Method for planning path of mobile node in Internet of Things based on firefly algorithm

Abstract

The invention provides a firefly algorithm-based mobile node path planning method in the Internet of Things, and the main research content is to balance the energy consumption of each mobile sensor node so as to prolong the life cycle of the network. Since the mobile sensor nodes cannot be charged, in order to support certain applications sensitive to time, a plurality of mobile sensors are used for collecting information of each node, so that the nodes complete information collection within specified time. The method mainly comprises the following steps of: (1) dividing a wireless sensor network area into a plurality of square grids according to a certain rule, and describing the structure of the whole network by using an incomplete graph model; and (2) according to the requirement of time sensitivity, the appropriate number of the mobile sensor nodes is set, so that the mobile sensor nodes can effectively complete data collection work within specified time. And (3) in order to balance the energy consumption of each mobile sensor, planning a reasonable data collection route for the mobile sensors by adopting multi-mobile sensor node path planning of a firefly algorithm.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method for mobile node path planning in the Internet of Things based on a firefly algorithm. Background Art

[0002] In recent years, as the Internet of Things (IoT) has developed toward intelligence and ubiquity, path planning for mobile network nodes (such as drones, autonomous vehicles, mobile sensors, and automated guided vehicles (AGVs) in complex environments has become one of the core technologies supporting the implementation of intelligent applications. The diversity of IoT scenarios (such as industrial manufacturing, smart cities, and emergency rescue) has given rise to differentiated demands for three-dimensional spatial planning, anti-interference path optimization, and multi-robot collaborative strategies. This has driven the evolution of path planning technology toward lightweight, intelligent, and scenario-specific customization, becoming a key technical bridge connecting the perception layer and the application layer.

[0003] With the intertwined evolution of communications, embedded systems, and sensor technologies, miniature wireless sensors—combining environmental sensing, data processing, and wireless communication—have become a research hotspot in information science. The emergence and development of these devices is not only the result of the integration of advanced technologies such as microelectromechanical systems (MEMS), wireless communication protocols, and low-power embedded computing, but also a typical product of the deep integration of multidisciplinary technologies. Despite their miniaturization, wireless sensors possess a comprehensive functional architecture: integrating high-precision sensing modules for real-time collection of multi-dimensional environmental data such as temperature, humidity, noise, hazardous gas concentrations, and water parameters; relying on on-chip microprocessors for local processing such as data filtering and feature extraction; and utilizing wireless communication units for remote data transmission. This integrated design offers broad application value in industrial monitoring, smart cities, and ecological protection.

[0004] The application of wireless sensors relies on the collaborative architecture of wireless sensor networks (WSNs). This network deploys a vast number of microsensor nodes across a target monitoring area through self-organizing protocols, forming an intelligent system with distributed sensing capabilities. A typical architecture involves nodes self-organizing into a multi-hop network via wireless communication links. Collected environmental data (such as temperature, humidity, and stress signals) are aggregated in the form of data packets to sink nodes (central base stations) via single-hop direct transmission or multi-hop relaying. While this distributed network architecture enables wide-area monitoring, it also faces severe energy efficiency challenges. As the monitoring area expands, remote nodes must forward data via multi-hop routing, burdening intermediate nodes with a significant amount of data relay work and exponentially increasing energy consumption. Sink nodes near base stations, in particular, are prone to forming "hotspots" due to their need to process network-wide data traffic. Their energy consumption is 3-5 times higher than that of edge nodes. This uneven energy distribution can lead to "energy holes"—critical nodes prematurely exhaust their energy and fail, resulting in connectivity gaps or monitoring blind spots, significantly shortening the network's lifespan. The essence of this problem is that the existing routing mechanism fails to effectively balance the relationship between node load and energy consumption. It is urgent to introduce energy-aware routing strategies into the network layer protocol design, and build a more energy-efficient network architecture through technical means such as dynamically adjusting data forwarding paths and optimizing node sleep scheduling.

[0005] Sensor nodes in wireless sensor networks are typically powered by internal batteries. Limited by the size of the sensors, their battery capacity is limited. Therefore, energy consumption is a significant factor limiting their development. In many complex environments, maintaining proper sensor node operation and timely data upload is crucial. Therefore, reducing energy consumption within sensor nodes, balancing energy consumption across the entire network, and maintaining a healthy network are key research issues in wireless sensor networks. This situation changes somewhat in networks that use mobile sensor nodes as data collection tools. Mobile sensor nodes bear a significant load, and if a mobile sensor node runs out of energy, the entire network will cease to function. Therefore, balancing energy consumption across multiple mobile sensor nodes in such networks is a key research priority. Summary of the Invention

[0006] Based on the problems existing in the background technology, the present invention proposes a method for mobile node path planning in the Internet of Things based on the firefly algorithm. On the basis of a hybrid sensor network, the present invention studies the planning of mobile paths for collecting data for all mobile nodes in the network when the mobile sensor nodes also have limited energy reserves, thereby balancing the workload of each mobile node and making their energy exhaustion as close as possible.

[0007] The present invention proposes a method for mobile node path planning in the Internet of Things based on the firefly algorithm, comprising:

[0008] Sensor Network Region Partitioning: Network region partitioning techniques have been widely used in wireless sensor network monitoring activities. First, static sensor nodes are deployed within the region. Since it's impossible to evenly distribute sensors across a large area, methods such as aerial seeding are used to randomly distribute sensor nodes throughout the network. The monitored area is divided into multiple square subregions using a regular quadrilateral method, with each subregion containing a certain number of static nodes. Each square subregion varies in location and number of sensors, and information from these subregions can be presented in a specific manner.

[0009] Setting the number of mobile sensors: Considering that all mobile sensor nodes have the same mobility and energy consumption characteristics, the time and energy consumed by a mobile sensor node moving from one vertex to the nearest vertex are certain. Let E represent the energy consumed by a mobile sensor node during its movement. If the total distance moved by the node in the working cycle is large, the E value of the node is large. E. represents the energy value consumed by a mobile sensor node during the data collection process of the working cycle. If the area passed by the node contains more static sensor nodes, the E value of the node is large. Therefore, the total energy consumed by a mobile sensor node in the network life cycle is EC i =E mi +E gi . In the scenario involved in the present invention, during the data collection process, the time taken for transmitting data after two nodes establish a connection is extremely short, which is almost negligible compared to the movement time of the mobile sensor node, which means that the time required for a sensor node to complete its work is only related to the total distance it moves. T represents the maximum time limit required for all static sensor nodes to upload their data once when collecting data for an area. This is an artificially set parameter. For example, in some scenarios, in order to obtain valid data, data from the entire network must be collected once every hour. N represents the number of nodes contained in the graph after the area is abstracted into a graph. V represents the movement speed of the mobile sensor. t represents the time required for a mobile sensor to move from a node to its nearest neighbor node at a speed of v. M represents the number of mobile sensor nodes, which is the parameter we want to obtain. Under the assumption, we believe that the M mobile sensor nodes evenly divide the N vertices, and the total distance traveled by each mobile sensor node when traversing all the vertices it is responsible for is the shortest. We can obtain an inequality: This formula gives the range of M. Since M may not be an integer, we set M to the smallest integer greater than M. Therefore, we assign N nodes to M mobile sensors, with each mobile sensor receiving an average of N / M nodes. Since N / M may not be an integer, N / M = C...D, where C represents the quotient and D represents the remainder. In this case, the first D mobile sensor nodes receive C+1 nodes, while the remaining mobile sensors receive C nodes.

[0010] Research focuses on the optimization of the firefly algorithm and its application in vehicle path planning. The flaws of the classic firefly algorithm are analyzed, aiming to address issues such as the algorithm's tendency to fall into local optimality and slow convergence. An improved strategy (DsFA) based on adaptive step size and attraction adjustment is proposed: By dynamically adjusting the step size, a larger step size is adopted in the early stages of the iteration to enhance global search capabilities, and a smaller step size is adopted in the later stages to improve local search accuracy. A control structure is introduced to dynamically determine the timing of step size changes based on the distance between the optimal individuals of two generations to avoid premature convergence. The initial attraction is set to 0.4 to maintain population diversity and enhance the algorithm's search capabilities. Subsequently, to address the discrete nature of VRPTW, the firefly individuals are discretized and encoded, and a two-dimensional matrix is ​​designed to represent customer allocation and path order, addressing decimal and out-of-bounds issues. A local search operator is introduced to optimize the decoded solution.

[0011] In its path planning algorithm design, the Firefly Algorithm (GLA) takes energy balance as its core objective and constructs a composite objective function encompassing both mobility energy (path length) and communication energy variance. This algorithm employs an attractivity model to replace the traditional pheromone mechanism. The attractivity of each vertex pair is dynamically adjusted based on the frequency of firefly visits and path quality. The attractivity of high-brightness paths is strengthened, while the attractivity of low-quality paths decays over time. The GLA models the multi-mobile node path planning problem as a multiple traveling salesman problem (MTSP), allowing mobile sensors to depart from any starting point without returning to their initial location, thus adapting to the flexible deployment requirements of real-world data collection. Through brightness-guided collaborative search, the algorithm optimizes the shortest single-node path while minimizing the energy variance between nodes, ensuring balanced energy consumption across multiple mobile sensors and effectively extending the overall network lifecycle.

[0012] The simulation experiment of the present invention was carried out in rectangular areas of different sizes (such as 450 meters by 300 meters and 1000 meters by 750 meters). 1000 to 10,000 static sensor nodes were randomly deployed. The mobile sensor communication radius was set to 50 meters and the moving speed was set to 5 meters per second. The energy consumption variance V and convergence time were used as indicators to verify the algorithm performance. The results show that the firefly algorithm can effectively reduce the energy consumption variance (in most cases, V < 400) and achieve node energy consumption balance. The firefly algorithm also converges faster and has greater stability in large-scale areas. Compared with traditional intelligent optimization algorithms, the firefly algorithm proposed in this invention has a smaller energy consumption variance under seven different data sets, verifying its superiority in complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A parallel diagram of the Firefly system is shown for a method of mobile node path planning in the Internet of Things based on the Firefly algorithm.

[0014] Figure 2 Shows the logic flow chart of the firefly algorithm.

[0015] Figure 3 The figure shows the operation flow chart of mobile node path planning in the Internet of Things.

[0016] Figure 4 A diagram showing the process of abstracting a network area into a graph. DETAILED DESCRIPTION

[0017] The present invention will be further described below with reference to the accompanying drawings.

[0018] according to Figure 1 First, the firefly population is divided into multiple subpopulations based on spatial regions, computational tasks, or iteration rounds, and these subpopulations are assigned to different computing nodes. Each node independently initializes the subpopulation's position and calculates its luminous intensity, then synchronizes the global optimal solution. In each iteration, each node processes the movement and brightness updates of the fireflies within the subpopulation in parallel. A master-slave architecture can be used, with the master node aggregating the global optimal solution and broadcasting it to the slave nodes, or a ring or fully connected network can be used to allow nodes to directly exchange local optimal solutions. A load balancing mechanism is introduced to ensure similar computational load across nodes, and synchronization points are set to periodically merge subpopulation states to avoid search direction deviations. After the iteration, the master node integrates the results from each subpopulation and outputs the global optimal solution. This approach accelerates convergence through parallel computing and improves the efficiency of solving large-scale optimization problems.

[0019] Example 1: A method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm is provided. The method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm includes addressing the key challenge of limited sensor battery capacity, dividing the network area into square grids, abstracting the network structure with a non-complete graph model, considering time sensitivity requirements, and calculating the appropriate number of mobile sensors by establishing inequalities to ensure that the data collection task is completed within the specified time. A path planning strategy based on the firefly algorithm is designed, comprehensively considering the energy consumption of the movement process and the communication process, optimizing the path with the objective function, and achieving energy consumption balance of each node. Simulation experiments verify the effectiveness of the algorithm. The results show that the proposed algorithm can effectively balance energy consumption, meet timeliness, and effectively solve the path planning problem of mobile nodes in the Internet of Things.

[0020] according to Figure 2 First, the number, position, maximum number of iterations, brightness attenuation coefficient and attraction parameters of the fireflies are initialized. The initial position of each firefly is randomly generated in the solution space and its objective function value is calculated as the initial luminous brightness. Then, for each firefly, by comparing the luminous brightness, it is moved to all fireflies brighter than itself. During the movement, the attraction decreases according to the attenuation coefficient as the distance increases. At the same time, a random movement mechanism is combined to balance the exploration and development capabilities. Then, the position of the firefly is updated and the luminous brightness of the new position is recalculated to evaluate the quality of the solution. Finally, it is checked whether the termination condition is met (such as reaching the maximum number of iterations or the solution tends to be stable). If not, it returns to continue iteration until the optimal solution is output.

[0021] Example 2: A method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm is provided. The method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm includes addressing the key challenge of limited sensor battery capacity, dividing the network area into square grids, abstracting the network structure with a non-complete graph model, considering time sensitivity requirements, and calculating the appropriate number of mobile sensors by establishing inequalities to ensure that the data collection task is completed within the specified time. A path planning strategy based on the firefly algorithm is designed, comprehensively considering the energy consumption of the movement process and the communication process, optimizing the path with the objective function, and achieving energy consumption balance of each node. Simulation experiments verify the effectiveness of the algorithm. The results show that the proposed algorithm can effectively balance energy consumption, meet timeliness, and effectively solve the path planning problem of mobile nodes in the Internet of Things.

[0022] according to Figure 3, using multi-objective discrete fireflies to delineate different planning areas, collect data from the mobile communication network node hierarchy, and establish independent virtual firefly beacon nodes. The distance between each individual beacon node is kept constant. Using a conversion platform, the mobile communication network node data is converted into virtual data, and a three-dimensional planning matrix is ​​established within the mobile communication network's bidirectional node hierarchy model. The firefly algorithm is used to calculate and verify the node path limits within the three-dimensional planning matrix, forming a limit standard. This limit standard is used to adjust the mobile communication network's coverage. Within a reasonable range, virtual firefly beacons are used to correlate mobile communication network node paths, achieving optimal path planning.

[0023] Example 3: A method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm is provided. The method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm includes addressing the key challenge of limited sensor battery capacity, dividing the network area into square grids, abstracting the network structure with a non-complete graph model, considering time sensitivity requirements, and calculating the appropriate number of mobile sensors by establishing inequalities to ensure that the data collection task is completed within the specified time. A path planning strategy based on the firefly algorithm is designed, comprehensively considering the energy consumption of the movement process and the communication process, optimizing the path with the objective function, and achieving energy consumption balance for each node. Simulation experiments verify the effectiveness of the algorithm. The results show that the proposed algorithm can effectively balance energy consumption, meet timeliness, and effectively solve the path planning problem of mobile nodes in the Internet of Things.

[0024] according to Figure 4 ,abstract the square sub-network area, regard each square sub-area as a vertex in the graph, and the distance between the center points of two square areas is the edge between the two vertices, which is used to represent the distance relationship between any two square sub-areas. ,The process of representing a wireless sensor network area with 4 square sub-areas as a graph, where each vertex corresponds to each sub-area in the network area, and the corresponding edge represents the distance relationship between the corresponding sub-areas.

[0025] Example 4: A method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm is provided. The method for path planning of mobile nodes in the Internet of Things based on the firefly algorithm includes addressing the key challenge of limited sensor battery capacity, dividing the network area into square grids, abstracting the network structure with a non-complete graph model, considering time sensitivity requirements, and calculating the appropriate number of mobile sensors by establishing inequalities to ensure that the data collection task is completed within the specified time. A path planning strategy based on the firefly algorithm is designed, comprehensively considering the energy consumption of the movement process and the communication process, optimizing the path with the objective function, and achieving energy consumption balance for each node. Simulation experiments verify the effectiveness of the algorithm. The results show that the proposed algorithm can effectively balance energy consumption, meet timeliness, and effectively solve the path planning problem of mobile nodes in the Internet of Things.

Claims

1. A method for mobile node path planning in the Internet of Things based on the firefly algorithm, characterized in that: The following steps are involved: S1: Grid division of sensor network; S2: Abstracting grid area into graph; S3: Using firefly algorithm to plan paths for mobile sensor network nodes; S4: Introducing evolutionary computation model with inertia weight; S5: Introducing population mutation operation with Gaussian distribution; S6: Draw a path planning map.

2. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1, characterized in that: In the S1 process, the purpose of grid division is to effectively manage static sensor nodes. The entire network area is divided into multiple subareas, each of which contains a certain number of static sensor nodes. Using subareas as basic units, we manage all sensor nodes within them in a unified manner, reducing the actual workload and improving the efficiency of the entire network. The network area is divided into multiple, identical square subareas using the square area division method. To improve the efficiency of the entire network area, the square subareas need to be appropriately sized.

3. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1, characterized in that: In the process S2, the area to be monitored is divided into a plurality of square sub-areas according to a regular quadrilateral division method, and each sub-area has a certain number of static nodes. According to the relevant theories of graph theory, the square sub-network area is abstracted, and each square sub-area is regarded as a vertex in the graph. The distance between the center points of two square areas is the edge between the two vertices, which is used to represent the distance relationship between any two square sub-areas.

4. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1, characterized in that: In the S3 process, in the firefly algorithm, the main purpose of fireflies emitting light is to serve as a signal system to attract other firefly individuals. The assumption is: 1) Fireflies are gender-neutral and are attracted to brighter fireflies. 2) A firefly's attractiveness is proportional to its brightness. For any two fireflies, one will move toward the brighter one, but brightness decreases with distance. 3) If no firefly brighter than a given one is found, it will move randomly. The firefly algorithm incorporates two elements: brightness and attractiveness. Brightness reflects the quality of a firefly's position and determines its direction of movement, while attractiveness determines the distance it will travel. By continuously updating brightness and attractiveness, the goal is optimized.

5. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1, characterized in that: During the S4 process, as the population continues to iterate, the distance between individuals decreases, while the relative attraction between them increases, reducing the algorithm's local search capability. The addition of a random term with a specific coefficient increases the search range and prevents the algorithm from prematurely falling into a local optimum. However, this may require multiple iterations to achieve the required accuracy, making it impossible to meet the required accuracy when the number of iterations is limited. To improve both the local and global search capabilities of the algorithm, inertia weights are introduced, increasing the guiding role of the optimal individual in the population on other individuals.

6. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1 is characterized in that During S5, a firefly population trapped in a local optimum is characterized by evolutionary stagnation: multiple iterations without a change in the population's optimal value. If, after six consecutive iterations, the population's global optimal value remains unchanged, its evolution is considered stagnant, indicating it has been trapped in a local optimum. To help the firefly population escape the constraints of its local optimum, a Gaussian distribution is used to mutate the firefly population. All firefly individuals in the population are ranked by the value of the objective function. The top 10% × n fireflies are then used to replace the bottom 10% × n fireflies in the population. Gaussian mutation is then performed on the updated firefly population's state.

7. The method for mobile node path planning in the Internet of Things based on the firefly algorithm according to claim 1, characterized in that: In the process of S6, the algorithms and improved solutions of rights S4, S5, and S6 are applied to the wireless sensor network node path planning to draw a path planning graph.

Images