Publication

Abstract : The growing accessibility of UAVs has sparked interest in deploying drones for multiple purposes. This created an alarming situation for the unauthorized use of drones. To address this concern, this work focuses on the accurate positioning of unauthorized drones with the help of patrolling Unmanned Aerial Vehicles (UAVs) equipped with radar systems via adaptive Intelligent Reflective Surfaces (IRS). A two-stage framework is proposed for significantly enhancing the accurate positioning of the invader drone, considering practical constraints such as quivering and mutual coupling (MC) effects. First stage encompasses estimation and optimization of the parameters such as Doppler frequency, Direction of Arrival (DoA), and time delay using the hybrid SAGE-Cuckoo Search Algorithm. The second stage comprises a Back Propagation Neural Network (BPNN), which is employed as a feedback mechanism to reduce error from quivering and MC effects. In this, the IRS plays a significant role by adapting the configuration of its elements, such as Uniform Linear Array (ULA), Uniform Rectangular Array (URA), and Uniform Circular Array (UCA). This empowers substantial control over the signal propagation, offering diverse angular and path configurations, which directly affect the estimated parameters for localization. Simulation results highlight the robustness and the superiority of the proposed method as Root Mean Square Error (RMSE) concerning the SNR is reduced upto 24.20%, 41.27%, 56.52%; 16.43%, 30.19%, 44.44% for Doppler frequency, 87.22%, 89.66%, 91.3%; 71.1%, 74.67%, 78.65% for DoA and 25.93%, 37.44%, 51.29%; 10.79%, 21.96%, 37.96% for time delay estimation compared to the MUSIC and SAGE algorithm, at SNR − 30dB, 0 dB, and 30 dB, respectively. Similarly, the RMSE with respect to the relative time delay is reduced upto 60.85%; 63.90%; 66.07% and 50.21%, 53.45% 55.81% for Doppler Frequency estimation, 50.19%, 83.82%, 84.30% and 34.31%, 71.43%, 72.16% for DoA estimation and 64%, 71.43%, 75% and 52.63%, 60%, 64.29% for Time Delay estimation compared to the MUSIC and SAGE algorithm at 20 ns, 40 ns, and 60 ns relative time delay, respectively and demonstrating the impact of various practical aspects on IRS configurations. This framework validates promising potential for autonomous drone detection applications in security, enhancing system consistency and operational efficacy through automation.