UAV Tracking/avoiding Airborne Target Real Time Motion Guidance Method Research

Realizing autonomous control of UAVS is an important development trend of UAVS.It requires UAVS to have the ability of autonomous target perception,tracking and avoiding in order to making sure that its safe flight and executing task successfully.UAV air target tracking and avoiding have the same technology needs to airborne guidance and control system.First both of them need to get the measurement of the target using onboard sensors and estimate the state of the target.Then guidance the motion of UAV and control it to track/avoid the airborne target.This dissertation integrated these two issues into the same framework.Two key aspects of these two missions are researched,in terms of target state estimation and the motion guidance of the UAV.The main work and contributions of this dissertation are summarized as follows:Firstly,a solving framework based the way of guidance law to the two problems is presented.UAV autonomously collision avoidance and target tracking can respectively be seen as “sense and avoid” and “sense and tracking” process.The factors that influence the two tasks and the basic characteristics in the problems are first analyzed in detail.Meanwhile the mathematical model including UAV platforms,targets kinematic,and airborne sensors measurement are formalized respectively.During the two tasks,UAV needs to act in real time and both of the UAV and the target may be high maneuver.So,we chose the guidance law way as the main motion guidance in fine time of UAV and established a solving framework for UAV autonomously collision avoidance and target tracking problem.Base on this solving framework,we do not have to forecast the future trajectory of the target and only the current state of target need to be estimated.Secondly,a target state estimation algorithm(D-PF)based on SIS which can take the UAV own state uncertainty into consideration is proposed for UAV autonomously collision avoidance and target tracking.A common fundamental necessity in solving the airborne collision avoidance and tracking problem is the need to estimate the current state of the intruder.This dissertation first modified the trun rate of the coordinated by mean-adaptive acceleration model and give the intruder motion model which consists of a modified 5-state coordinated turn model for the horizeontal filter and a decoupled 3-state Singer acceleration model for the vertical filter based on some prior knowledge of the aircraft flying in the nation air space(NAS).Further more,during the whole task,UAV use onboard sensor like radar to get measurements of the target,which are relative vector between the ownship and the intruder.Both of the measurement noise and the accuracy of the ownship's state will affect the state estimation of intruder.This dissertation presented a state estimation algorithm based on SIS particle filter which took the UAV own state uncertainty into consideration.The feasibility and performance of the proposed approach is verified through simulation.Thirdly,guidance law for uav collision avoidance and target tracking are designed based on Lyapunov method.First,a guidance law in the 2D for collision avoidance is designed.It begins with the geometric characteristic of the collision avoidance and chose the virtual point and the collision avoidance vector.Then convert the rendezvous with the virtual point into the realization of the parallel navigation and using the Lyapunov method to obtain the guidance law for collision avoidane.The guidance law has the inherent simplicity and robustness of the PN guidance law and can have shorter stabilizing time of the system than PN guidance law.In 3D space,analytic geometry is used to choose the collision avoidance vector among countless vectors which lead to minimum acceleration command.Guidance laws for vertical maneuvers and maneuvering horizontally are design respectively.For the target tracking guidance,in additionto the coincidence of a virtural target and UAV,the UAV's velocity equals to the virtural target.So,motion of the UAV was decoupled into two components,lateral and longitudinal and design the guidance based on the Lyapunov method which is similar to the collision avoidance law design.The feasibility and performance of the guidance law is not only mathematically analyzed,but also verified through simulation.Lastly,simulation experiment is carried on Gazebo platform which is based on the Robot operationg system to verify the feasibility of the guidance law.In the simulation experiment,we use two quadrotor aircrafts to take place of the UAV and the airborne target.We set three different initial conditions for the collision avoidance.First case is that two uavs fly face to face and uav avoid on the horizontal plane.The second and the third case are that two uavs will have a lateral encounter and uav avoid on the horizontal and vertical separately.Moreover,we let one uav to track an airborne target which is on a coordinate turing moton.We can conclude through the simulation results that the guidance laws given by our approach are valid.