Target positioning method and system based on distributed initial value constraint particle swarm optimization algorithm

Abstract

This invention discloses a target localization method and system based on a distributed initial value constrained particle swarm algorithm. The target localization method includes: S1, obtaining the distance between each sensor and the moving target, and calculating the moving target m. j The location estimation sensor is deployed in the space to be located to communicate with the moving target and obtain its own position relative to the moving target. j The distance S2 between them, and the feasible region of the moving target determined by wireless ranging, form the search area S. PSO S3, with search area S PSO To constrain the positions of particles in a particle swarm, a particle swarm optimization (PSO) algorithm is used to obtain the position of a moving target. This method can accelerate the positioning speed and improve the positioning accuracy.

Description

Technical Field

[0001] This invention belongs to the technical field of moving target localization and tracking, specifically relating to a target localization method and system based on a distributed initial value constraint particle swarm algorithm. Background Technology

[0002] The widespread adoption of next-generation wireless networks has created a ubiquitous demand for location-based services. Determining the location of a moving target is crucial for identifying its location at a specific time or guiding it along a predetermined trajectory. The applications of location-based services are currently limited to our imagination, ranging from single-target tracking to multi-target collaboration, requiring a readily deployable, distributed, and error-free localization method. Thanks to the ubiquitous sensing and communication capabilities of wireless nodes, wireless sensor networks can achieve distributed information interaction between the logical and physical worlds. Wireless localization, whether utilizing inter-node signals, can be categorized into range-based localization and range-independent localization. Range-based wireless localization primarily involves solving nonlinear equations extracted from wireless signals; however, various measurement and environmental noises lead to unstable localization accuracy. Currently, researchers are improving wireless sensor network localization by enhancing range accuracy, refining deployment methods, and innovating localization algorithms. However, existing technologies have the following drawbacks:

[0003] (1) A single positioning algorithm has fluctuating accuracy throughout the positioning process. In particular, wireless distributed solution is prone to large positioning errors due to interference from uncertain measurements or even ill-conditioned matrices. It is also sensitive to noise. Therefore, it is necessary to combine wireless distributed solution with intelligent search algorithms, including particle swarm optimization.

[0004] (2) The classic particle swarm optimization algorithm searches the entire positioning space by randomly generated particles. It has problems such as slow convergence speed and long search time. At the same time, the particle search process is prone to getting trapped in local optima, which increases the positioning error of the target to a certain extent. Therefore, it is necessary to improve the wireless positioning based on particle swarm optimization. Summary of the Invention

[0005] Purpose of the invention: To address the problems existing in the prior art, this invention provides a target localization method based on a distributed initial value constraint particle swarm algorithm, which can accelerate the localization speed and improve the localization accuracy.

[0006] Technical solution: This invention discloses a target localization method based on a distributed initial value constrained particle swarm algorithm, comprising the following steps:

[0007] S1. Obtain the distance between each sensor and the moving target, and calculate the distance m between the moving target and the target. j Location estimate The sensor is deployed in the space to be located to communicate with the moving target and to obtain information about its own position relative to the moving target. j Distance between

[0008] S2. Determine the feasible region of the moving target based on wireless ranging, forming the search area S. PSO ;

[0009] S3, with search area S PSO To constrain the positions of particles in a particle swarm, the particle swarm optimization algorithm is used to obtain the position of the moving target.

[0010] Specifically, step S1 uses the least squares method to calculate the position value of the moving target, as follows:

[0011]

[0012] in a i Let be the coordinates of the i-th sensor in the space to be located, i∈[1,M], where M is the number of sensors; The superscript -1 indicates the matrix inversion operation, and the superscript T indicates the matrix or vector transpose operation.

[0013] Specifically, step S2 includes:

[0014] S21. Determine the spherical communication area for each sensor;

[0015] by Centered on the sphere, with sensor a i and Measured distance The spherical region with radius is sensor a. i A spherical communication area; in which This is the ranging error;

[0016] S22. The intersection of two spherical communication regions forms a sub-feasible region. Calculate the intersection between all pairwise spherical communication regions to obtain... Each feasible domain; The union of the individual feasible regions constitutes the total feasible region; the circumscribed cuboid of the total feasible region is the search region S. PSO ; This represents the number of combinations formed by choosing 2 elements from M elements.

[0017] Specifically, step S3 includes:

[0018] S31. Initialize a particle swarm of size n, where the position and velocity of particle p are both three-dimensional vectors. In the search space S... PSOThe positions of n particles are randomly initialized within the array, and the initial position of particle p is... Particle velocity V p Initialize to

[0019] Initialize the number of iterations k = 0, and set the maximum number of iterations k. max The value; the location is defined in X. p The fitness of the particle f(X) p (k) is:

[0020] Among them κ i For sensor a i Weighting coefficients:

[0021] Initialize the optimal position O of particle p p For X p The positions of n particles are used as parameters to calculate their fitness. The position of the particle with the smallest fitness value is selected as the initial global optimum position O. g ;

[0022] S32. Update the position and velocity of particle p in the particle swarm:

[0023]

[0024] X p (k+1)=X p (k)+V p (k+1)

[0025] Where w(k) is the inertia weight coefficient, w max and w min These are the maximum and minimum inertia weight coefficients, respectively, and α and β are the preset inertia weight coefficients minus the first and second parameters, respectively. and It is a random value;

[0026] If X p (k+1) Exceeds the search region S PSO Then calculate X p (k) Along vector V p (k+1) Direction and Search Area S PSO The intersection point X′ of the boundaries p (k) represents the updated position X of particle p. p (k+1);

[0027] After the update, recalculate the fitness of each particle. If f(X) p (k+1))<f(X p(k)), then O p The value is updated to X p (k+1); if f(X) p (k+1))<f(O g If the global optimal position O is determined, then the global optimal position O will be determined. g Updated to X p (k+1);

[0028] Increment the iteration count k by one;

[0029] S33. Repeat step S32 until the number of iterations k reaches k max Or the global optimal position O g If no update occurs after K consecutive iterations, the iteration ends; at this point, the global optimal position O is reached. g The value represents the position of the moving target.

[0030] On the other hand, the present invention also discloses a target localization system based on a distributed initial value constrained particle swarm algorithm, comprising:

[0031] Distance acquisition module 1 is used to acquire the distance between each sensor and the moving target, and to calculate the distance m between the moving target and the moving target. j Location estimate The sensor is deployed in the space to be located to communicate with the moving target and to obtain information about its own position relative to the moving target. j Distance between

[0032] Search area determination module 2 is used to determine the feasible region of the moving target based on wireless ranging, forming the search area S. PSO ;

[0033] The moving target positioning module 3 is used to locate the target in the search area S. PSO To constrain the positions of particles in a particle swarm, the particle swarm optimization algorithm is used to obtain the position of the moving target.

[0034] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described target positioning method.

[0035] The present invention also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described target localization method.

[0036] Beneficial effects: The target localization method and system based on distributed initial value constrained particle swarm algorithm disclosed in this invention have the following advantages:

[0037] 1. The results of distributed computation are used as the initial positions of particles. The search range is narrowed by using the feasible solution space identified by wireless ranging intersection, thereby improving the positioning accuracy and convergence speed of the classical particle swarm algorithm.

[0038] 2. The choice of fitness function determines whether better guidance can be generated. When the fitness function reaches its minimum value, the optimal coordinates of the moving node are updated. Considering the distance measurements from sensors with different weight coefficients, the fitness function is designed based on the difference between the actual distance and the estimated distance.

[0039] 3. To improve search accuracy and convergence speed, the inertia weight coefficient also needs to be dynamically adjusted during the particle search process. A larger inertia weight is used to increase the convergence speed in the early stages, while a gradually decreasing inertia weight ensures convergence to a stable state, thereby improving search accuracy. Attached Figure Description

[0040] Figure 1 This is a flowchart of the target localization method based on the distributed initial value constraint particle swarm algorithm disclosed in this invention;

[0041] Figure 2 This is a schematic diagram of the sensor deployment in the embodiment;

[0042] Figure 3 This is a flowchart illustrating the process of obtaining the position of a moving target using the particle swarm optimization algorithm.

[0043] Figure 4 This is a schematic diagram of the target localization system based on the distributed initial value constraint particle swarm algorithm disclosed in this invention.

[0044] Figure 5 This is a schematic diagram of the composition of the computer device disclosed in this invention. Detailed Implementation

[0045] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0046] This invention discloses a target localization method based on a distributed initial value constrained particle swarm algorithm, such as... Figure 1 As shown, the steps include:

[0047] S1. Obtain the distance between each sensor and the moving target, and calculate the distance m between the moving target and the target. j Location estimate The sensor is deployed in the space to be located to communicate with the moving target and to obtain information about its own position relative to the moving target. j Distance between

[0048] The position value of the moving target is calculated using the least squares method, specifically as follows:

[0049]

[0050] in a i Let be the coordinates of the i-th sensor in the space to be located, i∈[1,M], where M is the number of sensors; The superscript -1 indicates the matrix inversion operation, and the superscript T indicates the matrix or vector transpose operation.

[0051] like Figure 2 As shown, in this embodiment, eight sensors (a1, a2, a3, a4, a5, a6, a7, and a8) are deployed at the eight corners of the three-dimensional space to be located. The position of each sensor is known. Moving target m j The trajectory of the moving target is shown as the curve with arrows in the diagram. During the movement of the moving target, each sensor can obtain its own distance from the moving target by transmitting wireless signals and receiving feedback signals from the moving target.

[0052] S2. Determine the feasible region of the moving target based on wireless ranging, forming the search area S. PSO Specifically, it includes:

[0053] S21. Determine the spherical communication area for each sensor;

[0054] by Centered on the sphere, with sensor a i and Measured distance The spherical region with radius is sensor a. i A spherical communication area; in which This is the ranging error, which is typically Gaussian-distributed noise.

[0055] S22. The intersection of two spherical communication regions forms a sub-feasible region. Calculate the intersection between all pairwise spherical communication regions to obtain... Each feasible domain; The union of the individual feasible regions constitutes the total feasible region; the circumscribed cuboid of the total feasible region is the search region S. PSO ; This represents the number of combinations formed by choosing 2 elements from M elements.

[0056] The intersection of two spherical regions forms a polyhedron. In this embodiment, the pairwise intersections of the eight spherical communication regions corresponding to the eight sensors result in 28 polyhedra, thus yielding 28 sub-feasible regions. The union of these sub-feasible regions constitutes the total feasible region, and its circumscribed cuboid is the search region S. PSO .

[0057] S3, with search area S PSO To constrain the positions of particles in a particle swarm, the particle swarm optimization algorithm is used to obtain the position of the moving target, such as... Figure 3 As shown, it specifically includes:

[0058] S31. Initialize a particle swarm of size n, where the position and velocity of particle p are both three-dimensional vectors. In the search space S... PSO The positions of n particles are randomly initialized within the array, and the initial position of particle p is... Particle velocity V p Initialize to The initial value of the particle velocity can be set to 0 or a random value. In this embodiment, the particle velocity is initialized to a random value.

[0059] Initialize the number of iterations k = 0, and set the maximum number of iterations k. max The value; the location is defined in X. p The fitness of the particle f(X) p (k) is:

[0060] Among them κ i For sensor a i Weighting coefficients:

[0061] In this embodiment, the particle fitness is designed based on the difference between the estimated and measured distance between the sensor and the moving target. During the iteration process, the estimated value is continuously made to approximate the measured value, thereby obtaining the optimal estimate. Furthermore, different sensors have different weights in the fitness calculation; the closer the sensor is to the moving target and the more reliable the communication with the moving target, the greater its weight.

[0062] Initialize the optimal position O of particle p p For X p The positions of n particles are used as parameters to calculate their fitness. The position of the particle with the smallest fitness value is selected as the initial global optimum position O. g ;

[0063] S32. Update the position and velocity of particle p in the particle swarm:

[0064]

[0065] X p (k+1)=X p (k)+V p (k+1)

[0066] Where w(k) is the inertia weight coefficient, w max and w minThese are the maximum and minimum inertia weight coefficients, respectively. α and β are the first and second parameters for reducing the preset inertia weight coefficient, respectively. In this embodiment, α = 0.2 and β = 10. When the iteration number k increases from 0 to k... max At that time, the inertia weighting coefficient w(k) changes from w max Reduce to w min That is, it decreases as the number of iterations increases. A larger w(k) in the early stage of iteration is beneficial to speeding up the convergence speed, while a smaller w(k) in the later stage of iteration is beneficial to improving the convergence accuracy. and It is a random value; in this embodiment, it will be and The value is restricted to the range [0, c], that is... and A random number within the range [0, c]; where c > 0 is the preset upper limit for random numbers;

[0067] If X p (k+1) Exceeds the search region S PSO Then calculate X p (k) Along vector V p (k+1) Direction and Search Area S PSO The intersection point X′ of the boundaries p (k) represents the updated position X of particle p. p (k+1); This constrains the positions of the particles before and after the update within the search region S. PSO Inside or boundary.

[0068] After the update, recalculate the fitness of each particle. If f(X) p (k+1))<f(X p (k)), then the optimal position O of particle p is... p Updated to X p (k+1); if f(X) p (k+1))<f(O g If the global optimal position O is determined, then the global optimal position O will be determined. g Updated to X p (k+1);

[0069] Increment the iteration count k by one;

[0070] S33. Repeat step S32 until the number of iterations k reaches k max Or the global optimal position O g If no update occurs after K consecutive iterations, the iteration ends; at this point, the global optimal position O is reached. g The value represents the position of the moving target.

[0071] This invention also discloses a system for implementing the above-mentioned target localization method based on the distributed initial value constraint particle swarm algorithm, such as... Figure 4 As shown, it includes:

[0072] Distance acquisition module 1 is used to acquire the distance between each sensor and the moving target, and to calculate the distance m between the moving target and the moving target. j Location estimate The sensor is deployed in the space to be located to communicate with the moving target and to obtain information about its own position relative to the moving target. j Distance between

[0073] Search area determination module 2 is used to determine the feasible region of the moving target based on wireless ranging, forming the search area S. PSO ;

[0074] The moving target positioning module 3 is used to locate the target in the search area S. PSO To constrain the positions of particles in a particle swarm, the particle swarm optimization algorithm is used to obtain the position of the moving target.

[0075] The distance measurement module 1 uses the least squares method to calculate the position value of the moving target, specifically:

[0076]

[0077] in a i Let be the coordinates of the i-th sensor in the space to be located, i∈[1,M], where M is the number of sensors; The superscript -1 indicates the matrix inversion operation, and the superscript T indicates the matrix or vector transpose operation.

[0078] The search area determination module 2 determines the search area S using the following steps. PSO :

[0079] S21. Determine the spherical communication area for each sensor;

[0080] by Centered on the sphere, with sensor a i and Measured distance The spherical region with radius is sensor a. i A spherical communication area; in which This is the ranging error;

[0081] S22. The intersection of two spherical communication regions forms a sub-feasible region. Calculate the intersection between all pairwise spherical communication regions to obtain... Each feasible domain; The union of the individual feasible regions constitutes the total feasible region; the circumscribed cuboid of the total feasible region is the search region S. PSO ; This represents the number of combinations formed by choosing 2 elements from M elements.

[0082] The moving target localization module 3 obtains the position of the moving target through the following steps:

[0083] S31. Initialize a particle swarm of size n, where the position and velocity of particle p are both three-dimensional vectors. In the search space S... PSO The positions of n particles are randomly initialized within the array, and the initial position of particle p is... Particle velocity V p Initialize to

[0084] Initialize the number of iterations k = 0, and set the maximum number of iterations k. max The value; the location is defined in X. p The fitness of the particle f(X) p (k) is:

[0085] Among them κ i For sensor a i Weighting coefficients:

[0086] Initialize the optimal position O of particle p p For X p The positions of n particles are used as parameters to calculate their fitness. The position of the particle with the smallest fitness value is selected as the initial global optimum position O. g ;

[0087] S32. Update the position and velocity of particle p in the particle swarm:

[0088]

[0089] X p (k+1)=X p (k)+V p (k+1)

[0090] Where w(k) is the inertia weight coefficient, w max and w min These are the maximum and minimum inertia weight coefficients, respectively, and α and β are the preset inertia weight coefficients minus the first and second parameters, respectively. and It is a random value;

[0091] If X p (k+1) Exceeds the search region SPSO Then calculate X p (k) Along vector V p (k+1) Direction and Search Area S PSO The intersection point X′ of the boundaries p (k) represents the updated position X of particle p. p (k+1); This limits the position of the particles before and after the update within the search region S. PSO Inside or boundary.

[0092] After the update, recalculate the fitness of each particle. If f(X) p (k+1))<f(X p (k)), then O p The value is updated to X p (k+1); if f(X) p (k+1))<f(O g If the global optimal position O is determined, then the global optimal position O will be determined. g Updated to X p (k+1);

[0093] Increment the iteration count k by one;

[0094] S33. Repeat step S32 until the number of iterations k reaches k max Or the global optimal position O g If no update occurs after K consecutive iterations, the iteration ends; at this point, the global optimal position O is reached. g The value represents the position of the moving target.

[0095] The present invention also discloses a computer device, such as Figure 5 As shown, it includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described target localization method.

[0096] The present invention also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described target localization method.

[0097] This invention discloses an improved particle swarm optimization (PSO) localization method with distributed initial value constraints. Considering the drawback of randomly generated particles in PSO, this method uses wireless distributed computation as the initial search value to generate particles. To improve the localization accuracy and convergence speed of the PSO algorithm, the search range is narrowed by utilizing the feasible solution space identified through wireless ranging intersection. Considering the ranging values ​​between nodes with different weight coefficients, a fitness function is designed based on the difference between the actual distance and the estimated distance. To improve search accuracy and convergence speed, the inertia weight coefficient is dynamically adjusted during the particle search process. This invention effectively combines the advantages of wireless distributed computation and PSO, improving the performance of wireless target localization.

[0098] The above description outlines the basic principles, main features, and advantages of this invention. Those skilled in the art should understand that this invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to this invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A target positioning method based on a distributed initial value constraint particle swarm optimization algorithm, characterized in that, Including the following steps: S1. Obtain the distance between each sensor and the moving target, and calculate the moving target distance. Location estimate The sensor is deployed in the space to be located, and is used to communicate with the moving target and obtain information about itself and the moving target. Distance between ; S2, determining the feasible region of the moving target based on wireless ranging, forming a search region ; S3, with search area The position of the mobile target is obtained by using the particle swarm algorithm as the position constraint of the particles in the particle swarm. Step S2 specifically includes: S21. Determine the spherical communication area for each sensor; by Centered on the ball, with the sensor and Measured distance A spherical region with radius is the sensor A spherical communication area; in which , This is the ranging error; S22. The intersection of two spherical communication regions forms a sub-feasible region. Calculate the intersection between all pairwise spherical communication regions to obtain... Each feasible domain; The union of the individual feasible regions constitutes the total feasible region; the circumscribed cuboid of the total feasible region is the search region. ; From The number of combinations of two elements chosen from a set of elements; Step S3 specifically includes: S31. Initialize the population size as follows: Particle swarm, particles Position and velocity are both three-dimensional vectors in the search space. Internal random initialization The position of each particle, the particle The initial position is Particle velocity Initialize to ; Initialize the number of iterations And set the maximum number of iterations. The value; the location is defined in Particle fitness for: ; wherein is a weight coefficient of the sensor ; and ; Initialize particles optimal position for ;Will The positions of each particle are used as parameters to calculate the fitness, and the position of the particle with the smallest fitness value is selected as the initial global optimum position. ; S32, update the position and velocity of the particles in the particle group ​ in This is the inertia weighting coefficient. , and These are the maximum inertia weight coefficient and the minimum inertia weight coefficient, respectively. The first and second parameters are reduced respectively from the preset inertia weight coefficients; and It is a random value; If the search area is exceeded , then calculate the intersection of the direction of the vector with the boundary of the search area and the new position of the particle as the intersection point updated position of the particle ; After the update, recalculate the fitness of each particle, if , update the value of to ; if , update the global optimal position to ; number of iterations plus one; S33, repeat step S32 until the iteration number reaches , or the global optimal position is continuously updated for the next iteration, end the iteration; at this time, the value of the global optimal position is the position of the moving target.

2. The object positioning method according to claim 1, characterized in that, Step S1 uses the least squares method to calculate the position value of the moving target, specifically as follows: ； in , For the first in the space to be located The coordinates of each sensor, , The number of sensors; , ; superscript The superscript indicates the matrix inversion operation. This represents the transpose operation of a matrix or vector.

3. A target localization system based on a distributed initial value constrained particle swarm algorithm, characterized in that, include: The distance acquisition module (1) is used to acquire the distance between each sensor and the moving target, and to calculate the distance between the moving target and the moving target. Location estimate The sensor is deployed in the space to be located, and is used to communicate with the moving target and obtain information about itself and the moving target. Distance between ; The search area determination module (2) is used to determine the feasible region of the moving target based on wireless ranging, and form the search area. ; The moving target localization module (3) is used to locate the target in the search area. To constrain the positions of particles in a particle swarm, the particle swarm optimization algorithm is used to obtain the position of the moving target. The search area determination module (2) determines the search area using the following steps. : S21. Determine the spherical communication area for each sensor; by Centered on the ball, with the sensor and Measured distance A spherical region with radius is the sensor A spherical communication area; in which , This is the ranging error; S22. The intersection of two spherical communication regions forms a sub-feasible region. Calculate the intersection between all pairwise spherical communication regions to obtain... Each feasible domain; The union of the individual feasible regions constitutes the total feasible region; the circumscribed cuboid of the total feasible region is the search region. ; From The number of combinations of choosing two elements from a set of elements; The moving target positioning module (3) obtains the position of the moving target through the following steps: S31. Initialize the population size as follows: Particle swarm, particles Position and velocity are both three-dimensional vectors in the search space. Internal random initialization The position of each particle, the particle The initial position is Particle velocity Initialize to ; Initialize the number of iterations And set the maximum number of iterations. The value; the location is defined in Particle fitness for: ; in For sensors Weighting coefficients: ; Initialize particles optimal position for ;Will The positions of each particle are used as parameters to calculate the fitness, and the position of the particle with the smallest fitness value is selected as the initial global optimum position. ; S32, Update particles in the particle swarm Position and velocity: in This is the inertia weighting coefficient. , and These are the maximum inertia weight coefficient and the minimum inertia weight coefficient, respectively. The first and second parameters are reduced respectively from the preset inertia weight coefficients; and It is a random value; if Beyond search area Then calculate Along vector Direction and search area intersection of boundaries As particles Updated location ; After the update is complete, recalculate the fitness of each particle. Then The value is updated to ;if Then the global optimal position will be Updated to ; Number of iterations Add one; S33. Repeat step S32 until the number of iterations is reached. achieve or the global optimal position continuous If no update occurs in any of the next iterations, the iteration ends; at this point, the global optimal position is... The value represents the position of the moving target.

4. The target positioning system according to claim 3, characterized in that, The distance measurement module (1) calculates the position value of the moving target using the least squares method, specifically: ； in , For the first in the space to be located The coordinates of each sensor, , The number of sensors; , ; superscript The superscript indicates the matrix inversion operation. This represents the transpose operation of a matrix or vector.

5. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the target localization method as described in any one of claims 1-2.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the target localization method as described in any one of claims 1-2.

Images