Research On Hybrid Visual Servo Control Method For Mobile Robot

Visual servo refers to the closed-loop control of the robot by processing the image information which is obtained by the visual sensor as a part of the robot control system.Robot visual servo involves a wide range of aspects,which is an important topic in the field of robotics.In this paper,a hybrid visual servo system about mobile robot positioning and navigation is designed.The hybrid visual servo system involved in this paper is a double closed-loop control system: the closed-loop motion control system consisting of motor,encoder and PID controller is used as the inner loop;the closed-loop control system consisting of position and attitude estimation system,basic motion system and path planning system for mobile robot based on visual information is used as the outer loop.This article mainly carries on the systematic elaboration to the external ring:1.Aiming at the problem that the hybrid visual servo system relies heavily on the reference object,a method to reduce the dependence on the reference object by adding other positioning systems is proposed.The method of locating mobile robot based on dual-frame images can use natural images(without reference)to locate mobile robot.The extraction of target feature points is the key of this method.The SIFT and SURF feature point extraction algorithms are studied,and the SURF feature point extraction algorithm is improved.By comparing and analyzing the three algorithms through experiments,the performance of the improved SURF algorithm is better than that of SIFT and SURF algorithm.2.Based on single frame image,double frame(multi-frame)image and strapdown inertial navigation system(SINS)localization methods for mobile robot are studied.An extended Kalman filter(EKF)is introduced to design a multi-sensor data fusion localization method for mobile robot.3.A hybrid visual servo system is designed and simulated.The simulation results show that the system can achieve the positioning control of mobile robot with high accuracy.The experimental platform is built and the performance of the servo controller of mobile robot based on dual-frame(multi-frame)image,strapdown inertial navigation system and multi-sensor data fusion positioning system is compared.The experimental results show that the performance of the mobile robot servo controller based on multi-sensor data fusion positioning system is better than the other two methods.An artificial beacon used as a reference for hybrid visual servo system is designed.According to the structural characteristics of the artificial beacon,the detection and recognition method of the artificial beacon is studied.Based on hybrid visual servo system,a hybrid visual servo control system is designed by adding multi-sensor data fusion positioning system.The experiment is designed to test the hybrid visual servo control system.The experimental results show that the multi-sensor data fusion positioning system and the singleframe image positioning system complement each other.Multi-sensor data fusion positioning system can reduce the dependence of hybrid visual servo control system on reference objects.Single-frame image positioning system solves the problem that the cumulative error of multisensor data fusion positioning system is too large.