Research On Pose Measurement And Self-Positioning Method Of Mobile Print Robot

Mobile printing robot is a high-precision micro-robot,which combines robot technology with printing technology to realize the mobile print.It solves some problems of traditional printer,such as large size,inconvenient to carry and limited printing area at one-time.To realize the high-precision printing,it is significant that high-precision control and location for mobile printing.Meanwhile,the high-precision position is the significant precondition of control.Using the machine vision to construct pose measurement system,which can detect the motion trajectory and provides a data base for the positioning experiment.It is of great significance to research the self-positioning method of mobile printing robot.In order to realize the high precision self-positioning of the robot,the motion mechanism of robot is analyzed,which include designing structure,constructing dynamic model and kinematic model.According to the motion mechanism,the sliding experiments of robot on the thesis surface are analyzed and steadily running conditions of the robot are obtained.To measure the pose of mobile printing robot,the visual measurement platform based on machine vision principle is built.The mapping relationship between image coordinates and world coordinates is constructed,which is based on monocular vision measurement principle.The world pose of target can be obtained through tracking the detected objects.In order to solve the slip phenomenon,BP neural network positioning system is constructed.To training the BP neural network,the distance of the moving trajectory and the corresponding pulse number are collected as training data.By using the trained neural network,the trajectory of the moving wheel could be obtained according to the number of pulses of the motor.According to the trajectory of two wheel,we can realize the self-positioning in the process of moving.In this thesis,the kernel correlation filter(KCF)algorithm is introduced to locate the target.Aiming at the characteristics of mobile printing robot,such as high moving precision,small detection target and fast motion,the KCF algorithm is improved and the adaptive kernelized correlation filters(AKCF)algorithm is proposed.To solve the influence of slip positioning error,BP neural network is adopted to realize the self-positioning.The motion trajectory and printing timing of the mobile printing robot are analyzed by using the self-positioning system.It is of great significance to realize high-precision mobile printing.The designed pose measurement system can realize the high precision pose measurement of micro moving targets.It can provide the data base for designing self-positioning system of robot.The proposed AKCF algorithm solves the boundary effect of the original KCF algorithm.Neural network is used to realize self-positioning and solve the accumulative positioning error caused by wheel skidding.