IRIS Orientation Based Mobile Robot Steering Using Regional Division of Human Eye

Abstract

Paralysis in human being is a major curse. Quadriplegia or Tetraplegia is a form of paralysis caused from injuries or illness results in partial or complete loss of all four limbs in this condition a patient can only move his/her eyes. Different types of interfaces with wheelchair had been introduce for paralyzed patients i.e. joystick control, chin control and head control. But many of the patients cannot even utilize these devices due to limited movement. Eye controlled wheelchair is also introduced but most of them are difficult to operate during steering due to IRIS positioning as well as camera positioning. In this work a solution for IRIS tracking and mobile robot steering is developed. For which first step is face detection, second eye extraction, third IRIS detection and regional division and fourth mobile robot control. Regional division of human eye is based on IRIS orientation for which we use individualized IRIS localization, in which first step is to find the natural IRIS orientation of patient. Initially subject is told to look straight, left, and right which are then used to mark the eye left positions in three regions, eye right positions in three regions and straight up to a certain limit depending on eye natural straight orientation. Then the marked regions are used to control steering angle of mobile robot. All the image processing was done on raspberry pi using open cv. Images were captured using Pi-cam. Raspberry pi after taking decision of IRIS localization in the eye frame send commands to the mobile robot using Bluetooth module. Arduino controlled mobile robot is developed which receives commands of movement from Raspberry Pi via Bluetooth which are then decoded and used to operate four-wheel robot.