Research And Implementation Of Key Technologies Of Home Service Robot Based On Baxter Robot

In recent years,more and more service robots come into people's lives.However,due to the relatively late start of domestic research on service robots,the integration of some existing technical means is not high.The intelligence of robot needs to be further improved.In this paper,a home service robot is developed based on Baxter robot,which integrates the technologies of double arm cooperation,machine vision,map building navigation,web page design and so on.Users only need simple operation to control the robot to grab medicine,water cup,food and other objects,and use the box to transport the items to the location designated by users.When nursing the old,the weak,the sick,the disabled and other people,to a certain extent,it reduces the workload of the nursing staff,and plays a certain role in protecting the safety of the nursing staff in the process of nursing some infectious patients.This design has certain research significance and practical value to improve the intelligent degree of home service robot.In this paper,based on the DH analysis method,the kinematic equation of the manipulator is analyzed,and the trajectory planning simulation of the manipulator is carried out with MATLAB.The kinematic constraint equations of the two manipulators in the cooperative work of the two arms are studied.On the basis of this theory,the effect of Baxter's right arm following the left arm's moving position is realized,and the function of Baxter's double arm cooperating and smoothly lifting the medical Kit is developed also.Secondly,the principle of camera calibration is analyzed.The Kinect camera and Baxter hand RGB camera were calibrated.A visual servo scheme combining Kinect long-distance positioning and Baxter RBG camera short-distance precise positioning is designed.After the analysis and comparison of the target detection algorithm,the surf algorithm which is more suitable for the target function of this design is selected as the target detection algorithm,Then,according to the needs of the robot,the identification and location of drug bottle and drug box are realized.Then,by analyzing the applicable scenarios of cartographer,Hector slam and gmapping algorithm,we finally choose the gmapping algorithm based on particle filter algorithm to build the laser map in the laboratory.Based on A*algorithm and dynamic window method,the robot automatic navigation function is realized.In the process of navigation,there is a problem that the number of continuous corners is too large to cause the objects inside the cartridge to roll over.In this paper,the improved A*algorithm is used to optimize the navigation function and reduce the corners of the global planning.It is proved by software simulation and physical verification that the improved algorithm can make the robot run more smoothly,thus effectively reducing the risk of overturning in the box during the navigation process before optimization.Finally,the client control web page is designed.The web page is composed of landing page,video monitoring page,autonomous navigation page and object grabbing page.The web page integrates the functions of automatic/manual grabbing objects,autonomous navigation and transportation of objects,as well as video monitoring of control effect.The on-line test of the client web page and the robot shows that the web page is available and effective.The user can control the robot conveniently through the web page,grab and transport the medicine,water cup,food and other objects to the user's designated location.