Illegal Radio Location And Tracking Based On UAV Platform

Black broadcast(i.e.illegal radio)frequently occurs,illegally occupying normal broadcasting frequency channel.Its disrupt public order and civil aviation flight,seriously jeopardize that security of national communications and affects social stability.At present,however,the main method of discharging illegal radio is to combined the monitor vehicle with handheld equipment.This method is seriously affected by the multipath of ground and complex traffic environment lead to its slow speed and poor precision positioning.Unmanned aerial vehicle(UAV),on the other hand,because of its excellent characteristics such as communication,autonomous navigation and video capture attracted much attention.It has been successfully used in agriculture,transportation,aerial photography and power inspection.Therefore,the paper considers building an effective positioning and navigation model based on uav platform and using the theory of reinforcement learning and convex optimization design the algorithm.Finally,the paper realizes rapid and accurate illegal radio positioning and tracking.On the basis of fully understanding the research on illegal radio at home and abroad,the problem of illegal radio location and tracking under the platform of UAV is studied in the paper.Combined with the advantages of UAV,an effective mathematical positioning model is established and optimized.The main research work and innovation points of this paper are as follows:(1)An illegal radio localization model based on UAV platform is proposed.To solve the problem that the traditional illegal radio positioning method is seriously affected by the complex traffic environment and multipath fading,an reinforcement learning illegal radio model based on UAV platform is proposed.Compared with the traditional illegal radio localization method,the aerial flying of UAV can avoid the complex traffic environment and multipath fading.At the same time,the reinforcement learning algorithm is used to control the UAV to explore and learn in the unknown environment,which can realize the autonomous positioning and navigation of the UAV to illegal radio,greatly reducing the human and material costs.(2)A directional Q-learning algorithm based on directional antenna is proposed.It is not cooperation mode for that UAV to receive illegal radio information,the receive antenna can only acquire a received signal strength value,so the exploration process of reinforcement learning under the influence of noise will be longer.Considering the cost of information acquisition and the demand of real-time positioning,on the basis of(1),the reinforcement learning algorithm is redesigned and the directional antenna is introduced to obtain multiple received signal strength information.According to the signal strength information received in different directions,the reward are modified and the direction selection strategy is optimized,which effectively inhibits the impact of noise and reduces the number of explorations,enabling the UAV to locate illegal radio faster.(3)A localization model based on cutting plane algorithm is proposed.The positioning model based on the enhanced learning algorithm needs continuous exploration and learning.The flying distance of the UAV is limited,and the signal strength information received by the UAV is weak when it is far away from the illegal radio.The UAV needs more exploration process and even hovers.Therefore,illegal radio location over long distances is very important.In the paper,a long-distance localization model based on the cutting plane algorithm is proposed.By taking the directional angle of directional antenna as the cutting plane and looking for the analytic center of the constraint region as the anchor point,and then iteratively updating the anchor point,long-distance illegal radio localization can be realized.It can quickly locate that illegal radio by use the cut plane algorithm to search for the analytic center of the restriction area,and the position of the illegal radio can be approached through fewer iterations.