Semi-autonomous Patient Transfer System Assisted by Computer Vision

The paper presents a system designed to improve the mobility of patients for transfer operations with a high degree of independence from a caregiver; as such the proposed system consists of a mobility device and a 3D camera that scans the environment from possible hazards and, if needed, take over the control to actively avoid them. The paper is focused on the integration of a machine vision system with a mobility device, with a special care for the design of the dual input subassembly, that has to retain the ability of being human-controllable and also to accept the machine-computed corrections. For this purpose, a custom-made compliant linkage was designed, in which the degree of compliance might be adjusted in concordance with the level of control the patient is able to accomplish. The paper also briefly discusses the possibility of integration of other means of human input systems, more useful for people with an even higher degree of disability than the hypothesized use case scenario, presumed challenges and possible solutions.