| The technology of drone is growing rapidly,which brings more convenience to the society.And at the same time,the drone brings danger to the jail,airport,and many other gathering.With the increasing of control distance,and easier operation,the surveillance demand of drone is stronger.RF based drone detection method is a good method of drone detection,it detect the drone by signal between drone and its remote control.In this paper,we develop a detection method to solve the defect of present drone detection method,and is aimed at detect the DJI drone which account for over 75%share of commercial drone market.A complete drone detection system is designed in this paper.The main work of this paper can be summarized as follows:(1)Drone signal collector based on FPGA and DDR3 SDRAM.Due to high speed of RF signal,to get the continuous sample data of drone,a collector based on cache forwarding mechanism is designed.The signal of drone is received by antenna array,and sampled by multi-channel AD with sample rate of 250M.Then the data is written into DDR3 SDRAM.When the SDRAM is full,the data inside SDRAM is read and transferred through ethernet network to the computer,so finally the continuous data of drone RF signal with large capacity is received by antenna array is available in computer.The system need the design of connector of optical fiber,DDR3 SDRAM as well as the constraint of FPGA.(2)Design of drone detection method based on square wave match,which is the RF feature of DJI drones.To enhance the square feature of remote drone,the method of feature extraction,multi-resolution analysis is designed,and finally the detection probability and distance is raised.Besides the single channel enhancement method,the beamforming method is developed to raise the SNR ratio in certain direction,which is the weighted sum of signal received by antennas in antenna array.When the beams can cover 360 degree,the distance of detection can be raised further,and finally the drone can be detected within 600m by experiment.(3)Design of drone detection system.The detection algorithm need s large amount of data of drone RF signal,and in order to implement the detection algorithm,a FPGA is needed to implement the algorithm as a pipeline logic.The work can be summarized as RTL coding,allocating of logic resources,and pipeline expansion.In final system,after transmitted to FPGA by AD,the RF signal is processed by the beamforming module,the FFT IP core,the feature extraction module and so on,a bool value which represent the detection result comes out.And in practice,the resource consumption of our logic is within normal range. |