Research On Obstacle-avoidance Algorithm For Quad-rotor Aircraft Based On Multi-sensor Information Fusion And Shape Identification

With the continuous deepening research for the quad-rotor aircraft in the field of geological exploration,agricultural production,logistics and transport,emergency rescue,people’s demand for the quad-rotor aircraft are improving constantly,so it is an inevitable trend to achieve autonomous flight control,and the autonomous obstacle-avoidance is the basis of autonomous flight control.In addition,the popularization of the quad-rotor aircraft has brought many security problems,such as collision and crash,which poses a great threat to personal and property safety.Therefore,it is of great significance to study the obstacle avoidance technology for protection of people’s lives,fuselage and buildings in time when the quad-rotor aircraft is sailing.Aiming at solving the obstacle avoidance problem of quad-rotor aircraft,a new obstacle avoidance algorithm based on multi-sensor information fusion and configuration recognition is proposed.The main contents of this paper are as follows:(1)The hardware platform of quad-rotor aircraft is built based on the STM32F407 main control chip,the MPU6050 inertial navigation unit and the GPS positioning system.The Kalman filter is used to fuse the range information of laser sensor and ultrasonic sensor,and the optimal forgetting factor b is obtained by using the Simulink module in MATLAB,then the Sage_Husa adaptive filter is designed.At last,the experiment of distance information fusion for three kinds of materials is carried out.(2)The research of obstacle shape recognition is carried out.The theory of obstacle shape recognition is introduced based on ultrasonic sensor three point linear arrangement method.According to the shortcoming,a new measure method that is ultrasonic sensor three point 120 degrees obtuse angle measurement method of isosceles arrangement is put forward.According to the five kinds of basic obstacle shapes,the measurement formula is derived by using the geometric method and the principle of shape recognition is described,which is the theoretical matting for the obstacle avoidance algorithm.(3)Based on the obstacle distance information and the obstacle shape information,the obstacle avoidance algorithm of quad-rotor aircraft is developed.Firstly,according to the fusion information of distance,the safety state of quad-rotor aircraft is judged,which determines whether the automatic obstacle avoidance system is open or not.Secondly,the obstacle shape is recognized automatically,based on the different shapes the different obstacle-avoidance measures to avoid obstacles are developed.Then,the PID algorithm is used to achieve the obstacle-avoidance behavior.Finally,the obstacle shape recognition and obstacle avoidance experiments are carried out.The experiments of multi-sensor information fusion show that compared with the data collected by a single sensor and original Kalman filtering,the data collected by the Sage_Husa adaptive Kalman filtering has the advantages of high accuracy,better convergence,more powerful and more stable filtering,and the new algorithm has very strong adaptability to different obstacles.The experimental results show that the obstacle avoidance algorithm can correctly identify the shape of obstacles,and the corresponding obstacle avoidance is completed to achieve the goal of avoiding obstacles.