Study On The Control System Of Home Service Robot Based On Visual Technology

With the increasing development of society, the phenomenon of population aging and the Only Child becomes more and more serious, which leads to the lack of care for many Empty-Nest elders. It’s necessary to find a certain solution to provide daily basic service for the elderly. With the progress of science and technology, the robot industry has been developed rapidly. And with the policy support of our States for home service robot, researching on intelligent service robot appropriate for home will become inevitable.This subject relies on the Shandong Province science key Projects Program "The research on remote medical care monitoring platform applied to the intelligent home based on robotics", and studies the motion control system of home service robot. In order to make the robot more smarter, the robot can obtain unknown environmental information from various sensors so that it can localize itself and recognize specific target object through sensors, thus completing the appropriate services. This subject mainly research on technology related to motion control system of home service robot, and its major tasks is shown as follows:(1) We build a home service robot system, which is made up several major components they are mobile platform with two-wheel differential drive, modular serial six degrees of freedom operation manipulator, sensing subsystem, control subsystem, and describe each part in detail in this paper.(2)As the robot kinematics analysis is the theoretical basis of robot motion control, we established two-wheel differential drive kinematics model and odometer model as well as the forward and inverse kinematics of 6 DOF manipulator, then use the MATLAB software to simulate its forward and inverse kinematics and verify its correctness.(3) Introduce a method for creating environment map based on linear feature in detail, which uses laser range finders to perceive the unknown environment information. Through scanning lines split, and merging the data points, fitting, matching, it can complete creating two-dimensional environment map, the global map and local map’s matching. First, analysis each feature of laser localization method and odometer localization method separately. Then combine this two methods, and bring out a more accurate expression related to proportion of errors for mobile robot localization in order to achieve more precise position of the robot.(4)Introduce a particular SIFT image matching algorithm in detail, which is applied in home service robot for recognizing specific objects. It can complete the object recognition task through completing the similarities matching between the detected image and the stored image. A set of image matching experiments in different circumstances is also carried out, through these experiments we verify the SIFT algorithm of image matching is effective. In the end part, we bring out a robot motion control experiment, and we can see that home service robot can move to the target position smoothly and recognize the specific object, then complete the crawling operation.