In indoor anti-terrorism combats,a single-mode unmanned aerial vehicle (UAV) cannot adapt to complex terrain and is easily destroyed.Aiming at this problem,a small UAV which can bounce on the ground-Scyther, is proposed.The composition,structure and main parameters are firstly given.Then,the main algorithms are explained,and the feasibility of the scheme is analyzed.A combination of coaxial twin-rotor and jumping UAV are used on Scyther.When approaching the target,the coaxial twin-rotor pops up from the main body and sticks to the high position.The camera mounted on it provides a monitoring field of view.The attitude of UAV is controlled through the momentum wheel,whose path is planned using improved ant colony algorithm.Using distributed context-aware algorithm,the UAV’s trajectory will be difficult to predict within the enemy’s visible range,thereby increasing the probability of getting close to the target.Compared with traditional UAVs,Scyther can adapt to a variety of complex terrains,can be controlled remotely, and has the advantages of small size,fast speed,low power consumption,and strong maneuverability.It is suitable for both single and cluster combats,thus is very beneficial for indoor anti-terrorism combats.In addition,Scyther can be equipped with a variety of warheads for different mission requirements,and is applicable for extensive scenarios.

A clustered reconnaissance and defense system is designed,which can be used in military reconnaissance and civil disaster relief detection.The cubic unit is the core of the system hardware.The subsystems including fixed-wing UAV(unmanned aerial vehicle),four-rotor UAV and UAV are integrated in the cubic unit system.The working background is set,and the function and the working flow of the system are designed.A four-level communication network is designed to realize the information exchanges between UAV subsystems,collaborative communications between cubic unit systems,and information exchanges with remote command center.To realize the detection and defense tasks,a three-level command and control network is designed to implement intelligent collaborative management of tasks,data processing,and task execution control.The system can realize the function of detection and the function of regional communication relay,which can play an important role in both civil disaster relief and military war area detection and defense.

Aiming at the cruise segment trajectory prediction problem of near space reentry gliding vehicle, a trajectory prediction algorithm based on control law prediction was proposed. By analyzing and deriving the unknowns in the trajectory prediction and studying the flight modes and control laws of the cruising segment, the feasibility of trajectory prediction was clarified. The discriminating methods of two kinds of flight trajectories are studied: balanced gliding and jumping gliding. On this basis, a trajectory prediction algorithm based on control law prediction was designed. The algorithm was verified by simulation. The influence of tracking time and tracking starting point on trajectory prediction accuracy was studied. The results showed that the algorithm had high accuracy for different types of trajectories; the tracking time had a great influence on trajectory prediction accuracy; but the impact of tracking start point was small.

In view of the incomplete evaluation of the underlying indicators and the operational action plan, based on evidence reasoning method, starting from the required capability of operational action, a belief rule base is built, and the multiple types of uncertainty data are converted into a unified belief structure. The results of the evaluation of the operational action plan are obtained by means of evidence reasoning.

Aiming at the problem that the traditional algorithm cannot initiate the slow and small target (SST) track effectively, a track initiation method for SST based on image processing Hough transform is proposed. The parameters of the minimum circumscribed rectangle, consisting of F₃, F₂ and F₁, are introduced as the weighting factors of Hough transform, and these parameters are extracted in accumulation region by image marking technique. SST is successfully obtained by controlling the weighting factor of Hough transform, which improves the detection performance of the radar to the SST. Simulation results verify the effectiveness of the proposed algorithm.