Simulation of Autonomous Driving for a Line-Following Robotic Vehicle: Determining the Optimal Manoeuvring Mode

Abstract

Mobile robotic systems offer valuable test platforms due to their shared features with autonomous vehicles, including features such as sensor technologies, navigation algorithms, and control systems. However, constraints in laboratory environments or technical resources, along with the need for extensive testing, often necessitate the use of virtual test laboratories. While line-following is a widely preferred application in mobile robotics, research on this topic within virtual laboratories is limited. This study pioneers the use of a car-like robotic vehicle in conducting line-following tests within a virtual laboratory environment. To facilitate these tests, a virtual simulator was developed to meet the requirements of realistic simulations. This simulator includes simulated elements, such as roads and environmental features, along with virtual sensors designed to collect and process dynamic motion data. An exceptional aspect of this study is its ability to collect consistent dynamic travel data by sampling realistic sensor information within a virtual environment. The developed line-following algorithm employs a controller to minimise lateral deviation while the robotic vehicle follows a road line during its movement. The study conducted virtual driving tests using two different manoeuvre modes on four distinct road segments, exploring how the manoeuvring style influences the driving quality. It was demonstrated that in the low manoeuvre mode, the ride is more comfortable, but exhibits a greater route deviation due to reduced oscillation, while the high manoeuvre mode exhibits the opposite behaviour.