A Study On UAV Detection And 3D Localization Based On CSI

In recent years,the drone industry has developed rapidly and there are more and more drones on the market.On the one hand,the rapid rise of the drone industry has brought a lot of convenience and pleasure to our daily lives;on the other hand,the illegal use of drones has also brought us many problems and losses,which seriously disrupts social security and people's lives.Therefore,it is extremely important to detect drones in places with high security requirements such as airports and private places.In addition,in many cases,2D plane localization can no longer meet our needs,and the demand for 3D positioning is becoming increasingly urgent.It can be imagined that in a tall building,if 3D localization can be performed,the cost of our search will be greatly reduced.In order to solve these problems,this paper proposes a drone detection system based on RF signal physical layer information and an AoA-based feature-points constrained PSO Localization algorithm.The drone detection system collects the communication signals of the drone and the controller,and then analyzes the drone's movement characteristics including drone's mobility,spatiality,and vibration which is useful for detecting the drones.For 3D localization,this paper analyzes the time complexity of 3D localization.The results show that if the 2D localization algorithm is directly applied to 3D localization,the time complexity will increase exponentially with the number of AP nodes participating in the positioning.This paper proposes an AoA-based feature point constrained PSO 3D localization algorithm.The algorithm converts the 3D localization problem into an optimization problem,and then calculates the feature points to constrain the search space,and finally use the PSO algorithm to find the global optimal solution,i.e.,the position of the target node..This paper conducts measured experiments to verify the performance of the drone detection system.The experimental results show that our drone detection system can achieve a precision of 87.3% and a recall of 85.8%.For the 3D localization algorithm,we use the Matlab tool for simulation experiments.The experimental results show that our proposed 3D positioning algorithm performs well and the average positioning accuracy can reach 0.7 meters.