055] The time-series mapping feature vector of the remaining energy of the wireless sensor and the corrected wireless sensor remaining energy time-series mapping feature vector by After the function is activated, the unit eigenvector is subtracted and the absolute value of each eigenvalue is calculated to obtain the time-series prior decoupling eigenvector of the remaining energy of the wireless sensor. and the time-series prior decoupled feature vector of the remaining energy of the corrected wireless sensor :

[0056] ;

[0057] ;

[0058] in It is a unit eigenvector. This represents vector subtraction.

[0059] Decouple the time-series prior feature vector of the remaining energy of the wireless sensor The square of the L2 norm and the time-series prior decoupling eigenvector of the corrected wireless sensor remaining energy The eigenvector term of the optimized, corrected wireless sensor residual energy temporal prior decoupled eigenvector is obtained by multiplying the inverses of the eigenvalues ​​bit by bit:

[0060] ;

[0061] in, This indicates multiplication by position. The eigenvector term representing the optimized, corrected, time-prior decoupling eigenvector of the remaining energy of the wireless sensor;

[0062] The eigenvector terms of the optimized and corrected time-series prior decoupling feature vector of the remaining energy of the wireless sensor are compared with the eigenvector terms of the corrected time-series prior decoupling feature vector of the remaining energy of the wireless sensor. The bias term obtained by squared L2 norm is then applied to obtain the optimized corrected residual energy temporal prior decoupling eigenvector of the wireless sensor. , This represents vector addition.

[0063] That is, for the time-series feature vector of the remaining energy of the wireless sensor and the corrected wireless sensor remaining energy time-series feature vector In the feature information extraction process, the fitting time-series correlation features of the prediction index are fused. Based on the idea of ​​calibration regularization, this is achieved by fusing the time-series feature vector of the remaining energy of the wireless sensor. and the corrected wireless sensor remaining energy time-series feature vector Outlier characterization decoupling realizes the time-series feature vector of the remaining energy of the wireless sensor. and the corrected wireless sensor remaining energy time-series feature vector Prior constraint recovery of the multidimensional data stream. Specifically, the temporal feature vector of the remaining energy of the wireless sensor is reconstructed using a suboptimal enhancement mechanism. and the corrected wireless sensor remaining energy time-series feature vector The manifold topology structure is used to achieve the time-series feature vector of the remaining energy of the wireless sensor. and the corrected wireless sensor remaining energy time-series feature vector The fitting samples and prediction results are correlated through a collaborative feature enhancement mapping, thereby maintaining the time-series feature vector of the remaining energy of the wireless sensor. and the corrected wireless sensor remaining energy time-series feature vector While maintaining consistency in the expression of local temporal correlations, enhanced feature representation is achieved through spatial distribution correlation mapping. This effectively improves the overall expression of the optimized and corrected temporal feature vector of remaining energy of the wireless sensor, thereby enhancing the accuracy of the classification results obtained by the classifier. In this way, when selecting cluster head nodes, the remaining energy of the nodes, the network spatial topology, and the distance information between the nodes and the base station are considered. Furthermore, the optimal cluster number does not need to be estimated before the algorithm runs, and clustering routing can be adaptively performed at different data fusion rates, thereby improving energy utilization efficiency and extending the network lifetime.

[0064] Further, in step S1712, passing the plurality of optimized wireless sensor remaining energy time-series feature vectors through the classifier to obtain the plurality of probability values ​​includes: using the classifier's plurality of fully connected layers to perform fully connected encoding on the plurality of optimized wireless sensor remaining energy time-series feature vectors to obtain a plurality of encoded classification feature vectors; and passing the plurality of encoded classification feature vectors through the classifier's Softmax classification function to obtain the plurality of probability values.

[0065] As you can understand, the role of a classifier is to learn classification rules and classifiers using given categories and known training data, and then classify (or predict) unknown data. Logistic regression and SVM are commonly used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are needed to form the multi-class classification. However, this is prone to errors and inefficient. A commonly used multi-class classification method is the Softmax classification function.

[0066] It's worth noting that fully connected encoding is a technique that maps input data to fixed-length encoded vectors. In this step, multiple fully connected layers of the classifier are used to fully connect and encode the optimized temporal feature vector of the wireless sensor's remaining energy, resulting in multiple encoded categorical feature vectors. Fully connected layers are a common layer type in deep neural networks, where each neuron is connected to all neurons in the previous layer. In this case, the fully connected layers map the input data to a new feature space by multiplying the input feature vector by the weight matrix and applying an activation function. Multiple fully connected layers can progressively extract higher-level feature representations. The purpose of fully connected encoding is to transform the optimized feature vector into a more representative encoded vector. Through the combination of multiple fully connected layers, the network can learn more abstract and discriminative feature representations. These encoded categorical feature vectors better describe the temporal characteristics of the wireless sensor's remaining energy and provide the classifier with richer information to more accurately predict the probability of each wireless sensor becoming a cluster leader. In subsequent stages, these encoded categorical feature vectors are passed through the classifier's Softmax classification function, transforming the encoded feature vectors into corresponding probability values. The Softmax function maps the value of each encoded categorical feature vector to a probability value between 0 and 1, representing the likelihood that the wireless sensor will become a cluster leader. These probability values ​​can be used to select the wireless sensor with the highest probability as the cluster leader, enabling energy-efficient routing in wireless sensor networks.

[0067] In summary, the energy-saving routing method for wireless sensor networks based on intelligent algorithms, as described in this application, can improve energy utilization efficiency and extend the network's lifespan.

[0068] Figure 4 A block diagram of an energy-efficient routing system 100 for wireless sensor networks based on intelligent algorithms, according to an embodiment of this application, is shown. Figure 4As shown, the energy-saving routing system 100 for wireless sensor networks based on intelligent algorithms according to an embodiment of this application includes: a modeling module 110 for modeling the wireless sensor network to obtain a network space topology matrix; a degree matrix construction module 120 for constructing a degree matrix of the wireless sensor network relative to a base station; a remaining energy value acquisition module 130 for acquiring the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period; a vectorization module 140 for arranging the remaining energy values ​​of each wireless sensor at multiple predetermined time points within the predetermined time period into input vectors according to the time dimension to obtain a time-series input vector of the remaining energy of multiple wireless sensors; and time-series feature extraction. Module 150 is used to extract features from the time-series input vectors of the remaining energy of the multiple wireless sensors using a time-series feature extractor based on a deep neural network model to obtain time-series feature vectors of the remaining energy of the multiple wireless sensors; mapping module 160 is used to map each time-series feature vector of the remaining energy of the multiple wireless sensors to the feature space of the network space topology matrix and then to the feature space of the degree matrix to obtain multiple corrected time-series feature vectors of the remaining energy of the multiple wireless sensors as time-series features of the remaining energy of the multiple wireless sensors; and cluster head determination module 170 is used to determine the cluster head based on the time-series features of the remaining energy of the multiple wireless sensors.

[0069] In one possible implementation, the values ​​at each off-diagonal position in the network spatial topology matrix are the spatial distance values ​​between the corresponding two wireless sensors.

[0070] In one possible implementation, the values ​​at each position on the diagonal of the degree matrix represent the spatial distance between the corresponding wireless sensor and the base station.

[0071] Here, those skilled in the art will understand that the specific functions and operations of each unit and module in the above-described intelligent algorithm-based wireless sensor network energy-saving routing system 100 have been referenced above. Figures 1 to 3 The energy-saving routing method for wireless sensor networks based on intelligent algorithms has been described in detail, and therefore, its repeated description will be omitted.

[0072] As described above, the energy-saving routing system 100 for wireless sensor networks based on intelligent algorithms according to embodiments of this application can be implemented in various wireless terminals, such as servers with energy-saving routing algorithms for wireless sensor networks based on intelligent algorithms. In one possible implementation, the energy-saving routing system 100 for wireless sensor networks based on intelligent algorithms according to embodiments of this application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the energy-saving routing system 100 for wireless sensor networks based on intelligent algorithms can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the energy-saving routing system 100 for wireless sensor networks based on intelligent algorithms can also be one of many hardware modules of the wireless terminal.

[0073] Alternatively, in another example, the smart algorithm-based wireless sensor network energy-saving routing system 100 and the wireless terminal can also be separate devices, and the smart algorithm-based wireless sensor network energy-saving routing system 100 can be connected to the wireless terminal via wired and / or wireless networks, and transmit interactive information in accordance with an agreed data format.

[0074] Figure 5 An application scenario diagram illustrating an energy-saving routing method for wireless sensor networks based on intelligent algorithms according to an embodiment of this application is shown. For example... Figure 5 As shown, in this application scenario, firstly, the data from each wireless sensor in the wireless sensor network (e.g., ...) is acquired. Figure 5 The remaining energy values ​​of N1 (as shown) at multiple predetermined time points within a predetermined time period (e.g., Figure 5 As shown in D1), corresponding to two wireless sensors (e.g., Figure 5 Spatial distance values ​​between (e.g., N1 as shown) Figure 5 As shown in D2), and the corresponding wireless sensor (e.g., Figure 5 N1 as shown in the diagram) and the base station (e.g., Figure 5 Spatial distance values ​​between (e.g., N2 as shown) Figure 5 As shown in D3), the remaining energy values ​​of each wireless sensor at multiple predetermined time points within a predetermined time period, the spatial distance values ​​between corresponding two wireless sensors, and the spatial distance values ​​between the corresponding wireless sensor and the base station are then input to a server deployed with a wireless sensor network energy-saving routing algorithm based on intelligent algorithms (e.g., Figure 5In the S shown, the server can use the smart algorithm-based wireless sensor network energy-saving routing algorithm to process the remaining energy value of each wireless sensor at multiple predetermined time points within a predetermined time period, the spatial distance value between two corresponding wireless sensors, and the spatial distance value between the corresponding wireless sensor and the base station to obtain multiple probability values. Then, the wireless sensor corresponding to the largest of the multiple probability values ​​is taken as the cluster head.

[0075] Furthermore, this application also provides a wireless sensor network clustering routing method based on the cultural gene algorithm, which includes the following steps: (1) Model the wireless sensor network, initialize the network topology and wireless communication energy consumption model; sort all sensor nodes in order of distance from the base station from near to far, the sorting is to allow sensor nodes far from the base station to select the cluster head first. (2) For all surviving nodes, call the cultural gene algorithm to select the cluster head and free nodes from the nodes with remaining energy higher than the median value, based on the optimization principle of reducing the total energy consumption in this round and balancing the energy consumption of each node, and the remaining nodes become cluster member nodes. (3) Cluster member nodes select the cluster head to form a cluster, and communicate with the base station in one round through the cluster head forwarding method, while free nodes communicate directly with the base station in one round, and calculate the energy consumption of each node in this round. (4) Update the remaining energy of all nodes, determine whether the sensor node has died, if there are still nodes surviving, jump to step (2), otherwise the program ends.

[0076] Specifically, the process of step (1) is as follows: Initialize the length and width of the monitoring area, base station coordinates, number of sensor nodes, coordinates, initial energy, data fusion rate, and packet size; initialize the wireless communication energy consumption parameters; calculate the distance between each sensor node and the base station; and store the position coordinates of all sensor nodes in array S in order of distance from the base station from closest to farthest. Specifically, the process of step (2) can use the MPCESSALS algorithm.

[0077] Furthermore, a method for calculating the energy consumption of sensor nodes in a single transmission round is also provided. Specifically, this includes: (1) Cluster establishment: using Boolean vectors The type of sensor node in the network is represented by n, where n represents the number of surviving nodes. Each component of x corresponds to a sensor node. 1 indicates that the node is a candidate cluster head node, and 0 indicates that the node is a cluster member node. There are two conditions for a candidate cluster head to be selected as the cluster head: first, the remaining energy is equal to or greater than the median of the remaining energy of all nodes; second, there is a member node that selects this candidate cluster head as the cluster head. Member nodes select their own cluster head according to the previous sorting order and the principle of proximity. If the cluster members connected to the nearest cluster head have exceeded the limit, the next nearest cluster head is selected, and so on. Candidate nodes that are not selected as cluster heads and other nodes that do not belong to any cluster become free nodes. (2) Transmission energy consumption calculation: Member nodes send data to the cluster head. After the cluster head collects the data of all member nodes, it performs data fusion and then forwards it to the base station. Free nodes communicate directly with the base station. The energy consumption of each node in this process can be calculated according to the wireless communication energy consumption model.

[0078] It should be understood that this application uses the Cultural Gene Algorithm to select cluster heads. The number of cluster heads does not need to be specified before the algorithm runs; it is automatically determined by the Cultural Gene Algorithm through optimization. The objective function of the Cultural Gene Algorithm is a weighted sum of the total network energy consumption in this round and the standard deviation of the remaining energy of the nodes. While considering energy consumption balance, it adopts a greedy strategy to minimize the total energy consumption of the entire network in each round, thereby extending the network's lifetime. This application can also consider the energy consumption of data fusion in the objective function, and the number of cluster heads affected by the data fusion rate is also determined by the algorithm through optimization. Therefore, the algorithm can adapt to changes in the data fusion rate.

[0079] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0080] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A power-saving routing method for wireless sensor networks based on intelligent algorithms, characterized in that, include: Model the wireless sensor network to obtain the network space topology matrix; Construct the degree matrix of the wireless sensor network relative to the base station; Obtain the remaining energy values ​​of each wireless sensor in the wireless sensor network at multiple predetermined time points within a predetermined time period; The remaining energy values ​​of each wireless sensor at multiple predetermined time points within a predetermined time period are arranged into input vectors according to the time dimension to obtain the time sequence input vector of the remaining energy of multiple wireless sensors. The time-series feature vectors of the remaining energy of the multiple wireless sensors are obtained by extracting features from the time-series input vectors of the multiple wireless sensors using a time-series feature extractor based on a deep neural network model. Each of the remaining energy time-series feature vectors of the multiple wireless sensors is first mapped to the feature space where the network space topology matrix is ​​located, and then mapped to the feature space where the degree matrix is ​​located to obtain multiple corrected remaining energy time-series feature vectors of the wireless sensors as multiple remaining energy time-series features of the wireless sensors; And based on the time-series characteristics of the remaining energy of the multiple wireless sensors, the cluster head is determined; Based on the time-series characteristics of the remaining energy of the multiple wireless sensors, the cluster head is determined, including: The time-series feature vectors of the remaining energy of the multiple corrected wireless sensors are passed through a classifier to obtain multiple probability values; and the wireless sensor corresponding to the largest of the multiple probability values ​​is taken as the cluster head. The time-series feature vectors of the remaining energy of the multiple corrected wireless sensors are passed through a classifier to obtain multiple probability values, including: Based on the remaining energy time-series feature vectors of each wireless sensor, the corresponding corrected remaining energy time-series feature vectors of the wireless sensors are optimized to obtain multiple optimized remaining energy time-series feature vectors of the wireless sensors; and the multiple optimized remaining energy time-series feature vectors of the wireless sensors are passed through the classifier to obtain the multiple probability values; The optimized time-series feature vectors of remaining energy of the wireless sensors are passed through the classifier to obtain the multiple probability values, including: The classifier uses multiple fully connected layers to fully connect and encode the time-series feature vectors of the remaining energy of the multiple optimized wireless sensors to obtain multiple encoded classification feature vectors; and the multiple encoded classification feature vectors are then passed through the classifier's Softmax classification function to obtain the multiple probability values.

2. The energy-saving routing method for wireless sensor networks based on intelligent algorithms according to claim 1, characterized in that, The values ​​at each off-diagonal position in the network space topology matrix represent the spatial distance between the corresponding two wireless sensors.

3. The energy-saving routing method for wireless sensor networks based on intelligent algorithms according to claim 2, characterized in that, The values ​​at each position on the diagonal of the degree matrix represent the spatial distance between the corresponding wireless sensor and the base station.

4. The energy-saving routing method for wireless sensor networks based on intelligent algorithms according to claim 3, characterized in that, The temporal feature extractor based on the deep neural network model is a temporal feature extractor based on a one-dimensional convolutional layer.

Images