A hierarchical weighted translation DV-HOP node localization method based on RSSI correction

Abstract

This paper proposes a hierarchical weighted translation DV-HOP node positioning method based on RSSI correction. This method uses self-positioning of anchor nodes to derive the optimal error estimate for the entire network. It then performs a step-by-step positioning of unknown nodes across the entire network. For the initial positioning results, error circles and RSSI circles are constructed, and translation vectors and translation weights are calculated. Weighted translation corrections are then applied to the initial positioning coordinates. This method is well-suited for random networks, particularly anisotropic networks, achieving significantly improved positioning accuracy compared to the classic DV-HOP positioning algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless sensor networks, and in particular to a hierarchical weighted translation DV-HOP node positioning method based on RSSI correction. Background Art

[0002] In recent years, with the rapid development of the Internet of Things (IoT) and the widespread application of various new sensor devices, wireless sensor networks have been widely used in smart homes, industrial control, smart ranches, environmental monitoring, military confrontation, and other fields. The effectiveness of the various environmental and image information monitored by wireless sensor networks requires the support of location information.

[0003] Wireless sensor network node localization technologies can be categorized as ranging and non-ranging based on the ranging method. Ranging-based positioning methods primarily include time of arrival (TOA), time difference of arrival (TDOA), angle of arrival (AOA), and received signal strength indicator (RSSI). Common non-ranging positioning algorithms include the APIT algorithm, the centroid algorithm, the MDS-MAP algorithm, and the classic DV-HOP positioning algorithm. These non-ranging positioning algorithms require no additional hardware and outperform ranging positioning methods in dense networks and complex environments. Furthermore, their low energy consumption and ease of deployment meet the operational requirements of wireless sensor networks, making them suitable for resource-constrained wireless sensor networks. The classic DV-HOP positioning algorithm, the most widely used non-ranging positioning algorithm, offers superior positioning results even with a limited number of anchor nodes, making it a highly efficient and cost-effective positioning method.

[0004] The classic DV-HOP localization algorithm uses the hop count information obtained during the flooding phase and the known coordinates of anchor nodes to calculate the average hop distance. The unbiased average hop distance estimated by neighboring anchor nodes is used as the average hop distance for locating unknown nodes. The estimated distance from the unknown node to the anchor node is then calculated. When the number of anchor nodes involved in localization exceeds two, a least-squares method is used to solve for the unknown node coordinates using a multilateration equation. During this process, the minimum hop path between anchor nodes is inconsistent with the minimum hop path from the location node to the anchor node. As a result, the average hop distance of the anchor nodes does not reflect the average hop distance of the unknown nodes. This error is amplified in anisotropic networks.

[0005] Currently, most optimization schemes for the classic DV-HOP localization algorithm focus on correcting the average hop distance of anchor nodes. In this process, most improvement methods correct the average hop distance of anchor nodes as a whole, without considering the differences in the average hop distances of a single anchor node and other anchor nodes in different directions. Due to the anisotropy of the network topology of wireless sensor networks, some anchor nodes may cause large errors when participating in the average hop distance calculation. In addition, when the unknown node is far away from the nearest anchor node, the average hop distance of the anchor node has little correlation with the average hop distance of the unknown node. At this time, using the average hop distance of the anchor node as the average hop distance of the unknown node to calculate the estimated distance will lead to error accumulation. Summary of the Invention

[0006] Aiming at the problems of low positioning accuracy of the classic DV-HOP positioning algorithm and the significant influence of the anisotropy of the minimum hop path in the network on the average hop distance calculation, the present invention proposes a hierarchical weighted translation DV-HOP node positioning method based on RSSI correction. First, the anchor node is self-positioned and the best estimation error of the entire network is calculated using the quartile statistical method. Then, the unknown nodes in the entire network are graded and a step-by-step positioning method starting from the first-level node to the last-level node is adopted. The average hop distance estimated by the non-unbiased neighboring nodes is used as the average hop distance of the corresponding unknown node. After the initial positioning of each level node is completed, the RSSI estimated distance and the best error estimate of the entire network are used to construct RSSI circles and best estimation error circles around the initial positioning coordinates of the neighboring nodes and the corresponding unknown nodes respectively. The end point coordinates are obtained by the intersection of the two circles and the translation vector is calculated. The translation weight is calculated by the maximum level of the unknown node and the current level of the unknown node. Finally, the initial positioning coordinates of the unknown node are weighted translated to obtain the corrected coordinates of the unknown node.

[0007] In order to achieve the above object, the technical solution of the present invention is implemented as follows: a hierarchical weighted translation DV-HOP node positioning method based on RSSI correction is characterized by comprising the following steps:

[0008] S1: The classic DV-HOP positioning algorithm is used to self-locate the anchor nodes with known positions in the network. The anchor node self-localization error is calculated based on the estimated position and the actual position. The quartile method is used to statistically analyze the distribution of the anchor node self-localization error and obtain the best estimated error.

[0009] S2: Classify the unknown nodes in the entire network according to the minimum hop count array from the unknown node to the nearest anchor node after flooding, and mark the neighbor nodes of each unknown node.

[0010] S3: Positioning unknown nodes step by step: Starting from the first-level nodes, all unknown nodes in the entire network are positioned step by step to the last-level nodes. The positioning process for each level of nodes is divided into initial positioning and initial positioning correction. Initial positioning: Use the average hop distance of the non-unbiased estimate of neighboring nodes as the average hop distance of the corresponding unknown node, calculate the estimated distance from the corresponding unknown node to all anchor nodes, and use the standard least squares method to solve the initial positioning coordinates of the corresponding unknown node. Initial positioning correction: Construct the RSSI circle and the best estimate error circle, calculate the translation vector and translation weight, and perform a weighted translation of the initial positioning coordinates along the direction of the specific correction vector to obtain the corrected coordinates.

[0011] In a preferred solution, the calculation process of the best estimation error in S1 is:

[0012] (1) Anchor node self-positioning error calculation:

[0013]

[0014] Among them, Error j is the self-positioning error of the j-th anchor node, (X o , Y o ) is the estimated position coordinate of the j-th anchor node calculated by the classic DV-HOP positioning algorithm, (X j , Y j ) is the coordinate of the j-th anchor node.

[0015] (2) Calculate the best estimate error:

[0016]

[0017] Among them, Error Q2 is the best estimation error, q is the number of anchor nodes, and the anchor node self-localization error sample dataset X = {Error1, Error2, ..., Error j ,...,Error q}, sort the data from small to large, and get the ordered sequence X′={m1,m2,...,m k ,...,m q}, where m k Error is the self-positioning error of the anchor node at the kth position after sorting. Q2 Take the median of the permutation.

[0018] In a preferred solution, the unknown node classification and neighbor node marking method in S2 is:

[0019] (1) Unknown node classification method: mark the unknown node that is one hop away from the anchor node as a first-level node, the unknown node that is two hops away from the anchor node as a second-level node, and so on.

[0020] (2) Neighbor node marking method: The anchor node with the smallest RSSI estimated distance between the marker and the first-level node is the neighbor node of the corresponding first-level node; the first-level node with the smallest RSSI estimated distance between the marker and the second-level node is the neighbor node of the corresponding second-level node, and so on.

[0021] In a preferred solution, the RSSI estimated distance in S32 is calculated as follows:

[0022]

[0023] in, is at a distance d est The received signal strength at is the signal strength received at the reference distance d0, σ is the path loss exponent, d0 is the reference distance, d est is the distance between the receiver and the transmitter, which is the RSSI estimated distance, and Xσ represents the blocking factor.

[0024] In a preferred solution, the initial positioning steps for each level of unknown nodes in S3 are:

[0025] (1) Calculate the average hop distance of the neighboring nodes in the non-unbiased estimate:

[0026]

[0027] Among them, i is the number of unknown nodes, n′ represents the neighbor node of the current nth unknown node, and Dhop n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s average hop distance, DIS n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s Euclidean distance, HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j ), (X j , Y j ) is the coordinate of the jth anchor node, (X n′ , Y n′ ) is the coordinate of the neighbor node n′ of the current nth unknown node.

[0028] (2) Calculation of the distance from unknown nodes to anchor nodes:

[0029] d nj =DhoP n′j ·HM nj ;

[0030] Among them, d nj is the estimated distance from unknown node n to the jth anchor node, HM nj is the minimum number of hops from unknown node n to the jth anchor node.

[0031] (3) Construct the matrix form of the difference equation:

[0032]

[0033] Among them, the matrix P is the design matrix, which contains the anchor node coordinates and differential relationship, and the matrix H is the constant matrix, which contains the square difference of the anchor node distance and the difference of the sum of the square of the anchor node coordinates; C j =(X j , Y j ) is the coordinate of the j-th anchor node, represents the sum of squares of the j-th anchor node coordinates.

[0034] (4) Use the least squares method to solve the initial positioning coordinates of the unknown nodes:

[0035] X=(P T P) -1 P T H;

[0036] Where X is the initial positioning coordinate of the unknown node.

[0037] In a preferred solution, the initial positioning correction steps for each level of unknown nodes in S3 are:

[0038] (1) Construct RSSI circle and best estimation error circle:

[0039]

[0040] Among them, the coordinate system is established with x as the horizontal coordinate and y as the vertical coordinate. oe , Y oe ) is the initial positioning coordinate of the unknown node, Error Q2 is the best estimation error, (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, R min The RSSI estimated distance between the current unknown node and its neighboring nodes.

[0041] (2) Intersection relationship between RSSI circle and best estimation error circle and calculation of end point coordinates:

[0042]

[0043] Among them, (x t ,y t ) is the end point coordinate, (x s1 ,y s1 ), (x s2 ,y s2 ) is the coordinate of the intersection point where the RSSI circle intersects with the best estimation error circle. When the RSSI circle intersects with the best estimation error circle, (x t ,y t ) has two solutions. By using the regional frequency analysis method, (x s1 ,y s1 ), (x s2 ,y s2 ) selects the end point coordinates; t is the proportional coefficient, and D is the distance between the center of the RSSI circle and the best estimation error circle. n′ +t(X oe -X n′ ), Y n′ +t(Y oe -Y n′ )) is the coordinate of the intersection of the RSSI circle and the center point of the line segment passing through the RSSI circle and the best estimation error circle when the two circles do not intersect.

[0044] (3) Translation vector calculation:

[0045]

[0046] in, represents the translation vector, It is a vector pointing from the initial coordinates of the current unknown node to the coordinates of the end point.

[0047] (4) Translation weight calculation:

[0048]

[0049] Among them, w(γ) is the translation weight, HM max is the maximum level of the current network, w min is the minimum translation weight of the current network, and γ is the unknown node level of the translation operation performed this time.

[0050] (5) Estimated position coordinates of unknown nodes after weighted translation correction:

[0051]

[0052] Among them, (x new ,y new ) is the estimated position coordinate of the unknown node after weighted translation correction.

[0053] In a preferred embodiment, the regional frequency analysis method selects the end point coordinate (x t ,y t ) process is:

[0054] (1) Establish a coordinate system with the neighbor node of the current nth unknown node as the zero point:

[0055]

[0056] Among them, (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, (X j , Y j ) is the jth anchor node coordinate, for each anchor node (X j , Y j ), confirmed (X j , Y j ) relative to the neighboring node (X n′ , Y n′ ) is located in the quadrant Q n′ .

[0057] (2) Update the distribution probability function:

[0058]

[0059] in, The nth unknown node is located at the Qth n′ Quadrant probability value, initial value is 0, diag(Q n′ ) is Q n′ diagonal quadrant of HM nj For the current unknown node (X n , Y n ) to the anchor node (X j , Y j ), HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j ) minimum hop value.

[0060] (3) Calculate the point (x s1 ,y s1 ), (x s2 ,y s2 ) quadrant distribution, according to the distribution probability function, the point in the high probability quadrant is taken as the end point coordinate (x t ,y t ).

[0061] Compared with the existing technology, the present invention has the following beneficial effects: the present invention uses the quartile statistical method to infer the optimal error estimate of the unknown nodes in the entire network through anchor node self-positioning, thereby improving the stability of the algorithm in random networks; in addition, the weighted translation correction of the initial positioning of unknown nodes is a dynamic correction process based on different network topologies, which can dynamically correct the results of the initial positioning of unknown nodes at all levels according to the network node distribution and the overall error distribution, greatly improving the node positioning accuracy and the robustness of the algorithm. The algorithm of the present invention uses a step-by-step positioning method that uses the positioning result of the previous node as the benchmark for the initial positioning and weighted translation process of the next node, thereby improving the error accumulation problem caused by the huge difference between the average hop distance of the anchor node and the average hop distance of the unknown node when the unknown node is multiple hops away from the anchor node, and significantly improving the positioning accuracy of the edge nodes. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 It is a flow chart of the present invention.

[0064] Figure 2 This is a correction effect diagram of the algorithm of the present invention based on the classic DV-HOP positioning algorithm.

[0065] Figure 3 This is a comparison chart of the positioning accuracy of the algorithm of the present invention and the classic DV-HOP positioning algorithm under 100 random networks.

[0066] Figure 4 This is a comparison chart of the positioning accuracy of the algorithm of the present invention and the classic DV-HOP positioning algorithm under different anchor node ratios.

[0067] Figure 5 This is a comparison chart of the positioning accuracy of the algorithm of the present invention and the classic DV-HOP positioning algorithm under different communication radius.

[0068] Figure 6 This is a comparison chart of the positioning accuracy of the algorithm of the present invention and the classic DV-HOP positioning algorithm at different summary points. DETAILED DESCRIPTION

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0070] A hierarchical weighted translation DV-HOP node localization method based on RSSI correction is proposed. The concept is as follows: first, anchor nodes are self-localized. The anchor node error dataset is processed using quartile statistics to estimate the network-wide best estimate error. Unknown nodes in the network are then classified according to the minimum number of hops from the unknown node to the nearest anchor node. Each node level is located stepwise from the first to the highest level. The average hop distance estimated by the non-unbiased neighboring nodes is used as the average hop distance for the corresponding unknown node to calculate the initial location coordinates of the unknown node. Next, the RSSI estimated distance is calculated using a logarithmic decay model. An RSSI circle is constructed with the neighboring node as the center and the RSSI estimated distance as the radius. A best estimate error circle is constructed with the corresponding unknown node as the center and the network-wide best estimate error as the radius. The destination coordinates are calculated based on the intersection of the two circles. Finally, a translation vector is calculated from the initial and destination coordinates of the unknown node. The translation weight is calculated using the maximum level in the network and the level to which the current unknown node belongs. The initial location coordinates of the unknown node are then weighted and translated to obtain the corrected coordinates of the unknown node.

[0071] like Figure 1 As shown in the figure: After the anchor node is self-positioned, the present invention obtains the best estimated error for the entire network, calculates the RSSI estimated distance between nodes within the communication radius and marks the neighboring nodes, and then locates the unknown nodes step by step from the first level to the highest level. In this process, the unknown node is initially positioned, and then the translation vector and translation weight are calculated by constructing the RSSI circle and error circle to complete the weighted translation correction of the initial positioning coordinates of the unknown node. The steps are as follows:

[0072] S1: By self-locating the anchor nodes at known locations in the network, the positioning error is calculated by comparing the estimated position with the actual position, and the best estimated error of the entire network is calculated using the quartile statistical method. The method for calculating the best estimated error of the entire network using the quartile method is:

[0073] S11: Anchor node self-positioning error calculation:

[0074]

[0075] Among them, Error j is the self-positioning error of the j-th anchor node, (X o , Y o ) is the estimated position coordinate of the j-th anchor node calculated by the classic DV-HOP positioning algorithm, (Xj , Y j ) is the coordinate of the j-th anchor node.

[0076] S12: Calculate the best estimate error:

[0077]

[0078] Among them, Error Q2 is the best estimation error, q is the number of anchor nodes, and the anchor node self-localization error sample data set X = {Error1, Error2, ..., Error j ,...,Error q}, sort the data from small to large, and get the ordered sequence X′={m1,m2,...,m k ,...,m q}, where m k Error is the self-positioning error of the anchor node at the kth position after sorting. Q2 Take the median of the permutation.

[0079] In real-world situations, the error distribution of the entire network varies greatly in different network topologies and environments. This step begins by self-locating the anchor nodes and counting the positioning errors. The positioning errors of unknown nodes in the entire network are inferred from the anchor node error distribution, providing a more accurate reference for subsequent position corrections after the initial positioning of unknown nodes, and improving the stability of the algorithm in random networks.

[0080] S2: Classify the unknown nodes in the entire network according to the minimum number of hops from the unknown node to the nearest anchor node after flooding: mark the unknown node with a distance of one hop from the anchor node as a first-level node, the unknown node with a distance of two hops from the anchor node as a second-level node, and so on.

[0081] The unknown nodes in the entire network are classified according to the minimum hop count information, and the unknown nodes closest to the anchor node are given the highest priority in positioning order. This can fully utilize the accuracy of the anchor node location information, greatly improving the positioning accuracy of the first-level nodes. Under the premise of high positioning accuracy of the first-level nodes, the corrected first-level nodes can be used as neighbor nodes of the second-level nodes to correct the positioning results of the second-level nodes, thereby improving the positioning accuracy of subsequent nodes.

[0082] S3: Calculate the RSSI distance between nodes within the communication radius using the logarithmic attenuation model and mark neighbor nodes:

[0083]

[0084] in, is at a distance d est The received signal strength at is the signal strength received at the reference distance d0, σ is the path loss exponent, d0 is the reference distance, d est is the distance between the receiver and the transmitter, which is the RSSI estimated distance, X σ Represents the blocking factor. The anchor node with the smallest RSSI estimated distance between the marker and the primary node is the neighbor node of the corresponding primary node. The primary node with the smallest RSSI estimated distance between the marker and the secondary node is the neighbor node of the corresponding secondary node, and so on.

[0085] Using RSSI information for distance estimation eliminates the need for additional hardware. RSSI values ​​provide a direct distance measurement method, providing a reliable distance estimate between neighboring nodes. Distance estimates between neighboring nodes are crucial for subsequent position corrections of unknown nodes. Neighbor nodes are selected based on the minimum RSSI estimated distance between nodes. This improves the accuracy of using the average hop distance of neighboring nodes as the average hop distance for unknown nodes, compared to the classic DV-HOP approach, which uses the minimum number of hops from an unknown node to an anchor node to select neighboring nodes.

[0086] S4: Position unknown nodes step by step: Position unknown nodes in the entire network starting from the first-level nodes to the last-level nodes step by step, and perform initial positioning and initial positioning correction on each level of nodes. The initial positioning process of each level of nodes is as follows:

[0087] S41: Calculate the average hop distance of the neighboring nodes:

[0088]

[0089] Among them, i is the number of unknown nodes, n′ represents the neighbor node of the current nth unknown node, and Dhop n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s average hop distance, DIS n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s Euclidean distance, HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , X j ), (X j , Y j ) is the coordinate of the jth anchor node, (X n′ , Y n′) is the coordinate of the neighbor node n′ of the current nth unknown node.

[0090] S42: Calculate the estimated distance from the unknown node to the anchor node:

[0091] d nj =Dhop n′j ·HM nj ;

[0092] Among them, d nj is the estimated distance from unknown node n to the jth anchor node, HM nj is the minimum number of hops from unknown node n to the jth anchor node.

[0093] S43: Construct the matrix form of the difference equation:

[0094]

[0095] Among them, the matrix P is the design matrix, which contains the anchor node coordinates and differential relationship, and the matrix H is the constant matrix, which contains the square difference of the anchor node distance and the difference of the sum of the square of the anchor node coordinates; C j =(X j , Y j ) is the coordinate of the j-th anchor node, represents the sum of squares of the j-th anchor node coordinates.

[0096] S44: Use the least squares method to solve the initial positioning coordinates of the unknown nodes:

[0097] X=(P T P) -1 P T H;

[0098] Where X is the initial positioning coordinate of the unknown node.

[0099] The advantage of this step is that it uses the non-unbiased estimated average hop distance of neighboring nodes as the average hop distance of the corresponding unknown node. It effectively solves the problem of error accumulation caused by the huge difference between the average hop distance of edge nodes and the average hop distance of anchor nodes when simply using the non-unbiased estimated average hop distance, and significantly improves the positioning accuracy of edge nodes.

[0100] S5: The initial positioning correction process of each level of unknown nodes is:

[0101] S51: Construct RSSI circle and best estimation error circle:

[0102]

[0103] Among them, the coordinate system is established with x as the horizontal coordinate and y as the vertical coordinate. oe , Y oe) is the initial positioning coordinate of the unknown node, Error Q2 is the best estimation error, (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, R min The RSSI estimated distance between the current unknown node and its neighboring nodes.

[0104] S52: Intersection relationship between RSSI circle and best estimation error circle and calculation of end point coordinates:

[0105]

[0106] Among them, (x t ,y t ) is the end point coordinate, (x s1 ,y s1 ), (x s2 ,y s2 ) is the coordinate of the intersection point where the RSSI circle intersects with the best estimation error circle. When the RSSI circle intersects with the best estimation error circle, (x t ,y t ) has two solutions. By using the regional frequency analysis method, (x s1 ,y s1 ), (x s2 ,y s2 ) selects the end point coordinates; t is the proportional coefficient, and D is the distance between the center of the RSSI circle and the best estimation error circle. n′ +t(X oe -X n′ ), Y n′ +t(Y oe -Y n′ )) is the intersection coordinate of the line segment passing through the RSSI circle and the center of the best estimated error circle with the RSSI circle when the two circles do not intersect. When the RSSI circle intersects the error circle, it is necessary to t ,y t ) to find the correct end point coordinates from the two solutions using the regional frequency analysis method:

[0107] Establish a coordinate system with the neighbor node of the current nth unknown node as the zero point:

[0108]

[0109] Among them, (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, (X j , Y j ) is the jth anchor node coordinate, for each anchor node (X j , Y j ), confirmed (Xj , Y j ) relative to the neighboring node (X n′ , Y n′ ) is located in the quadrant Q n′ .

[0110] Update the distribution probability function:

[0111]

[0112] in, The nth unknown node is located at the Qth n′ Quadrant probability value, initial value is 0, diag(Q n′ ) is Q n′ diagonal quadrant of HM nj For the current unknown node (X n , Y n ) to the anchor node (X j , Y j ), HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j ) minimum hop value.

[0113] Calculate the point (x s1 ,y s1 ), (x s2 ,y s2 ) quadrant distribution, according to the distribution probability function, the point in the high probability quadrant is taken as the end point coordinate (x t ,y t ).

[0114] S53: Translation vector calculation:

[0115]

[0116] in, represents the translation vector, It is a vector pointing from the initial coordinates of the current unknown node to the coordinates of the end point.

[0117] S54: Translation weight calculation:

[0118]

[0119] Among them, w(γ) is the translation weight, HM max is the maximum level of the current network, w min is the minimum translation weight of the current network, and γ is the unknown node level of the translation operation performed this time.

[0120] S55: Estimated position coordinates of unknown nodes after weighted translation correction:

[0121]

[0122] Among them, (x new ,y new )

[0123] are the unknown node coordinates after weighted translation.

[0124] The advantage of this step is that it can dynamically calculate the translation vector of each unknown node based on the distance between neighboring nodes and the anchor node self-positioning error obtained from the RSSI value, so that most unknown nodes can move closer to their true positions. Similarly, the calculation of the translation weight is also based on the current unknown node level γ and the current network maximum level HM max ,This dynamic correction method can not only improve the node positioning ,accuracy, but also significantly improve the stability and robustness of the ,algorithm of the present invention in random networks.

[0125] 1. Experimental environment:

[0126] In order to verify the performance of the improved algorithm, according to the method of the present invention, MatlabR2022a simulation software is used for simulation, and the experimental environment is set as follows: 100 sensor nodes are randomly distributed in an area of ​​100m×100m, the number of anchor nodes is 30, the number of unknown nodes is 70, and the maximum communication radius of the node is 30m; in order to avoid the randomness of the simulation results, the positioning situation of the algorithm in 100 random network topologies is statistically analyzed.

[0127] 2. Evaluation indicators:

[0128] This embodiment uses the average normalized error as an indicator to measure the performance of the algorithm, and its expression is as follows:

[0129]

[0130] Among them, ANE is the average normalized error, i is the number of unknown nodes, error n is the positioning error of the nth unknown node, and R is the maximum communication radius.

[0131] 3. Experimental results:

[0132] Figure 2The following figure shows the effect of the classic DV-HOP positioning algorithm and the improved DV-HOP positioning algorithm of the present invention under the same network topology. In the figure, the circular nodes are the true positions of the anchor nodes, the triangles are the true positions of the unknown nodes, the crosses are the positioning results of the classic DV-HOP positioning algorithm, the black solid line is the line connecting the true position of the unknown node to the positioning result of the classic DV-HOP positioning algorithm, the rectangles are the positioning results of the improved algorithm of the present invention, and the black dotted line is the line connecting the true position of the unknown node to the positioning result of the improved algorithm of the present invention. As can be seen from the figure, the positioning accuracy of the algorithm of the present invention is significantly higher than that of the classic DV-HOP positioning algorithm. Almost all the corrections of the unknown nodes make the estimated position of the unknown node closer to the true position.

[0133] Figure 3 The figure shows the positioning errors of the classic DV-HOP positioning algorithm and the algorithm of the present invention in 100 randomly generated networks. The average normalized error is used to measure the algorithm performance. It can be seen from the figure that the positioning accuracy of the positioning algorithm of the present invention is much higher than that of the classic DV-HOP positioning algorithm as a whole, and the normalized error of the positioning algorithm of the present invention has a smaller fluctuation range. The stability in random networks is significantly better than that of the classic DV-HOP positioning algorithm.

[0134] Figure 4 The following figure compares the positioning accuracy of the proposed algorithm and the classic DV-HOP positioning algorithm at different anchor node ratios. In a 100x100 grid, the number of summary nodes is set to 100, the maximum communication radius is 30m, and the anchor node ratio is increased from 10% to 45%. As can be seen from the figure, the mean normalized error of both algorithms decreases with increasing anchor node ratio, reaching a plateau at 35%. The proposed algorithm has higher positioning accuracy than the classic DV-HOP positioning algorithm at all anchor node ratios.

[0135] Figure 5 The following figure compares the positioning accuracy of the proposed algorithm and the classic DV-HOP positioning algorithm at different communication radii. In a 100x100 grid, the number of summary nodes is set to 100, the anchor node ratio is set to 30%, and the maximum communication radius is increased from 15m to 45m. As can be seen from the figure, the mean normalized error of both algorithms decreases with increasing communication radius and gradually levels off. The proposed algorithm has higher positioning accuracy than the classic DV-HOP positioning algorithm at all communication radii.

[0136] Figure 6This figure compares the positioning accuracy of the algorithm of the present invention and the classic DV-HOP positioning algorithm at different numbers of summary nodes. In a 100*100 grid, the number of summary nodes is set to increase from 100 to 350, the anchor node ratio is 30%, and the maximum communication radius is 30m. As can be seen from the figure, the average normalized error of the classic DV-HOP positioning algorithm fluctuates with the increase in the number of summary nodes. Overall, the positioning accuracy shows an upward trend. The average normalized error of the algorithm of the present invention decreases with the increase in the number of summary nodes, and is generally more stable. The positioning accuracy of the algorithm of the present invention is higher than that of the classic DV-HOP positioning algorithm at different total node numbers.

[0137] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A hierarchical weighted translation DV-HOP node positioning method based on RSSI correction, characterized in that: The steps are as follows: S1: Use the classic DV-HOP positioning algorithm to self-locate the anchor nodes at known locations in the network, calculate the anchor node self-localization error based on the estimated position and the actual position, and use the quartile method to calculate the distribution of the anchor node self-localization error to obtain the best estimated error; S2: Classify the unknown nodes in the entire network according to the minimum hop count array from the unknown node to the nearest anchor node after flooding, and mark the neighbor nodes of each unknown node; S3: Positioning unknown nodes step by step: Starting from the first-level nodes, the unknown nodes in the entire network are positioned step by step to the last-level nodes. The positioning process of each level of nodes is divided into initial positioning and initial positioning correction. Initial positioning: Use the average hop distance of the non-unbiased estimate of the neighboring nodes as the average hop distance of the corresponding unknown node, calculate the estimated distance from the corresponding unknown node to all anchor nodes, and use the standard least squares method to solve the initial positioning coordinates of the corresponding unknown node; Initial positioning correction: Construct the RSSI circle and the best estimation error circle, calculate the translation vector and translation weight, and perform a weighted translation on the initial positioning coordinates along the direction of the specific correction vector to obtain the corrected coordinates.

2. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 1 is characterized in that: The calculation process of the best estimation error in S1 is: S21: Anchor node self-positioning error calculation: Among them, Error j is the self-positioning error of the j-th anchor node, (X o ,Y o ) is the estimated position coordinate of the j-th anchor node calculated by the classic DV-HOP positioning algorithm, (X j ,Y j ) is the coordinate of the j-th anchor node; S22: Calculate the best estimate error: Among them, Error Q2 is the best estimation error, q is the number of anchor nodes, and the anchor node self-localization error sample data set X = {Error1, Error2, ..., Error j ,...,Error q }, sort the data from small to large, and get the ordered sequence X′={m1,m2,...,m k ,...,m q }, where m k Error is the self-positioning error of the anchor node at the kth position after sorting. Q2 Take the median of the permutation.

3. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 1 is characterized in that: The unknown node classification and neighbor node marking method in S2 is: S31: unknown node classification method: mark the unknown node that is one hop away from the anchor node as a first-level node, the unknown node that is two hops away from the anchor node as a second-level node, and so on; S32: Neighbor node marking method: mark the anchor node with the smallest RSSI estimated distance to the first-level node as the neighbor node of the corresponding first-level node; The first-level node with the smallest RSSI estimated distance between the marker and the second-level node is the neighbor node of the corresponding second-level node, and so on.

4. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 3 is characterized in that: The calculation method of the RSSI estimated distance in S32 is: in, is at a distance d est The received signal strength at is the signal strength received at the reference distance d0, σ is the path loss exponent, d0 is the reference distance, d est is the distance between the receiver and the transmitter, which is the RSSI estimated distance, X σ Represents the blocking factor.

5. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 1 is characterized in that: The initial positioning steps for each level of unknown nodes in S3 are: S51: Calculate the average hop distance of the neighboring nodes' non-unbiased estimate: Among them, i is the number of unknown nodes, n′ represents the neighbor node of the current nth unknown node, and Dhop n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s average hop distance, DIS n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s Euclidean distance, HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j ), (X j , Y j ) is the coordinate of the jth anchor node, (X n′ , Y n′ ) is the coordinate of the neighbor node n′ of the current nth unknown node; S52: Calculation of the distance from the unknown node to the anchor node: d nj =Dhop n′j ·HM nj ; Among them, d nj is the estimated distance from unknown node n to the jth anchor node, HM nj is the minimum number of hops from unknown node n to the jth anchor node; S53: Construct the matrix form of the difference equation: Among them, the matrix P is the design matrix, which contains the anchor node coordinates and differential relationship, and the matrix H is the constant matrix, which contains the square difference of the anchor node distance and the difference of the sum of the square of the anchor node coordinates; C j =(X j , Y j ) is the coordinate of the j-th anchor node, represents the sum of squares of the j-th anchor node coordinates; S54: Use the least squares method to solve the initial positioning coordinates of the unknown nodes: X=(P T P) -1 P T H; Where X is the initial positioning coordinate of the unknown node.

6. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 1 is characterized in that: The initial positioning correction steps for each level of unknown nodes in S3 are: S61: Construct RSSI circle and best estimation error circle: Among them, the coordinate system is established with x as the horizontal coordinate and y as the vertical coordinate. oe , Y oe ) is the initial positioning coordinate of the unknown node, Error Q2 is the best estimation error, (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, R min RSSI estimated distance between the current unknown node and its neighboring nodes; S62: Intersection relationship between RSSI circle and best estimation error circle and calculation of end point coordinates: Among them, (x t ,y t ) is the end point coordinate, (x s1 ,y s1 ), (x s2 ,y s2 ) is the coordinate of the intersection point where the RSSI circle intersects with the best estimation error circle. When the RSSI circle intersects with the best estimation error circle, (x t ,y t ) has two solutions. By using the regional frequency analysis method, (x s1 ,y s1 ),(x s2 ,y s2 ) selects the end point coordinates; t is the proportional coefficient, D is the distance between the center of the RSSI circle and the best estimation error circle; where (X n′ +t(X oe -X n′ ),Y n′ +t(Y oe -Y n′ )) is the coordinate of the intersection of the RSSI circle and the center of the line segment passing through the RSSI circle and the best estimation error circle when the two circles do not intersect; S63: Translation vector calculation: in, represents the translation vector, It is a vector pointing from the initial coordinates of the current unknown node to the coordinates of the end point; S64: Translation weight calculation: Among them, w(γ) is the translation weight, HM max is the maximum level of the current network, w min is the minimum translation weight of the current network, and γ is the unknown node level of the translation operation performed this time; S65: Estimated position coordinates of unknown nodes after weighted translation correction: Among them, (x new ,y new ) is the estimated position coordinate of the unknown node after weighted translation correction.

7. The hierarchical weighted translation DV-HOP node positioning method based on RSSI correction according to claim 6 is characterized in that: The regional frequency analysis method in S62 selects the end point coordinate (x t ,y t ) process is: S71: Establish a coordinate system with the neighboring node of the current nth unknown node as the zero point: Where (X n′ , Y n′ ) is the neighbor node coordinate of the current nth unknown node, (X j ,Y j ) is the jth anchor node coordinate, for each anchor node (X j , Y j ), confirmed (X j , Y j ) relative to the neighboring node (X n′ , Y n′ ) is located in the quadrant Q n′ ; S72: Update distribution probability function: in, The nth unknown node is located at the Qth n′ Quadrant probability value, initial value is 0, diag(Q n′ ) is Q n′ diagonal quadrant of HM nj For the current unknown node (X n , Y n ) to the anchor node (X j , Y j ), HM n′j For neighbor nodes (X n′ , Y n′ ) to the anchor node (X j , Y j )’s minimum hop value; S73: Calculate the point (x s1 ,y s1 ), (x s2 ,y s2 ) quadrant distribution, according to the distribution probability function, the point in the high probability quadrant is taken as the end point coordinate (x t ,y t ).

Images