Scenario Analysis for Road Following Using JetBot

Abstract

Recently, the application of Unmanned Ground Vehicle (UGV) has been increasingly prevalent. It can be used in different applications both indoors and outdoors. However, for some scenarios, the Global Positioning System (GPS) may not be always precise and available due to signal attenuation and multi-path propagation. Hence, in this research, we use a novel vision-based architecture, including machine learning and edge computing, to analyze the geometric features to find the path. Furthermore, we also intend to perform a scenario analysis for obtaining the setting with better performance. According to the experimental results, the prototype demonstrates the appropriate performance and properties.