Disaster area rescue target positioning method based on BP neural network optimized by sparrow algorithm

Abstract

The invention relates to the technical field of target positioning, and discloses a disaster area rescue target positioning method for optimizing a BP neural network based on a sparrow algorithm, and the method comprises the steps: carrying out the preprocessing of unmanned plane record information during the cruising period of an unmanned plane; establishing a BP neural network model, initializing the weight and threshold of the BP neural network model, inputting the table into the BP neural network model, subtracting the output mobile phone longitude and latitude from the actual mobile phone longitude and latitude as the fitness of the sparrow, operating a sparrow optimization algorithm, and iteratively finding the optimal fitness value; taking the sparrow data corresponding to the optimal fitness value as a final weight and a final threshold value, taking the table as input to train a BP neural network model, and putting the trained BP neural network model into unmanned aerial vehicle remote control; and inputting the arranged data lines into the trained BP neural network model to obtain a final predicted position latitude and longitude, and marking the predicted mobile phone position on a map. According to the invention, accurate positioning is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of target positioning, and in particular to a disaster area rescue target positioning method based on a sparrow algorithm to optimize a BP neural network. Background Art

[0002] The sparrow algorithm is a heuristic optimization algorithm based on the behavior of bird flocks, simulating the behavior of sparrows in the process of foraging and reproduction. The algorithm searches for the optimal solution through the interaction between individuals in the group, and has the advantages of strong global search ability and fast convergence speed. Its core idea is to initialize the position and speed of the sparrow, calculate the fitness value (such as the error size of the BP neural network) according to the current position, and update the optimal position and the global optimal position according to the fitness value, and finally find the optimal solution by continuously adjusting the speed and position. Due to its efficient global search ability and fast convergence speed, the sparrow algorithm is widely used in various optimization problems, including but not limited to machine learning, data mining, image processing and other fields. BP neural network (Back Propagation Neural Network) is a commonly used machine learning model that is trained and optimized through the back propagation algorithm. The network consists of an input layer, a hidden layer and an output layer, and the mapping relationship between input and output is approximated by adjusting the weights and bias values ​​between each layer. BP neural network has the advantages of strong adaptability and strong nonlinear mapping ability, and is widely used in tasks such as classification and regression.

[0003] However, there are some problems in the training process of BP neural network, such as slow convergence speed and easy to fall into local minimum. These problems limit the performance and application scope of BP neural network. GPS positioning has disadvantages such as limited signal, accuracy problem, dependency problem, energy consumption and cost, as well as privacy and security risks. In disaster scenarios, many collapsed objects have a relatively large impact on GPS positioning, and mobile phones with accurate GPS positioning require relatively large costs. In disaster scenarios, nearby base stations will be damaged and cannot be used. WiFi positioning and Bluetooth positioning are only applicable to indoor scenarios and not outdoor scenarios. AOA technology determines the location by the angle of the signal source, and requires multiple sensors to be equipped on the device to meet the requirements. The equipment is expensive and has great technical difficulties. TDOA technology determines the location by the time difference of the signal arrival. The positioning effect in the ideal space is good, but when encountering buried terrain, the signal time difference will be greatly affected and the positioning accuracy is poor. Since AOA and TDOA use the signal strength or azimuth data to calculate the mobile phone position through formulas, when encountering obstacles, the interference coefficient of the signal strength is difficult to detect or calculate. At the same time, there are multipath problems in both technologies. Multipath refers to the situation where, after electromagnetic waves propagate through different paths, each component field arrives at the receiving end at different times, and they overlap with each other according to their respective phases, causing interference, which distorts the original signal or generates errors.

[0004] Therefore, it is necessary to provide a disaster area rescue target positioning method based on the sparrow algorithm to optimize the BP neural network to solve the problem that the positioning method in the prior art has poor accuracy and is easily affected by the environment. Summary of the invention

[0005] In view of this, the present invention proposes a disaster area rescue target positioning method based on the sparrow algorithm to optimize the BP neural network, aiming to solve the problem that the positioning method in the prior art has poor accuracy and is easily affected by the environment.

[0006] The present invention proposes a disaster area rescue target positioning method based on the sparrow algorithm to optimize the BP neural network, comprising:

[0007] Pre-process the mobile phones that are continuously logged in, collect mobile phone information through drones equipped with positioning payloads, pre-process the drone recorded information during the drone cruise, and record it in a table;

[0008] Establishing a BP neural network model, initializing the weights and thresholds of the BP neural network model, substituting the data of each sparrow into the BP neural network model, inputting the table into the BP neural network model, subtracting the latitude and longitude of the mobile phone output by the BP neural network model from the actual latitude and longitude of the mobile phone as the fitness of the sparrow, running the sparrow optimization algorithm, and iteratively finding the optimal fitness value;

[0009] The sparrow data corresponding to the optimal fitness value is used as the final weight and threshold of the BP neural network model, the table is used as input to train the BP neural network model, and the trained BP neural network model is put into the UAV remote control so that the UAV remote control processes the received data in real time;

[0010] Determine the scope of the disaster by remote sensing technology, automatically cruise the drone carrying the positioning payload over the disaster area, collect the mobile phone information, and transmit the mobile phone information and drone record information back to the drone remote controller;

[0011] Each group of collected mobile phone data is arranged and combined into three groups as a row to form a data row. The arranged data row is used as input and input into the trained BP neural network model to obtain the final predicted location longitude and latitude, and the mobile phone location predicted by the BP neural network model is marked on the map.

[0012] Furthermore, the pre-processing of the mobile phones that continue to log in includes:

[0013] Place the mobile phones that are continuously being recorded in a square flat area, bury them in an "inclined" manner using concrete bricks, and record the actual latitude and longitude of the phone.

[0014] Furthermore, the collecting of mobile phone information by a drone equipped with a positioning payload includes:

[0015] The positioning payload is installed directly below the drone, and the signal receiving side of the payload is kept facing downward. The remote-controlled drone automatically cruises in a square flat area and flies around the mobile phone. When the mobile phone is scanned, the drone exchanges data with the mobile phone to obtain the mobile phone information.

[0016] Furthermore, the mobile phone information includes: mobile phone serial number and mobile phone signal strength;

[0017] The drone recorded information includes: drone altitude, drone longitude and latitude, and mobile phone signal strength.

[0018] Furthermore, the pre-processing of the drone record information during the drone cruise and recording it in a table includes:

[0019] The drone record information during the drone cruise is recorded and exported through the remote control, the drone record information is processed in groups, the three groups and the actual longitude and latitude of the mobile phone are combined into a large group, and the recorded data is put into a table to be trained using a computer, wherein the mobile phone signal strength in the data of the three groups is for the same mobile phone.

[0020] Furthermore, the large group is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3, actual mobile phone longitude, actual mobile phone latitude).

[0021] Furthermore, the process of establishing the BP neural network model, initializing the weights and thresholds of the BP neural network model, and substituting the data of each sparrow into the BP neural network model includes:

[0022] The number of input layer nodes of the BP neural network model is n, the number of hidden layer nodes is m, and the number of output layer nodes is p, then the number of weights is N = n*m + m*p, and the number of thresholds is M = m + p;

[0023] Set a number of sparrows, each sparrow has M+N data and contains all weights and thresholds, and initialize the data of each sparrow.

[0024] Furthermore, the collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is input as input to the trained BP neural network model to obtain the final predicted location longitude and latitude, including:

[0025] Each group of collected mobile phone data is arranged and combined into three groups as a row to form a data row of 12 data. The arranged data rows are used as input and input into the trained BP neural network model, where the data behavior is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3).

[0026] Furthermore, the collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is input as input to the trained BP neural network model to obtain the final predicted location longitude and latitude, including:

[0027] The predicted longitude and latitude output by the BP neural network model are sorted by size, the largest and smallest data, a total of 1 / 4, are removed, the middle data are taken for mean processing, and the final predicted position longitude and latitude are obtained.

[0028] Furthermore, the loss function of the BP neural network model is MSE, and the activation function is set to tans ig and pure in.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows: the method of establishing a fitting model by optimizing the neural network with sparrows perfectly solves the difficulties of low precision and high cost of the AOA and TDOA methods by using existing data to establish a model to predict the application data. It does not cost too much to establish a model through a large amount of data, and the accuracy is also guaranteed by the method of data fitting. The positioning method of the sparrow-optimized BP neural network model uses a large amount of data verification training to enable the model to cope with most situations in the training scenario, and will not cause positioning deviation due to individual special signal strengths. On the multipath problem, the method of positioning the sparrow-optimized BP neural network model, the collected data before training the model will also have multipath propagation. In the process of fitting the training model to these data, the signal multipath propagation problem is also fitted, so that the model has a good anti-interference ability for the multipath propagation of the signal, which greatly reduces the impact of the signal multipath propagation. The output processing of multiple groups of data in the prediction process further handles the interference of the signal and improves the accuracy of positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0031] Figure 1 A flow chart of a method for locating disaster area rescue targets based on a sparrow algorithm optimizing a BP neural network provided by an embodiment of the present invention;

[0032] Figure 2 A flow chart of the sparrow optimization BP neural network algorithm provided by an embodiment of the present invention;

[0033] Figure 3 A schematic diagram of the flight route of a drone and the location of a mobile phone provided in an embodiment of the present invention;

[0034] Figure 4 A network diagram of the BP neural network model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0035] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0036] In some embodiments of the present application, see Figure 1 As shown, this embodiment provides a disaster area rescue target positioning method based on the sparrow algorithm to optimize the BP neural network, including the following steps:

[0037] S100, pre-processing the mobile phones that are continuously logged in, collecting mobile phone information through a drone equipped with a positioning payload, pre-processing the drone recorded information during the drone cruise, and recording it in a table;

[0038] S200, establishing a BP neural network model, initializing the weights and thresholds of the BP neural network model, substituting the data of each sparrow into the BP neural network model, inputting the table into the BP neural network model, subtracting the latitude and longitude of the mobile phone output by the BP neural network model from the actual latitude and longitude of the mobile phone as the fitness of the sparrow, running the sparrow optimization algorithm, and iterating to find the optimal fitness value;

[0039] S300, using the sparrow data corresponding to the optimal fitness value as the final weight and threshold of the BP neural network model, using the table as input to train the BP neural network model, and putting the trained BP neural network model into the drone remote control, so that the drone remote control processes the received data in real time;

[0040] S400, determining the disaster-affected area through remote sensing technology, automatically cruising a drone carrying a positioning payload over the disaster-affected area, collecting the mobile phone information, and transmitting the mobile phone information and drone record information back to the drone remote controller;

[0041] S500, arranging and combining each group of collected mobile phone data into three groups as a row to form a data row, and inputting the arranged data row as input into the trained BP neural network model to obtain the final predicted location longitude and latitude, and marking the mobile phone location predicted by the BP neural network model on the map.

[0042] Preferably, the continuously logged-in mobile phone refers to a mobile phone that can normally receive and send signals.

[0043] Specifically, Figure 2 Optimize the BP neural network algorithm flow chart for Sparrow.

[0044] In some embodiments of the present application, the preprocessing of the mobile phone that continues to log in includes:

[0045] Place the mobile phones that keep getting numbers on a square flat ground, bury them in an "inclined" manner using concrete bricks, and record the actual latitude and longitude of the phones.

[0046] Specifically, the post-earthquake burial environment is classified into three types: "inclined type", "collapse type" and "composite type". Among them, the "inclined type" burial environment has a large overall living space, the strongest stability, and the highest survival rate of buried personnel. Therefore, the "inclined type" environment is used to collect data.

[0047] In some embodiments of the present application, the collecting of mobile phone information by a drone equipped with a positioning payload includes:

[0048] The positioning payload is installed directly below the drone, and the signal receiving side of the payload is kept facing downward. The remote-controlled drone automatically cruises in a square flat area and flies around the mobile phone. When the mobile phone is scanned, the drone exchanges data with the mobile phone to obtain the mobile phone information.

[0049] In some embodiments of the present application, the mobile phone information includes: mobile phone serial number and mobile phone signal strength;

[0050] The drone recorded information includes: drone altitude, drone longitude and latitude, and mobile phone signal strength.

[0051] Specifically, the serial number of the SIM card in the mobile phone can distinguish different mobile phone cards, and the data will not be confused.

[0052] In some embodiments of the present application, the pre-processing of the drone recorded information during the drone cruise and recording in a table includes:

[0053] The drone record information during the drone cruise is recorded and exported through the remote control, the drone record information is processed in groups, the three groups and the actual longitude and latitude of the mobile phone are combined into a large group, and the recorded data is put into a table to be trained using a computer, wherein the mobile phone signal strength in the data of the three groups is for the same mobile phone.

[0054] In some embodiments of the present application, the large group is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3, actual mobile phone longitude, actual mobile phone latitude).

[0055] Specifically, the three groups plus the actual longitude and latitude of the mobile phone are combined into a large group, which is: drone altitude 1, drone longitude 1, drone latitude 1, mobile phone signal strength 1, drone altitude 2, drone longitude 2, drone latitude 2, mobile phone signal strength 2, drone altitude 3, drone longitude 3, drone latitude 3, mobile phone signal strength 3, actual longitude of the mobile phone, actual latitude of the mobile phone. Each set of data is for one mobile phone. The principle is that the actual location of a mobile phone can be determined through the signal data of three points.

[0056] See also Figure 3 As shown: (x, y) represents the location of the mobile phone, (xi, yi) represents the latitude and longitude of the drone at point i, i = 1, 2, 3; di represents the distance of the drone from the mobile phone at point i. The signal strength ei received by the drone has a mapping relationship with the distance si between the two signal sources. The mapping relationship is recorded as si = a*ei + b, where a and b are conversion coefficients, which will be affected by the signal strength consumption during signal transmission. Through the fitting function of the neural network, a and b can be fitted into specific data, so as to achieve the effect of converting the signal strength ei received by the drone into the distance si. According to the Pythagorean theorem of the triangle, the plane distance di can be obtained through the distance si between the drone and the mobile phone and the height hi of the drone. Finally, the specific position (x, y) of the mobile phone can be obtained through the three-point positioning principle diagram in the above figure. The specific formula is as follows:

[0057] si=a*ei+b

[0058] d 2 +hi 2 =si 2

[0059] (x-x1) 2 +(y-y1) 2 =d1 2

[0060] (x-x2) 2 +(y-y2) 2 =d2 2

[0061] (x-x3) 2 +(y-y3) 2 =d3 2 .

[0062] In some embodiments of the present application, the step of establishing a BP neural network model, initializing the weights and thresholds of the BP neural network model, and substituting the data of each sparrow into the BP neural network model includes:

[0063] The number of input layer nodes of the BP neural network model is n, the number of hidden layer nodes is m, and the number of output layer nodes is p, then the number of weights is N = n*m + m*p, and the number of thresholds is M = m + p;

[0064] Set a sparrows, each sparrow has M+N data and contains all weights and thresholds, and initialize the data of each sparrow.

[0065] Specifically, Figure 4 For the BP neural network diagram, drone altitude 1, drone longitude 1, drone latitude 1, mobile phone signal strength 1, drone altitude 2, drone longitude 2, drone latitude 2, mobile phone signal strength 2, drone altitude 3, drone longitude 3, drone latitude 3, mobile phone signal strength 3, so the input layer is set to 12 nodes, the output is the actual longitude and latitude of the mobile phone, so the output layer is set to 2 nodes, and the number of hidden layer nodes can be set to 5 to 12. The number of input layer nodes is n (here is 12), the number of hidden layer nodes is m, and the output layer is p (here is 2). Then the number of weights is: N = n*m + m*p, and the number of thresholds is: M = m + p. Set a sparrow, each sparrow has M + N data, and contains all weights and thresholds, and initialize the data of each sparrow to maintain randomness.

[0066] In some embodiments of the present application, the collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is used as input to the trained BP neural network model to obtain the final predicted location longitude and latitude, including:

[0067] Each group of collected mobile phone data is arranged and combined into three groups as a row to form a data row of 12 data. The arranged data rows are used as input and input into the trained BP neural network model, where the data behavior is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3).

[0068] Specifically, due to the large amount of training data, the influence of multipath signals can be avoided as much as possible. The training rounds and step size of the neural network can be set by yourself, the loss function is MSE, and the activation function is set to 'tans ig' and 'pure line' to achieve good training results.

[0069] In some embodiments of the present application, the collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is used as input to the trained BP neural network model to obtain the final predicted location longitude and latitude, including:

[0070] The predicted longitude and latitude output by the BP neural network model are sorted by size, the largest and smallest data, a total of 1 / 4, are removed, the middle data are taken for mean processing, and the final predicted position longitude and latitude are obtained.

[0071] Specifically, since the collected data will have multiple outputs after being input into the model through permutations and combinations, and we only need to process the output data for one predicted location, we need to sort the longitude and latitude by size, remove the largest and smallest 1 / 4 data, take the middle data for average processing, and get the final predicted location longitude and latitude. This data processing step is mainly used to remove the influence of signal strength outliers and enhance the model's anti-interference ability.

[0072] In some embodiments of the present application, the loss function of the BP neural network model is MSE, and the activation function is set to tans ig and pure lin.

[0073] In summary, the method of establishing a fitting model by optimizing the neural network with existing data to predict the application data perfectly solves the difficulties of low precision and high cost of the AOA and TDOA methods. It does not cost too much to establish a model through a large amount of data, and the accuracy is also guaranteed by the data fitting method. The positioning method of the sparrow optimization BP neural network model uses a large amount of data verification training to enable the model to cope with most of the situations in the training scenario, and will not cause the positioning to be offset due to the strength of individual special signals. In terms of multipath problems, the method of optimizing the BP neural network model positioning by the sparrow optimizes the data collected before the training model will also have multipath propagation. In the process of fitting the training model with these data, the signal multipath propagation problem is also fitted, so that the model has a good anti-interference ability for the multipath propagation of the signal, which greatly reduces the impact of the multipath propagation of the signal. The output processing of multiple sets of data in the prediction process further handles the interference of the signal and improves the positioning accuracy.

[0074] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0075] The present application is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0076] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A disaster area rescue target positioning method based on the sparrow algorithm to optimize the BP neural network is characterized by: include: Pre-process the mobile phones that are continuously logged in, collect mobile phone information through drones equipped with positioning payloads, pre-process the drone recorded information during the drone cruise, and record it in a table; Establishing a BP neural network model, initializing the weights and thresholds of the BP neural network model, substituting the data of each sparrow into the BP neural network model, inputting the table into the BP neural network model, subtracting the latitude and longitude of the mobile phone output by the BP neural network model from the actual latitude and longitude of the mobile phone as the fitness of the sparrow, running the sparrow optimization algorithm, and iteratively finding the optimal fitness value; The sparrow data corresponding to the optimal fitness value is used as the final weight and threshold of the BP neural network model, the table is used as input to train the BP neural network model, and the trained BP neural network model is put into the UAV remote control so that the UAV remote control processes the received data in real time; Determine the scope of the disaster by remote sensing technology, automatically cruise the drone carrying the positioning payload over the disaster area, collect the mobile phone information, and transmit the mobile phone information and drone record information back to the drone remote controller; Each group of collected mobile phone data is arranged and combined into three groups as a row to form a data row. The arranged data row is used as input and input into the trained BP neural network model to obtain the final predicted location longitude and latitude, and the mobile phone location predicted by the BP neural network model is marked on the map.

2. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 1 is characterized in that: The pre-processing of the mobile phones that continue to register includes: Place the mobile phones that are continuously receiving numbers in a square flat area, bury them in an "inclined" manner using concrete bricks, and record the actual latitude and longitude of the phone.

3. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 2 is characterized in that: The collecting of mobile phone information by using a drone equipped with a positioning payload includes: The positioning payload is installed directly below the drone, and the signal receiving side of the payload is kept facing downward. The remote-controlled drone automatically cruises in a square flat area and flies around the mobile phone. When the mobile phone is scanned, the drone exchanges data with the mobile phone to obtain the mobile phone information.

4. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 3 is characterized in that: The mobile phone information includes: mobile phone serial number and mobile phone signal strength; The drone recorded information includes: drone altitude, drone longitude and latitude, and mobile phone signal strength.

5. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 4 is characterized in that: The preprocessing of the drone recorded information during the drone cruise and recording it in a table includes: The drone record information during the drone cruise is recorded and exported through the remote control, the drone record information is processed in groups, the three groups and the actual longitude and latitude of the mobile phone are combined into a large group, and the recorded data is put into a table to be trained using a computer, wherein the mobile phone signal strength in the data of the three groups is for the same mobile phone.

6. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 5 is characterized in that: The large group is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3, actual mobile phone longitude, actual mobile phone latitude).

7. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 6 is characterized in that: The step of establishing the BP neural network model, initializing the weights and thresholds of the BP neural network model, and substituting the data of each sparrow into the BP neural network model includes: The number of input layer nodes of the BP neural network model is n, the number of hidden layer nodes is m, and the number of output layer nodes is p, then the number of weights is N = n*m + m*p, and the number of thresholds is M = m + p; Set a number of sparrows, each sparrow has M+N data and contains all weights and thresholds, and initialize the data of each sparrow.

8. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 7 is characterized in that: The collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is input into the trained BP neural network model as input to obtain the final predicted location longitude and latitude, including: Each group of collected mobile phone data is arranged and combined into three groups as a row to form a data row of 12 data. The arranged data rows are used as input and input into the trained BP neural network model, where the data behavior is (UAV altitude 1, UAV longitude 1, UAV latitude 1, mobile phone signal strength 1, UAV altitude 2, UAV longitude 2, UAV latitude 2, mobile phone signal strength 2, UAV altitude 3, UAV longitude 3, UAV latitude 3, mobile phone signal strength 3).

9. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 8 is characterized in that: The collected mobile phone data is arranged and combined into three groups per row to form a data row, and the arranged data row is input into the trained BP neural network model as input to obtain the final predicted location longitude and latitude, including: The predicted longitude and latitude output by the BP neural network model are sorted by size, the largest and smallest data, a total of 1 / 4, are removed, the middle data are taken for mean processing, and the final predicted position longitude and latitude are obtained.

10. The disaster area rescue target positioning method based on the sparrow algorithm optimization BP neural network according to claim 9 is characterized in that: The loss function of the BP neural network model is MSE, and the activation function is set to tansig and purelin.

Images