DESIGN AND DEVELOPMENTS OF ROBOTIC AIDS FOR DIFFERENTLY ABLED PERSONS

Abstract

A person with a physical or mental impairment that leads to functional limitations, activity restrictions, or a physical handicap is considered differently abled. This condition arises from the interaction between a person's bodily characteristics and their environment's societal features. These individuals are often perceived as having limitations in their daily activities due to various factors, some linked to the individual and others influenced by social, environmental, and political arrangements. In recent decades, the concept of disability has evolved from an individual impairment perspective to a more socially oriented phenomenon. The Multiple Disabilities category includes individuals with more than one disability, such as those with both visual and hearing impairments. This research aims to develop robotic aids for assisting persons with hearing disabilities during communication with the hearing persons. The hearing and speech-impaired community use sign language, which is a structured form of gestures, used for communication. To assist these individuals, the process of conversion from sign language into speech and text and vice versa is required. Sign language varies across different regions and countries, and each country has its version with different variations for communication. In America, ASL is used as a medium of communication among speech and hearing-impaired individuals. ASL involves both static and dynamic gestures, primarily using hand and wrist movements. When creating an interpreter system for those who have hearing disabilities, the first step is to understand the language and alphabet used. In India, Indian Sign Language (ISL) is the primary language used by those with speech and hearing disabilities. Unlike other sign languages, ISL has more compound signs and uses signing space differently. Unfortunately, many parents of deaf children are not familiar with sign language and its ability to facilitate communication. There is a great need for Indian Sign Language practitioners. However, the challenge is the shortage of practitioners in India despite the sizeable deaf population and the need for comprehensive training. A specialized robotic hand interpreter system must be designed to create an effective communication solution among hearing impaired and hearing persons. This system is designed to generate sign language gestures corresponding to the complete alphabet, encompassing both American Sign Language (ASL) and Indian Sign Language (ISL). Its main goal is to create unique gestures that reflect the spoken words of individuals who do not have speech and hearing impairments. Considering the proposed methodology for the Indian and American Sign Language Generation, the simulation of the robotic hand is achieved by utilizing an Arduino board, Raspberry Pi 4 and connecting various external components to it. Arduino programming is employed to provide instructions to the microcontroller on the board, indirectly controlling the movements of the hands to produce meaningful sign language gestures. Python Programming is also employed using Raspberry Pi 4 along with other external components to perform the simulation of the Robotic hand. Experimental trials involve both speech and hearing-impaired individuals, often referred to as sign language interpreters, as well as individuals with normal hearing. The side-by-side comparison of actual signs and generated signs of the alphabet is also discussed in tabular form in the experimental results section. The Identification of the static, dynamic and distinct signs of the alphabet from A to Z during the process of Sign Generation, both for American and Indian Sign Language Generation, is also presented. Furthermore, the identification of the list of signs of the alphabet from A to Z that cannot be generated by the Robotic hand correctly due to similarities in their signs for the American and Indian Sign Language systems is also observed. The error estimation is further calculated during the Alphabet Generation from A to Z for both American and Indian Sign Language systems. The error variation in the case of ASL Generation, involving the robotic hand, ranges from 2 to 3 percent. The accuracy of the interpreter system consisting of the Robotic hand, based on error estimation during the generation of various alphabet signs from A to Z, varies between 97 to 98 percent. In contrast, the observed error variation during ISL Generation ranges from 1 to 1.5 percent. The accuracy of the Robotic Hand in the Indian Sign Language System, based on error estimation during the generation of alphabet signs from A to Z, varies between 98 to 99 percent. Performance evaluations were conducted by ten different individuals for the Robotic Hand. The Robotic Hand Interpreter system demonstrates an overall accuracy of 90 percent in evaluations for American Sign Language Generation. Conversely, in the context of Indian Sign Language Generation, the overall accuracy is 90.53 percent. Emphasizing Word Generation, a three-dimensional robotic hand design is implemented to animate gestures in both American and Indian Sign Language. This robotic hand model is created using 3D animation software like Blender. The robotic hand produces animations corresponding to voice commands spelled out by users in ASL and ISL. Individuals with speech and hearing impairments then observe these animations on a computer screen. Impressively, the robotic hand interpreter system achieves commendable overall accuracy in generating sign language gestures for the entire alphabet from A to Z. Similarly, the three-dimensional animated robotic hand model demonstrates reasonable overall accuracy according to the experimental results.