Three-dimensional passive direct positioning method based on TDOA and FDOA

Abstract

The invention relates to a three-dimensional passive direct positioning method based on TDOA and FDOA. The three-dimensional passive direct positioning method comprises the following steps: S0, acquiring initial position coordinates of a target positioning area and a signal source; the method comprises the following steps: S1, setting one of sensing nodes as a reference node, calculating TDOA vectors between the sensing nodes and the reference node, and extracting a plurality of sensing nodes corresponding to the minimum TDOA vectors as trust nodes; s2, setting a to-be-searched three-dimensional grid region based on the initial position coordinates of the signal source; and S3, constructing a positioning cost function according to the TDOA vector and the FDOA vector between each trust node and the reference node, and iteratively solving a signal source position coordinate enabling the value of the positioning cost function to be maximum by using a grid search direct positioning algorithm. According to the invention, positioning tasks under different application scenes can be coped with, and accurate positioning in a three-dimensional space is realized.

Description

Technical Field

[0001] The present invention relates to the field of passive positioning technology, and in particular to a three-dimensional passive direct positioning method based on TDOA and FDOA. Background Art

[0002] Existing UAV positioning methods are classified according to observation information, including direction of arrival (DOA), time difference of arrival (TDOA), frequency difference of arrival (FDOA) and received signal strength (RSS). In different application scenarios, the difficulty of obtaining each observation information, the credibility of the information obtained and the accuracy of the final positioning solution (estimate) are all different. Therefore, the positioning method based on a single observation information is difficult to cope with positioning tasks in different application scenarios.

[0003] The existing three-dimensional spatial positioning method mainly adopts the active positioning method. The active positioning method is similar to the principle of radar. The self-transmits a radio signal outward. The signal will be reflected back after encountering the target. The target is located by analyzing the received reflected signal. Although this method has high accuracy, it has security risks. The radio signal transmitted by the self-transmitting party can be easily detected by the enemy during the transmission process. Once detected, the enemy can track the self-transmitting party according to the transmission position and characteristics of the signal, thereby exposing the self-transmitting party's position and endangering the security of the entire system.

[0004] The traditional two-step positioning method (TSD) is prone to introduce errors in the positioning process. First, it needs to use the observed signal to estimate the intermediate measurement value required for positioning solution. This step may lead to inaccurate estimation of the intermediate measurement value due to factors such as noise and interference of the observed signal. Then, a nonlinear equation system is constructed for the intermediate measurement value. The linearization process of the nonlinear equation system will also introduce errors, because linearization is an approximate process and it is difficult to fully and accurately reflect the essential characteristics of the nonlinear equation system. Ultimately, these errors will accumulate in the position estimation result, resulting in a decrease in positioning accuracy. Especially in a low signal-to-noise ratio environment, the observed signal itself is subject to greater noise interference, and the estimation error of the intermediate measurement value will be even greater, which greatly reduces the positioning performance of the TSD method and cannot meet the needs of high-precision positioning. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a three-dimensional passive direct positioning method based on TDOA and FDOA, which can cope with positioning tasks in different application scenarios and achieve precise positioning in three-dimensional space.

[0006] The technical solution adopted by the present invention to solve the technical problem is: to provide a three-dimensional passive direct positioning method based on TDOA and FDOA, which is applied to a system including a signal source and a plurality of sensing nodes that observe the signal source multiple times during movement, and includes the following steps:

[0007] S0 obtains the initial position coordinates of the target positioning area and the signal source;

[0008] S1 sets one of the sensing nodes as a reference node, calculates the TDOA vector between each sensing node and the reference node according to the initial position coordinates of the signal source, and extracts the sensing nodes corresponding to the set number of minimum TDOA vectors as trust nodes;

[0009] S2: setting a three-dimensional grid area to be searched based on the initial position coordinates of the signal source, wherein the three-dimensional grid area includes a set number of three-dimensional grid units;

[0010] S3 constructs a positioning cost function according to the TDOA vector and the FDOA vector between each trusted node and the reference node, and uses a grid search direct positioning algorithm to iteratively solve the signal source position coordinates that maximize the value of the positioning cost function.

[0011] Furthermore, the method of iteratively solving the signal source position coordinates that maximize the value of the positioning cost function by using a grid search direct positioning algorithm includes:

[0012] The grid point coordinates of each three-dimensional grid unit are used as candidate position coordinates of the signal source;

[0013] According to the TDOA vector and FDOA vector between each trust node and the reference node, respectively construct the positioning cost function when the signal source is at different candidate position coordinates, and update the initial position coordinate of the signal source to the candidate position coordinate that maximizes the value of the positioning cost function;

[0014] The search step is calculated based on the size of the three-dimensional grid unit, and it is determined whether the current search step is greater than a set threshold. If it is greater than the set threshold, the three-dimensional grid unit is reduced and the process returns to step S0. Otherwise, the updated initial position coordinates of the signal source are output.

[0015] Furthermore, the method also includes calculating the average value of the initial position coordinates of all rounds of updated signal sources as the signal source positioning result.

[0016] Furthermore, constructing a positioning cost function when the signal source is at different candidate position coordinates includes:

[0017] For any candidate position coordinate, calculate the TDOA vector and FDOA vector between each trust node and the reference node when the signal source is located at the candidate position coordinate at the current observation time, and use the calculated TDOA vector and FDOA vector to compensate the corresponding trust node receiving signal to the reference node receiving signal, and obtain the compensated reference node receiving signals corresponding to all trust nodes as the receiving signal matrix, and then solve the maximum eigenvalue of the receiving signal matrix as the positioning eigenvalue at the current observation time;

[0018] The positioning cost function for constructing any candidate position coordinate is the sum of the positioning feature values ​​at all observation times.

[0019] Furthermore, the method of calculating the maximum eigenvalue of the received signal matrix is ​​achieved by respectively calculating the eigenvalue vectors of the received signal matrix at the current observation moment by using singular value decomposition and extracting the maximum eigenvalue therein.

[0020] Furthermore, when the search calculation is performed for the first time, the initial position coordinates of the signal source are set as the center coordinates of the target positioning area where the signal source is located.

[0021] Furthermore, when the search calculation is performed for the first time, a set number of perception nodes are randomly selected as trust nodes.

[0022] Furthermore, the number of the trusted nodes is not less than three.

[0023] Furthermore, the three-dimensional grid area includes three-dimensional grid units, where and are both integers greater than zero.

[0024] Furthermore, the search step length is the length of the shortest side of the three-dimensional grid unit.

[0025] Beneficial Effects

[0026] Due to the adoption of the above-mentioned technical scheme, the present invention has the following advantages and positive effects compared with the prior art: the present invention combines a variety of information types to build a model, uses the advantages of different observation information to achieve multi-information fusion positioning, copes with positioning tasks in different application scenarios, and can achieve better positioning effects under more application restrictions; for the case where the target itself is the signal source, the present invention only uses the sensing node to receive the signal, and realizes the positioning of the target by analyzing and processing the received signal. The sensing node does not need to transmit a signal to the outside, so it is not easy to be detected and tracked by the enemy, thereby improving concealment and security; the present invention adopts a direct positioning method, by constructing the relationship between the signal model and the position estimation, the position of the target is directly given, the error introduced by the intermediate measurement value estimation is reduced, it has higher accuracy, and has higher positioning robustness under low signal-to-noise ratio conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of an embodiment of the present invention;

[0028] Figure 2 is a schematic diagram of a swarm of drones flying into an area of ​​interest in an embodiment of the present invention;

[0029] Figure 3 is a schematic diagram of selecting three groups of drone nodes to collect data in an embodiment of the present invention;

[0030] Figure 4 is a schematic diagram of a drone group flying to the next moment in an embodiment of the present invention;

[0031] Figure 5 is a schematic diagram of strategically selecting drone nodes in an embodiment of the present invention;

[0032] Figure 6 is a schematic diagram of a group of drones flying away from a positioning area in an embodiment of the present invention;

[0033] Figure 7 is a schematic diagram of the first step of searching and positioning in an embodiment of the present invention;

[0034] Figure 8 is a schematic diagram of a fine search centered on the last result in an embodiment of the present invention;

[0035] Fig. 9 is a schematic diagram of a search range for three-dimensional spatial positioning in an embodiment of the present invention;

[0036] Fig.10 It is a flow chart of the positioning algorithm in the implementation manner of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.

[0038] The embodiments of the present invention relate to a three-dimensional passive direct positioning method based on TDOA and FDOA, which is applied to a system including a signal source and a plurality of sensing nodes that observe the signal source multiple times during movement, such as Figure 1 As shown, the following steps are included:

[0039] S0 obtains the initial position coordinates of the target positioning area and the signal source;

[0040] S1 sets one of the sensing nodes as a reference node, calculates the TDOA vector between each sensing node and the reference node according to the initial position coordinates of the signal source, and extracts the sensing nodes corresponding to the set number of minimum TDOA vectors as trust nodes;

[0041] S2: setting a three-dimensional grid area to be searched based on the initial position coordinates of the signal source, the three-dimensional grid area including a set number of three-dimensional grid units;

[0042] S3 uses the grid point coordinates of each three-dimensional grid unit as the candidate position coordinates of the signal source, and constructs the positioning cost functions when the signal source is at different candidate position coordinates according to the TDOA vectors and FDOA vectors between each trust node and the reference node, and updates the initial position coordinates of the signal source to the candidate position coordinates that maximize the value of the positioning cost function;

[0043] S4 calculates the search step size based on the size of the three-dimensional grid unit, and determines whether the current search step size is greater than a set threshold. If it is greater than the set threshold, the three-dimensional grid unit is reduced and returns to step S0, otherwise the updated initial position coordinates of the signal source are output.

[0044] The following takes a drone cluster as an example to further illustrate the principles and specific steps of this implementation.

[0045] Each drone in the drone cluster of this embodiment is equipped with an omnidirectional antenna and flies in the area of ​​interest where the unknown signal source is located to perform real-time signal perception detection. The drones are numbered from 1 to L. When the existence of the signal source of interest is detected, each drone starts to collect and store the intercepted signal raw data at the specified time and continues to fly. The time for collecting the raw data once is recorded as t s us. According to the wireless communication link bandwidth of the drone, select the appropriate t s us, usually the original data points selected for a sampling are greater than or equal to 512, and each drone transmits the sampling points back to the ground terminal through a wireless link.

[0046] When the signal source is first located, the data of three drones are randomly selected (three-dimensional spatial positioning requires at least three sets of collected data, and if necessary, more sets of data can be used to improve positioning accuracy) as the initial data set, and the three drones are renumbered in order from small to large according to their actual numbers. make The sampling data is used as the benchmark data, and the signal source position is directly located and solved by a three-dimensional grid search at the ground terminal; at the same time, the data sent back by all drones are simply processed to obtain the TDOA estimation result.

[0047] In addition to randomly selecting 3 drone sampling data before the first positioning solution to obtain the fuzzy location information of the unknown signal source, from the second positioning solution, 3 drone data with high credibility will be selected for positioning solution according to a specific strategy. That is, after the first estimation is completed, the position estimate of the signal source is obtained. And as the drone swarm has moved to the next moment's location, the next drone used for signal source positioning will be strategically selected based on the fuzzy position information.

[0048] This implementation uses the estimation results of TDOA to measure the credibility of all drone sampling data in the drone cluster. The specific method is as follows:

[0049] 1) When locating the signal source for the first time, based on the center coordinates P0 (x0, y0, z0) of the region of interest and the coordinates u of the L drones at the first observation time l,1 , calculate the delay D of the signal received by the lth UAV at the first observation time l,1 ,

[0050]

[0051] Where, l = 1,…,L, c represents the signal propagation speed, and ||·|| represents the Euclidean norm;

[0052] 2) The ground terminal calculates the TDOA vector between the lth UAV and the UAV numbered 1 at the first moment

[0053] ΔD1={ΔD 1,1 ,ΔD 2,1 ,…ΔD L,1}

[0054] in,

[0055] ΔD l,1 =D l,1 -D 1,1

[0056] ΔD k Sort by smallest to largest, and we get

[0057] 3) Take The drones corresponding to the smallest q values ​​are regarded as trust nodes and are renumbered in ascending order according to their actual numbers.

[0058] 4) In the i-th positioning solution, i ≥ 2, the coordinates are estimated based on the i-1-th signal source position The coordinates u of L drones at the kth observation time l,k , calculate the delay D of the lth UAV receiving the signal at the kth observation time l,k ,

[0059]

[0060] 5) The ground terminal calculates the TDOA vector between the lth UAV and the UAV numbered 1 at the kth moment

[0061] ΔD k ={ΔD 1,k ,ΔD 2,k ,…ΔD L,k}

[0062] in,

[0063] ΔD l,k =D l,k -D 1,k

[0064] ΔD k Sort by smallest to largest, and we get

[0065] 6) Take The drones corresponding to the smallest q values ​​are renumbered in ascending order according to their actual numbers.

[0066] The selection process of the trust node drone is as follows Figures 2 to 6 As shown, q=3 is selected in the figure.

[0067] This implementation adopts a three-dimensional grid search direct positioning method to solve the position of the signal source. The initial search range of the grid search needs to be input by the user, and the input area is divided into n×n×m three-dimensional grid points, where n represents the number of search points on the X-axis and Y-axis, and m represents the number of search points on the Z-axis.

[0068] The search process of grid search on the X-axis and Y-axis is as follows Figures 7 and 8 As shown in the figure, by assuming that the signal source is at each grid point, when it is at a certain grid point, the distance between the assumed position of the signal source and the sensing node can be calculated. When the distance is calculable, the TDOA information and FDOA information between different sensing nodes can be calculated. Then, the calculated signal is matched with the actual received signal. It is believed that the higher the matching degree, the closer the assumed position is to the unknown signal source. The cost function is used to measure the degree of signal matching. Generally speaking, the larger the value of the cost function, the higher the matching degree. The condition for the termination of the search of the grid search method is generally that the distance between the grid points is less than a certain threshold. When the number of iterations is very large and the distance between the grid points is close enough, the positioning accuracy can be extremely high.

[0069] like Fig. 9As shown, the search area during three-dimensional space search is a cube, and the search center is the centroid of the cube.

[0070] Specifically, the 3D grid search direct positioning solution method uses the user-specified area of ​​interest centroid as the starting point and the specified search range as the initial parameter, dividing the search range into n×n×m three-dimensional grid points. By assuming that the signal source is at each grid point, the TDOA and FDOA of the selected UAV are estimated at the assumed position, and the estimated values ​​are used to calculate the The received signal compensation is The received signal is used to form a received signal matrix, and the maximum eigenvalue of the received signal matrix is ​​calculated. The maximum eigenvalue calculated by k observation results is accumulated to obtain the extreme point of the positioning cost equation. This extreme point is the point where the search grid point is closest to the true position. Then the first positioning result is used as the center of the search range, the search area is narrowed, and the above process is repeated until the iteration termination condition is met, and the positioning result is output. Fig.10 As shown in Figure 1, the specific process of three-dimensional grid search includes:

[0071] 1) If this is the first time to locate the unknown signal source, enter the centroid coordinates P0 (x0, y0, z0) of the range of interest, otherwise the initial position of the iteration is

[0072] 2) Calculate the coordinates of the n×n×m grid points in the grid search based on the initial position P1(x1,y1,z1),P2(x2,y2,z2),…,P g (x g ,y g ,z g ),g=n×n×m, the search range is R×R×H;

[0073] 3) The current minimum search step is recorded as step = min{R / n,H / m};

[0074] 4) If step>η, execute step 5), otherwise execute step 14), where η is the iteration termination condition set according to the positioning accuracy requirement;

[0075] 5) Take the coordinates of a grid point in turn, assuming that the current coordinate is P g (x g ,y g ,z g );

[0076] 6) Calculate the TDOA vector ΔD at time k in sequence k and FDOA vector Δf k , k = 1, 2, ..., where

[0077] ΔDk ={ΔD 1,k , ΔD 2,k , ...ΔD L,k}

[0078] Δf k ={Δf 1，k , Δf 2，k , …Δf L，k}

[0079] ΔD l，k =D l，k -D 1，k

[0080] Δf l，k =f l，k -f 1，k

[0081] 7) According to the estimated TDOA vector ΔD k and FDOA vector Δf k Will The received signal compensation is For the convenience of calculation, this patent adopts the frequency domain compensation method.

[0082]

[0083] where l = 2, 3, ..., q, S l，k is the Fourier transform of the time domain signal point collected by the lth UAV at the kth sampling time, W l，k is the Fourier transform of the time domain noise;

[0084] 8) Calculate the compensation The receiving signal matrix is ​​set here is a column vector signal, then

[0085]

[0086] 9) Calculate svd(V T V), we get the eigenvalue vector Λ={λ1,λ2,λ3,...,λ q};

[0087] 10) Calculate G k = argmax(Λ);

[0088] 11) Calculate P g Unknown positioning cost function value C g =∑G k ;

[0089] 12) The positions P corresponding to the n×n×m search points gIn the middle, it is the positioning solution value under the current search range, recorded as

[0090] 13) Reduce the grid, set R = R × α, H = H × α, where 0 < α < 1, and repeat steps 3) to 10);

[0091] 14) Output the positioning solution result, that is, the positioning cost function value C g Maximum position coordinates

[0092] Finally, until the drone cluster flies away from the positioning area, multiple positioning solutions have been performed, and the solution results are recorded as In order to reduce the variance of the positioning error, the average value of multiple solution results is used as the final positioning result.

[0093]

[0094] Where N is the final number of positioning solutions.

Claims

1. A three-dimensional passive direct positioning method based on TDOA and FDOA, characterized in that: A system applied to a signal source and a plurality of sensing nodes for multiple observations of the signal source during movement, comprising the following steps: S0 obtains the initial position coordinates of the target positioning area and the signal source; S1 sets one of the sensing nodes as a reference node, calculates the TDOA vector between each sensing node and the reference node according to the initial position coordinates of the signal source, and extracts the sensing nodes corresponding to the set number of minimum TDOA vectors as trust nodes; S2: setting a three-dimensional grid area to be searched based on the initial position coordinates of the signal source, wherein the three-dimensional grid area includes a set number of three-dimensional grid units; S3 constructs a positioning cost function according to the TDOA vector and the FDOA vector between each trusted node and the reference node, and uses a grid search direct positioning algorithm to iteratively solve the signal source position coordinates that maximize the value of the positioning cost function.

2. The method according to claim 1, characterized in that: The method of iteratively solving the signal source position coordinates that maximize the value of the positioning cost function by using a grid search direct positioning algorithm includes: The grid point coordinates of each three-dimensional grid unit are used as candidate position coordinates of the signal source; According to the TDOA vector and FDOA vector between each trust node and the reference node, respectively construct the positioning cost function when the signal source is at different candidate position coordinates, and update the initial position coordinate of the signal source to the candidate position coordinate that maximizes the value of the positioning cost function; The search step is calculated based on the size of the three-dimensional grid unit, and it is determined whether the current search step is greater than a set threshold. If it is greater than the set threshold, the three-dimensional grid unit is reduced and the process returns to step S0. Otherwise, the updated initial position coordinates of the signal source are output.

3. The method according to claim 2, characterized in that The method also includes calculating the average value of the initial position coordinates of all rounds of updated signal sources as the signal source positioning result.

4. The method according to claim 2, characterized in that: The constructing of the positioning cost function when the signal source is at different candidate position coordinates includes: For any candidate position coordinate, calculate the TDOA vector and FDOA vector between each trust node and the reference node when the signal source is located at the candidate position coordinate at the current observation time, and use the calculated TDOA vector and FDOA vector to compensate the corresponding trust node receiving signal to the reference node receiving signal, and obtain the compensated reference node receiving signals corresponding to all trust nodes as the receiving signal matrix, and then solve the maximum eigenvalue of the receiving signal matrix as the positioning eigenvalue at the current observation time; The positioning cost function for constructing any candidate position coordinate is the sum of the positioning feature values ​​at all observation times.

5. The method according to claim 4, characterized in that The method of calculating the maximum eigenvalue of the received signal matrix is ​​achieved by using singular value decomposition to respectively calculate the eigenvalue vectors of the received signal matrix at the current observation moment and extracting the maximum eigenvalue therein.

6. The method according to claim 1, characterized in that During the first round of search calculation, the initial position coordinates of the signal source are set to the center coordinates of the target positioning area where the signal source is located.

7. The method according to claim 1, characterized in that During the first round of search calculation, a set number of perception nodes are randomly selected as trust nodes.

8. The method according to claim 1, characterized in that The number of the trust nodes is not less than three.

9. The method according to claim 1, characterized in that: The three-dimensional grid area includes n×n×m three-dimensional grid units, wherein n and m are both integers greater than zero.

10. The method according to claim 1, characterized in that The search step length is the length of the shortest side of the three-dimensional grid unit.

Images