Mobility Assistance for Vulnerable Road Users using Machine Learning

This paper presents the research for the development of a new assistance system to increase the independent mobility of vulnerable road users (VRUs). This assistance system helps the elderly mobile users in virtually screening of directions using GPS, tracking of sidewalks, identifying the traffic signals and other sign hoards. MATLAB/Simulink serves as the main software used in developing this application. Images and other types of data are captured using sensors on Android devices. AlexNet is used to identify and classify different traffic warning signs in real-time. A detailed description of this method and the results are presented in this paper. Once the MATLAB-based program is developed, it can he converted into Java codes when needed. Using Android Studio, the code can be used in the application. VRUs with mobility impairments and vision deficiencies often find it difficult to use wheelchair on sidewalks. This paper also presents a sidewalk tracking system with a departure warning. In this research, Bough Transform is used to present the detection of sidewalk. This sidewalk tracking system can provide the user with essential information that can minimize the risk of an accident. Elaborated implementation of this system and results are presented in this paper. Rerouting the user when they approach the end of the sidewalk is left for future work. The application will further be developed to provide a voice navigation informing the departure warning, a traffic signal or a recognized sign board.