Research On Automatic Docking Technology Of Intelligent Wheelchair/bed

Nowadays,with the rapid increasing of social aging population and the rising demands for health care,it is of great social significance to develop intelligent service robots for the elderly and the disabled so as to improve their quality of life.An intelligent wheelchair/bed is an intelligent system with the functions of wheelchair and bed,which can be deformed and separated.The intelligent wheelchair/bed system(IWBS)is used to improve self-care ability of the elderly and the motor handicapped as well as reduce the labor intensity of nursing staff.Therefore,the IWBS has become a research hotspot in service robots for the elderly and the disabled.When the wheelchair docks with the auxiliary bed,the manual control docking requires the user to master skilled operation skills,which affects the improvement of user experience.Consequently,automatic docking becomes a basic function of IWBS.Based on the omni-directional IWBS platform,the automatic docking algorithm of a robotic chair/bed system is deeply studied in this paper.In this thesis,the remote vision guidance of the wheelchair,the short-range bed/chair docking guidance and the docking control method of the wheelchair in the process of automatic docking are studied deeply.The main research contents of the dissertation are as follows:1)A remote vision guidance method for the wheelchair based on the artificial landmarkFor the problem that the distance between the intelligent wheelchair and the assistant bed is too far to directly collect the marker on auxiliary bed for docking by using the on-board vision system,a remote vision guidance method for the wheelchair based on artificial landmarks is proposed in this paper.Firstly,in order to identify the wheelchair in indoor environment accurately,a new artificial landmark is designed.Secondly,an artificial landmark recognition algorithm based on geometric and color features is presented.A target state estimation tracking method based on Kalman Filter(KF)is used to narrow the scope of target search and then greatly improve the efficiency of landmark recognition.Then,the matching of feature points is realized by utilizing the deflection angle of the landmark and the relationship between points and lines.Finally,the localization algorithm based on P3 P is used to realize the remote vision positioning of the wheelchair.Experimental results demonstrated that this method can effectively identify and track targets and realize the positioning guidance of the wheelchair.2)A method of short-range docking guidance for bed/chair based on vision and ultrasound sensorsWith regard to the problem of short-range docking guidance for IWBS,a method of docking guidance for bed/chair based on vision and ultrasound sensors is proposed in this thesis.The method is divided into visual guidance docking stage and ultrasonic ranging guidance docking stage.In the visual guidance stage,for the reliability of chessboard corners extraction in automatic docking process,an improved chessboard corners detection method is proposed.Combining the improved chessboard corners detection method with OpenCV's chessboard corners detection method,the reliability of chessboard corners detection is better than the single OpenCV's checkerboard corner extraction method.For the problem of short-range visual localization between the wheelchair and the bed,a method based on plane homography matrix is proposed to determine the relative position and attitude between the wheelchair and the auxiliary bed.When the distance between the on-board camera and the marker on the auxiliary bed is getting more and more smaller,and the visual guidance method is invalid,it is switched to the ultrasonic ranging guidance phase.At this moment,it is assumed that the wheelchair has been adjusted to the desired pose in the visual phase.Then,the relative distance and attitude between the wheelchair and the bed are adjusted according to the information of ultrasonic ranging,and the docking guidance is realized.Several docking experiments have been conducted and experimental results show that the short-range docking guidance method between the wheelchair and the bed is effectiveness and robustness.3)Automatic docking control for the omni-directional wheelchair based on mecanum wheelsFor the automatic docking control problem of the omni-directional wheelchair based on mecanum wheels,the visual servo docking control based on staged stabilization and the visual servo docking control based on point stabilization are proposed.For the research of the visual servo docking control based on staged stabilization,the attitude stabilization control is carried out firstly according to the information of visual feedback.And then the position stabilization control is carried out.Finally,the wheelchair is docked with the auxiliary bed.Towards the research of the visual servo docking control based on point stabilization,the position and orientation errors are obtained by combining visual feedback information.And then a stabilization controller is constructed based on pose error.Finally,the stability of the closed-loop system is proved by the stability principle.Simulation and real-world docking experimental results show that two stabilization docking control methods can effectively achieve the control tasks from different initial poses.