DEVELOPMENT OF ALGORITHM FOR DOOR DETECTION AND MANIPULATION BY MOBILE MANIPULATOR MOUNTED ON WHEELCHAIR

Abstract

There has been an increase in the number of people with disabilities either due to old age, disease or accident. These patients can be paraplegic (impairment in motor or sensory function of lower extremities), can have muscle diseases or have serious paralysis. In order to assist such people in their daily chores rehabilitation robots are being developed. With the use of these robots patients get a feeling of independence (as they no longer require the help from other people) and their self-esteem is restored. Intelligent wheelchair is one such robot which can be used for helping many types of disabled people. The user can command the wheelchair to navigate from one place to another place and the wheelchair is able to reach the destination without needing any more input from the patient. In an indoor environment the wheelchair has to navigate through doors in order to move from one room to another room. In order to successfully traverse through doors the wheelchair should be able to detect the door and properly open it. Therefore for a fully autonomous wheelchair door detection and door opening are important capabilities which increase its usefulness and make it effective for serving a variety of disabled people. This thesis is a part of overall research about the development of a fully functional intelligent wheelchair. The focus of this thesis is on the development of an algorithm for detecting doors, development of an algorithm to detect door opening direction and algorithms for manipulation of doors by mobile manipulator installed on the wheelchair. The door detection algorithm has been tested on real door images. The algorithm for detection of door opening direction has been developed and shows promising results. Two algorithms have been devised for door manipulation. These algorithms have been tested in OpenRAVE simulation environment.