Small unmanned aerial vehicle frequency spectrum identification and direction finding system and method

Abstract

The invention belongs to the technical field of unmanned aerial vehicle detection countering, and discloses a spectrum identification and direction finding system and method for a small unmanned aerial vehicle. The system comprises an antenna array element, a radio frequency switch, a data acquisition card, a CPU processing mainboard and an upper computer. According to the algorithm of the CPU processing mainboard, for unmanned aerial vehicle model identification and azimuth direction finding, based on a neural network algorithm model and an array structure model of a uniform circular array, rapid and accurate classification of different radio frequency signal types is realized through time-frequency diagram analysis, recessive advanced features of signals can be extracted from large-scale data, and a result can be directly output. And the unmanned aerial vehicle model is quickly identified. A DOA estimation algorithm based on deep learning is adopted, a DOA estimation value can be quickly given through an algorithm model based on covariance matrix decomposition, a final result is obtained, and an azimuth angle can be accurately given. The defect that a traditional algorithm is sensitive to the multi-path effect is overcome, and high-precision direction estimation can be achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle detection and countermeasure, and in particular relates to a spectrum recognition and direction finding system and method for a small unmanned aerial vehicle. Background Art

[0002] In recent years, with the rapid development of consumer drones, they have brought great convenience to society due to their small size, low cost and simple operation. At the same time, they have also brought many safety issues, and the number of drone "illegal flights" has also increased. Therefore, it is extremely important to adopt a reasonable and effective control measure to prevent drone "illegal flights".

[0003] Drone spectrum identification and direction finding technology is the key to monitoring "illegal" drone flights. Although drone manufacturers also restrict illegal drone flights by setting no-fly zones, areas that require temporary no-fly zones or involve state secrets are not included in the no-fly zone library by drone manufacturers, resulting in no flight restrictions in these areas. Therefore, the use of third-party equipment for drone detection and direction finding is of great practical significance. The detection technology uses a spectrum identification and classification algorithm to identify the spectrum of drone models, while the direction finding technology uses a DOA estimation algorithm to monitor the azimuth of the drone in real time, and finally provides an effective target for the jamming equipment.

[0004] The traditional method of detecting drones mainly involves first collecting drone signals, analyzing the drone spectrum with FFT, then extracting eigenvalues, saving them to a feature library file, and then comparing the eigenvalues ​​of the collected signals with the feature library to obtain the drone model. This method is not only slow in building a drone feature library, but the diversity of signals will make it difficult to extract drone signals, and the false alarm rate is high in complex electromagnetic environments, which can no longer meet the needs of actual applications. The direction finding technology used in the past was mainly amplitude comparison direction finding or phase comparison direction finding, which inferred the direction of the signal source based on the signal amplitude difference or phase difference output by each element of the direction finding antenna array receiving the incoming signal. This method also has some limitations. For example, the side blocked by the mountain may be affected by the multipath effect, and the signal propagation will be reflected, refracted and scattered, resulting in different paths and times for the same signal source to reach each element, thereby affecting the accuracy of the direction finding results. Summary of the invention

[0005] The purpose of the present invention is to provide a spectrum identification and direction finding system and method for a small UAV in order to solve the above problems. It is expected to improve the existing UAV detection methods, which have poor accuracy and cannot adapt to complex environments.

[0006] The technical solution adopted by the present invention is as follows: a spectrum identification and direction-finding system for a small unmanned aerial vehicle, the system comprising a radio frequency antenna array element, a radio frequency switching switch, a signal processing unit, an Ethernet interface and a host computer; the antenna array element receives a radio frequency signal, the radio frequency signal is switched and selected by the radio frequency switching switch, and the radio frequency signal is sent to the signal processing unit for signal processing operation; the operation result is transmitted to the host computer through the Ethernet interface; the host computer is used to display the result data, including the center frequency band, bandwidth, power value, model and azimuth of the unmanned aerial vehicle; wherein the signal processing operation includes low-pass filtering, signal preprocessing, and algorithm analysis, and the obtained operation result includes the model type and azimuth of the unmanned aerial vehicle, as well as signal characteristic parameters.

[0007] Furthermore, the antenna array elements include a 5-element antenna for the low frequency band and an 8-element antenna for the high frequency band, and the antenna array elements are divided into two areas, A and B; Area A includes elements 1, 2, 3, and 4 of the 8 elements and elements 9, 10, and 11 of the 5 elements, which are switched by an 8-to-1 matrix switch No. 1, and the signal processing unit is responsible for data collection; Area B includes elements 5, 6, 7, and 8 of the 8 elements and elements 12 and 13 of the 5 elements, which are switched by an 8-to-1 matrix switch No. 2, and the signal processing unit is responsible for data collection; the antenna array elements in areas A and B work independently and are selected by their respective 8-to-1 matrix switches; and the antenna array elements are connected to a WiFi processing board, and the WiFi processing board uploads the data.

[0008] Furthermore, in order to improve the timeliness of the system, two AD9361 RF signal collectors operate independently; the signal processing unit includes an AD9361 RF signal collector, an FPGA preprocessor, and a CPU algorithm analyzer; the RF signal enters the AD9361 RF signal collector, and the AD9361 RF signal collector performs parameter configuration and data collection of the AD9361, and preprocesses the IQ data to convert the RF signal into a digital IQ signal; the AD9361 RF signal collector is connected to the FPGA preprocessor through a high-speed interface, and the FPGA preprocessor performs signal judgment, classification, and signal preprocessing after receiving the signal; the FPGA preprocessor is connected to the CPU algorithm analyzer through a gigabit switch, and the signal is uploaded to the CPU algorithm analyzer through the gigabit switch, and the CPU algorithm analyzer performs data fusion and algorithm processing.

[0009] Furthermore, in order to improve the accuracy of recognition, the algorithm processing includes a spectrum recognition signal classification algorithm and a DOA direction finding estimation algorithm.

[0010] Furthermore, the data is processed by a CPU algorithm analyzer to obtain more accurate data parameters; the CPU algorithm analyzer parses the data, embeds the algorithm API interface, and uploads the data in an integrated manner; the data parsing process includes parsing the electronic compass data, GPS module data, and WiFi module data, and the parsed results and the calculated results are packaged and transmitted to the host computer; wherein the parsing and calculation results include azimuth data, drone model, and characteristic parameters of the signal.

[0011] Furthermore, the AD9361 radio frequency signal collector includes an acquisition card 1 and an acquisition card 2, which are used to respectively acquire data. The front ends of the acquisition cards 1 and 2 are connected to the radio frequency switching switch, and the back ends of the acquisition cards 1 and 2 are connected to the gigabit switch.

[0012] Furthermore, the host computer is connected to a CPU mainboard, which performs data processing and data fusion; and the CPU mainboard is connected to an electronic compass and a GPS module, which transmit data to the CPU mainboard for data processing and fusion; the data processed by the CPU mainboard is transmitted to the host computer for display.

[0013] In order to better identify the type and azimuth of drones, a spectrum recognition and direction finding method for small drones is proposed. The method includes:

[0014] S1: The host computer sends the monitoring frequency band, divides it by frequency band strategy, calculates the bandwidth, number of segments and center frequency, and then selects the antenna working mode according to the sent monitoring frequency band;

[0015] S2: Data acquisition, setting the parameters of the AD9361 RF signal collector, including acquisition frequency, bandwidth, sampling rate parameters, and performing data acquisition through polling switching of the matrix switch;

[0016] S3: Data analysis, the data is divided into three paths for analysis. One path is to perform FFT preprocessing on the original IQ data collected by the AD9361 RF signal collector to obtain signal characteristic parameter values, including center frequency, bandwidth, and power value; the collected characteristic parameters are packaged and uploaded; the other path is to slice, perform FFT, and calculate the amplitude of the original IQ data according to the algorithm requirements, and participate in the spectrum recognition algorithm; the last path is to add digital filters to the center frequency bands of the mixed IQ data with the obtained signal characteristic parameters to obtain different original IQ data, that is, signal classification, and finally package and upload to participate in the DOA direction finding estimation algorithm;

[0017] S4: upload the three-way data in S3 to the CPU mainboard through the Gigabit switch, and perform data fusion processing on the data of acquisition card 1 and acquisition card 2 in the AD9361 RF signal collector; then perform algorithm analysis, load the original IQ data for DOA estimation algorithm analysis, and obtain the direction finding angle; load the spectrum recognition algorithm, remove the WiFi signal and obtain the drone model;

[0018] S5: The angles and drone models measured in S4, as well as the characteristic parameters of the drone, are integrated and uploaded to the host computer for interface display.

[0019] Furthermore, the contents displayed by the host computer include warning list display, spectrum diagram, waterfall diagram display, direction finding angle display, equipment positioning, and automatic north calibration.

[0020] Furthermore, the working mode includes working mode 1 and working mode 2. Working mode 1 is to switch the switch to select the antenna of the high frequency band, and working mode 2 is to switch the switch to select the antenna of the low frequency band and the high frequency band.

[0021] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0022] 1. By judging and classifying the collected data and integrating algorithm processing technology, the specific model of the drone can be accurately identified and the azimuth data of the drone can be calculated, which is convenient for accurately locating the position of the drone and making judgments.

[0023] 2. The system includes antenna array elements, RF switches, data acquisition cards, CPU processing motherboards and host computers. In the algorithm of the CPU processing motherboard, the neural network algorithm model and the array structure model of the uniform circular array are respectively used for drone model identification and azimuth direction finding; the drone RF signal classification system based on the convolutional neural network can realize the rapid and accurate classification of different RF signal types through time-frequency diagram analysis, and can extract the implicit high-level features of the signal from large-scale data, directly output the results, and quickly identify the drone model. This model not only significantly improves the accuracy of recognition, but also can adapt to various complex environments and drone types, thereby improving the applicability of this system. Another direction finding algorithm upgrades the traditional direction finding algorithm mode and adopts a DOA estimation algorithm based on deep learning. Through the algorithm model based on covariance matrix decomposition, it can quickly give the DOA estimation value, obtain the final result, and accurately give the azimuth. It avoids the shortcomings of the traditional algorithm that is sensitive to multipath effects and can achieve high-precision direction estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the overall structure of the system of the present invention;

[0025] Figure 2 It is a system function framework diagram of the present invention;

[0026] Figure 3 It is a detailed structural diagram of the system of the present invention;

[0027] Figure 4 It is the functional block diagram of the CPU mainboard of the present invention;

[0028] Figure 5 The method flow of the present invention Figure 1 ;

[0029] Figure 6 The method flow of the present invention Figure 2 ; DETAILED DESCRIPTION

[0030] The present invention will be described in detail below in conjunction with the accompanying drawings.

[0031] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0032] like Figures 1 to 4 As shown in the figure, the system includes RF antenna array elements, RF switching switches, signal processing units, Ethernet interfaces and host computers; in normal working mode, after the radio signal is received by the RF antenna array elements, it is selected through the RF switching switch and sent to the signal processing unit for signal processing operations. The signal processing unit mainly includes AD9361 RF signal collector, FPGA preprocessor, CPU algorithm analyzer, and finally transmits the processing results to the host computer through the Ethernet interface. In order to ensure the accuracy of the data, the real-time processing of the data requires high timeliness, and the data acquisition link and data calculation and processing link need to control the timeliness.

[0033] The entire system consists of several functional modules from top to bottom: antenna area, array element logic selection, data acquisition, processing, transmission, algorithm processing, data fusion, and software function display. The functional block diagram is as follows: Figure 2 As shown:

[0034] The antenna area consists of a 5-element antenna for the low frequency band and an 8-element antenna for the high frequency band, and is divided into two areas, A and B.

[0035] Area A is mainly composed of 8-element array elements No. 1, 2, 3, and 4 and 5-element array elements No. 9, 10, and 11. It is mainly switched by No. 1 8-to-1 matrix switch, and acquisition card 1 is responsible for data acquisition.

[0036] Area B is mainly composed of 8-element array elements No. 5, 6, 7, and 8 and 5-element array elements No. 12 and 13. It is mainly switched by No. 2 8-to-1 matrix switch, and acquisition card 2 is responsible for data acquisition.

[0037] The antenna array elements in areas A and B work independently without interfering with each other, and are selected by a special matrix switch. A special acquisition card is responsible for data acquisition, processing and transmission. Finally, the signals in areas A and B are uploaded to the CPU motherboard through the Gigabit switch. The CPU motherboard performs data fusion and algorithm processing, and sends the processing results to the host computer for display.

[0038] Among them, the 2.4GHz and 5.8GHz omnidirectional antennas are connected to the WiFi processing board, and the WiFi processing board uploads the received WiFi channel data.

[0039] like Figure 3 As shown, the signal direction is from left to right. The whole equipment includes multi-element antenna array elements, RF switching switches, AD9361 acquisition card, Gigabit switch, CPU motherboard, and host computer software.

[0040] Its main working process is as follows: after the RF signal is received by the antenna array element, it is selected by the RF switch polling switch and sent to the AD9361 acquisition card. The acquisition card converts it into a digital IQ signal through low-pass filtering, acquisition, and digital processing, and sends it to the FPGA preprocessor through the high-speed interface. The FPGA preprocessor receives the signal and performs signal judgment, signal classification, and signal preprocessing, etc. The processed data is uploaded through the gigabit switch, and the CPU mainboard participates in data fusion and algorithm processing. The algorithms mainly include spectrum recognition signal classification algorithm and DOA estimation algorithm. Finally, all calculation results are sent to the host computer through the Ethernet interface.

[0041] The auxiliary workflow also includes GPS module data analysis, electronic compass data analysis and WiFi module board data analysis, which are transmitted to the host computer together with the CPU calculation results through the Ethernet interface.

[0042] Among them, the spectrum recognition signal classification algorithm includes the following processes: first define the problem to be solved, the collector collects data, pre-processes the collected data, then selects the model architecture, trains the selected model, and sends the data to the model for testing and evaluation. When the performance indicators are met, deployment and testing are carried out. If the performance indicators are not met, the data is re-verified and adjusted, and then tested and evaluated until the data meets the performance indicators.

[0043] The DOA estimation algorithm includes two processes: training and recognition. The output data of the two processes are sent to the weight model to calculate more accurate azimuth data through the DOA estimation algorithm.

[0044] like Figure 5-6 As shown in the figure, a spectrum identification and direction finding method for a small UAV is provided, and its working process includes the following: first, the host computer sends down the monitoring frequency band, divides the frequency band strategy, calculates the bandwidth, number of segments and center frequency, and then selects the working mode of the antenna according to the sent monitoring frequency band. Working mode 1 is to switch the switch to select the antenna of the high frequency band, and working mode 2 is to switch the switch to select the antenna of the low frequency band and the high frequency band.

[0045] Data acquisition, set the parameters of the acquisition card, including acquisition frequency, bandwidth, sampling rate and other parameters, and perform data acquisition through polling switching of the matrix switch.

[0046] The collected data is analyzed. One of the ways is to pre-process the original IQ data collected by the collector by FFT to obtain parameters, including center frequency, bandwidth, power value, etc., and then package and upload the collected characteristic parameters. Another way is to slice, perform FFT, and calculate the amplitude of the original IQ data according to the requirements, and participate in the spectrum recognition algorithm. The last way is to add digital filters to the center frequency band of the mixed IQ data with the obtained signal characteristic parameters to obtain different original IQ data, that is, signal classification, and finally package and upload to participate in the DOA direction finding estimation algorithm;

[0047] The three channels of data collected above are uploaded to the CPU mainboard through the Gigabit switch. The front ends of acquisition cards 1 and 2 are connected to the RF switch, and the back ends of acquisition cards 1 and 2 are connected to the Gigabit switch. The data of acquisition cards 1 and 2 are fused. Then, the algorithm analysis is performed, the original IQ data is loaded for DOA estimation algorithm analysis, and the direction finding angle is obtained; the spectrum recognition algorithm is loaded to eliminate the WiFi signal and obtain the drone model.

[0048] Finally, the measured angles, drone models, and characteristic parameters of the drone are integrated and uploaded to the host computer for interface display, including warning list display, spectrum diagram, waterfall diagram display, direction finding angle display, equipment positioning, automatic northing, etc.

[0049] In the prior art, the detection distance of this embodiment can reach 3 kilometers, or even 5 kilometers, and the system is easy to operate, has low cost, is more adaptable, and is easy to promote and use on a large scale.

[0050] 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 and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A spectrum identification and direction finding system for a small UAV, characterized in that: The system includes a radio frequency antenna array element, a radio frequency switch, a signal processing unit, an Ethernet interface and a host computer; The antenna array element receives the RF signal, which is switched and selected by the RF switch and sent to the signal processing unit for signal processing operation. The operation result is transmitted to the host computer through the Ethernet interface. The host computer is used to display the result data, including the center frequency band, bandwidth, power value, UAV model and azimuth of the UAV. Among them, the signal processing operations include low-pass filtering, signal preprocessing, and algorithm analysis. The calculation results obtained include the model type and azimuth of the drone, as well as signal characteristic parameters.

2. A small UAV spectrum identification and direction finding system according to claim 1, characterized in that: The antenna array elements include a 5-element antenna in the low frequency band and an 8-element antenna in the high frequency band, and the antenna array elements are divided into two areas, A and B; Area A includes elements No. 1, 2, 3, and 4 of the 8-element array and elements No. 9, 10, and 11 of the 5-element array, which are switched by a No. 1 8-to-1 matrix switch, and the signal processing unit is responsible for data acquisition; Area B includes elements No. 5, 6, 7, and 8 of the 8-element array and elements No. 12 and 13 of the 5-element array, which are switched by a No. 2 8-to-1 matrix switch, and the signal processing unit is responsible for data acquisition; The antenna elements in areas A and B work independently and are selected by their respective 8-to-1 matrix switches; The antenna array element is connected to the WiFi processing board, and the WiFi processing board uploads the data.

3. A small UAV spectrum identification and direction finding system according to claim 1, characterized in that: The signal processing unit includes an AD9361 radio frequency signal collector, an FPGA preprocessor, and a CPU algorithm analyzer; the radio frequency signal enters the AD9361 radio frequency signal collector, and the AD9361 radio frequency signal collector performs parameter configuration and data collection of the AD9361, and preprocesses the IQ data to convert the radio frequency signal into a digital IQ signal; the AD9361 radio frequency signal collector is connected to the FPGA preprocessor through a high-speed interface, and the FPGA preprocessor performs signal judgment, classification, and signal preprocessing after receiving the signal; the FPGA preprocessor is connected to the CPU algorithm analyzer through a gigabit switch, and the signal is uploaded to the CPU algorithm analyzer through the gigabit switch, and the CPU algorithm analyzer performs data fusion and algorithm processing.

4. A small UAV spectrum identification and direction finding system according to claim 3, characterized in that: The algorithm processing includes a spectrum recognition signal classification algorithm and a DOA direction finding estimation algorithm.

5. A small UAV spectrum identification and direction finding system according to claim 3, characterized in that: The CPU algorithm analyzer performs data analysis, embeds the algorithm API interface and integrates and uploads the data; Data analysis and processing includes analyzing the electronic compass data, GPS module data, and WiFi module data, and integrating and packaging the analyzed results and the calculated results and transmitting them to the host computer; Among them, the analysis and calculation results include azimuth data, drone model and characteristic parameters of the signal.

6. A small UAV spectrum identification and direction finding system according to claim 3, characterized in that: The AD9361 radio frequency signal collector includes an acquisition card 1 and an acquisition card 2, which are used to respectively acquire data. The front ends of the acquisition cards 1 and 2 are connected to the radio frequency switching switch, and the back ends of the acquisition cards 1 and 2 are connected to the gigabit switch.

7. A small UAV spectrum identification and direction finding system according to claim 1, characterized in that: The host computer is connected to the CPU mainboard, which performs data processing and data fusion; the CPU mainboard is connected to the electronic compass and GPS module, which transmit data to the CPU mainboard for data processing and fusion; the data processed by the CPU mainboard is transmitted to the host computer for display.

8. A spectrum identification and direction finding method for a small UAV, according to a spectrum identification and direction finding system for a small UAV according to any one of claims 1 to 7, characterized in that: The method comprises: S1: The host computer sends the monitoring frequency band, divides it by frequency band strategy, calculates the bandwidth, number of segments and center frequency, and then selects the antenna working mode according to the sent monitoring frequency band; S2: Data acquisition, setting the parameters of the AD9361 RF signal collector, including acquisition frequency, bandwidth, sampling rate parameters, and performing data acquisition through polling switching of the matrix switch; S3: Data analysis, the data is divided into three paths for analysis. One path is to perform FFT preprocessing on the original IQ data collected by the AD9361 RF signal collector to obtain signal characteristic parameter values, including center frequency, bandwidth, and power value; the collected characteristic parameters are packaged and uploaded; the other path is to slice, perform FFT, and calculate the amplitude of the original IQ data according to the algorithm requirements, and participate in the spectrum recognition algorithm; the last path is to add digital filters to the center frequency bands of the mixed IQ data with the obtained signal characteristic parameters to obtain different original IQ data, that is, signal classification, and finally package and upload to participate in the DOA direction finding estimation algorithm; S4: upload the three-way data in S3 to the CPU mainboard through the Gigabit switch, and perform data fusion processing on the data of acquisition card 1 and acquisition card 2 in the AD9361 RF signal collector; then perform algorithm analysis, load the original IQ data for DOA estimation algorithm analysis, and obtain the direction finding angle; load the spectrum recognition algorithm, remove the WiFi signal and obtain the drone model; S5: The angles and drone models measured in S4, as well as the characteristic parameters of the drone, are integrated and uploaded to the host computer for interface display.

9. A small UAV spectrum identification and direction finding method according to claim 8, characterized in that: The contents displayed on the host computer include warning list display, spectrum diagram, waterfall diagram display, direction finding angle display, equipment positioning, and automatic north calibration.

10. A small UAV spectrum identification and direction finding method according to claim 8, characterized in that: The working modes include working mode 1 and working mode 2. Working mode 1 is to switch the switch to select the antenna of the high frequency band, and working mode 2 is to switch the switch to select the antenna of the low frequency band and the high frequency band.

Images