Distributed Estimation And Fast Path Planning For Moving Target Tracking Using UAVs

Unmanned aerial vehicles(UAVs)have the advantages of low cost,no casualties,mobility,and usableness,which play an important role in modern military application as well as civil application.In the most of current research concerning path planning for UAV,the desired trajectory is preset for UAV before carrying out missions.Nevertheless,the desired trajectory cannot be obtained ahead in some cases,which is needed to be estimated online during the flight.Therefore,it is of significance to study the estimation of unknown moving target.In addition,in view of the high velocity of UAV,fast path planning is demanded to meet the requirement of high real time.The content of research can be divided into several parts.In the case of unknown moving target,squared-root cubature information filtering is applied to estimate the target state based on the measurement from sensor aboard for each UAV,resulting in prediction value and increments for information vector and information matrix.Then,considering the communication topology,appropriate weighting coefficients and iteration step are chosen to deal with consensus problem for the results obtained above,leading to more accurate estimation for moving target.UAV plans it trajectory using model predictive control.In view of the characteristic of online solving repeatedly and long calculation time,the fast model predictive control(FMPC)is studied in this paper to make it feasible to adapt the traditional model predictive control into UAV control system,meeting the real time requirement.The controlled object model and constraints are linearized.The objective function is reconfigured.And then a linear optimization problem is formed.Finally,the first-order KKT optimality condition and Newton step are used to update optimal variables.So the fast path planning of UAV is realized accordingly.At last,the algorithms for target estimation and fast model predictive control are combined to achieve the cooperative control of multiple UAVs for ground moving target tracking.Target estimation is related to UAVs' state and the path planning of UAVs has connection with the results of target estimation vice versa.The avoidance constraints between UAVs at the same level and the change of topology among UAVs at different levels are taken into account during path planning.The control block diagram and algorithm steps are also shown in the paper.The simulation results demonstrate the performance of the control system.