Flying ad-hoc networks （FANET） using directional transmissions face challenges during network initialization due to frequent topology changes and fluctuations in link quality. To address these issues， a sensing-assisted rapid neighbor discovery method is proposed. The main idea of this method is to use a sensing mechanism to obtain neighbor node information， which accelerates network efficiency. In addition， this paper describes the adoption of a new integrated sensing and communication waveform， known as multiple-input multiple-output （MIMO） orthogonal time-frequency space （OTFS），which combats the Doppler effect in fast-changing channels and improves link quality. For FANET scenarios， a multi-target detection technique based on MIMO-OTFS integrated sensing and communication waveforms is studied at the physical layer. This physical layer sensing scheme is then mapped to the upper layer network， and a sensing-assisted efficient neighbor discovery algorithm is designed. Additionally， a integrated sensing and communication mechanism for collaborative discovery of multi-point is proposed. This mechanism indirectly senses and discovers potential targets by interacting with sensing information and neighbor discovery tables among neighboring nodes. The proposed mechanism enhances the initial networking efficiency of FANET. Simulation results demonstrate that the proposed scheme outperforms pure communication protocols. This leads to a significant reduction in the initial networking time of FANET， an increase in target sensing accuracy， and an overall improvement in network performance.

Multi-rotor drones are increasingly becoming a threat to critical facilities while bringing convenience. This article focuses on the difficulty of tracking and detecting low slow small unmanned aerial vehicles （UAVs）， and develops a UAV tracking， detection， recognition， and positioning system based on microphone array sound source localization. The system acquires UAV noise signals in real time， and then uses a wavelet transform-based beamforming time-frequency sound source localization algorithm for UAV noise source localization， and finally integrates the position of the multi-rotor UAV obtained by inversion of the microphone array and the image captured by the video. Tests conducted in a fully anechoic room have shown that the positioning system can accurately track the lateral flight and up-and-down flight of drones， accurately detect the position of drones even in blind spots of cameras， and effectively compensate for video detection. And the positioning accuracy of the system is 92.2%， which is better than that of similar systems.

Multi-agent reinforcement learning algorithms fail to form effective collaborative policy under partially observable conditions. In view of this problem， a policy transfer reinforcement learning method based on centralized training and decentralized execution （CTDE） paradigm was proposed. Firstly， under global observation， the teacher module was trained to explore good collaborative policy. Then， under partially observable conditions， the student module was trained online with the expectation of maximizing cumulative returns as the objective function， and at the same time， policy distillation techniques were used to transfer policy from the teacher module and adaptively adjust the proportion of teacher policy affecting student policy. Finally， the proposed method was verified by simulation in multiple map scenarios. The experimental results show that under partially observable conditions， the success rate of student modules is higher than that of the baseline algorithms. The research results can be applied to multi-agent collaborative tasks， improving the collaborative performance of agents in decentralized execution.

In order to satisfy impact angle constraints and reduce energy cost under salvo attacks against high-speed non-maneuvering targets， a three-dimensional coupled model is established， and a 3D energy optimal guidance law for non-linear systems with large constrained impact angles is proposed. The reasons for the low accuracy of 3D decoupled model are discussed， and a 3D coupled model is established. Based on this model， the appropriate control variables are chosen， and the 3D energy optimal guidance law under large impact angle constraints is derived， based on minimum principle. Experimental results indicate that the performance of the proposed coupled model on miss distance， impact angle error and energy cost has improved 0.7 m，1.1°，200 g compared with decoupled model. It can prove the advantages of the guidance law on control precision and energy cost.