Research On Vision Based Quadrotor Indoor Location And Control System

Quad-rotor is a typical unmanned multi-rotor aircraft.Its three-dimensional attitude and three-dimensional position can be maneuvered quickly.Refer to properties of less effect caused by terrain,quad-rotor has good adaptability,which is extensively applied in military and civilian industry.At the same time,the quad-rotor system is a 6-DOF,nonlinear,under-actuated and coupled system.It poses a lot of difficulties to the design of control system.Hence,the research of quad-rotor control has been a hot spot in academic circles.At first,this paper studies a vision-based indoor location system which can discriminate image information of quad-rotor with an external camera and make the location information of quad-rotor by the position solution in host computer.Based on wireless bluetooth communication technology,the location information can be delivered to quad-rotor in real time to make autonomous flight.At the same time,a flight control experiment platform and an attitude control experiment platform are designed and set up successfully for indoor environment.A series of experiments including attitude control,hovering control and trajectory tracking control are given respectively to verify the effectiveness of the indoor vision-based location system.Finally,to further improve the control performance of quadrotor in special environments like indoor,this paper considers the case of flight velocity constraint,and presents a finite-time position tracking control for quad-rotor system with velocity constraint.The controller design steps are given.By skillfully constructing a suitable Lyapunov function and using the homogeneous domination idea,it is proven that the time-varying closed-loop system is finite-time stable.Results of simulation are given to verify the availability and advantages of proposed position controller with velocity constraint.The major innovations of this research are as follows:(1)In this paper,the parameter calibration of the vision-based location system can be reduced to a kind of multi-dimensioned complex optimization problem with setting up the database and establishing the cost function.Since the conventional optimization methods are easily appeared premature convergence and trapped in the local optimum,a beetle swarm optimization algorithm with adaptive mutation is proposed to optimize the parameters of the vision-based location system,which introduces a perception mechanism of surrounding for individuals,iterates with adaptive length and step size.Besides,the algorithm adjusts mutation probability according to population aggregation.It improves capability of local exploration and adaptability to different population density during different iteration period,which avoids premature convergence and local optimum leading to higher calibration accuracy.(2)Compared with the existing control methods for quad-rotor aircraft,the proposed finite-time position tracking control for quad-rotor system with velocity constraint in this paper can guarantee that the desired position trajectory can not only be tracked in a finite time but also the velocity is always satisfied the constrained condition.Moreover,the dynamic performance and anti-interference performance of the control system are greatly improved.